In December 2025, a report emerged that Foxconn had hired 30,000 people over 8-9 months at its new iPhone assembly plant outside Bengaluru. It was the fastest hiring ramp at any Indian manufacturing facility on record. The Devanahalli plant sits on roughly 300 acres. About 80 percent of the workforce is women, most of them between 19 and 24, with an average monthly salary of around Rs 18,000. Foxconn is investing close to Rs 20,000 crore in the project and expects to scale to 50,000 workers at peak production.

Within the same city, in a 13,000 square-foot lab, a company called CynLr, which has so far raised ~$15 mn, is doing something that looks like the opposite of Devanahalli. At each desk, an engineer trains a three-armed robot to pick objects it has never seen before.

Behind the screens, visualisations show how the robot sees and interprets each shape it encounters. The company, in 2024, had about 25 such robot systems and was on course to build 50 by 2026. Its first customers are Denso and General Motors. Across town at the prestigious Indian Institute of Science (IISc), a neuroscience lab is recording the neural activity of monkeys handling unfamiliar objects, in a partnership designed to reverse-engineer the principles of primate vision into industrial robotic systems.

These two facts coexist roughly within the same city, say, a few tens of kilometers apart. They appear to contradict each other. The largest single-location women’s workforce in Indian manufacturing has been assembled at the exact moment a credible Indian robotics ecosystem is starting to emerge.

The Devanahalli plant is not evidence that automation is irrelevant in India. It is evidence that the economics still favor manual assembly at this scale, today, for this product. iPhone assembly is high-mix, dexterous, quality-sensitive, and labor-cost-bound. The arithmetic that puts 30,000 women on a Foxconn line is the same arithmetic that has slowed industrial automation in India for two decades. What is different now is that the arithmetic is changing, and changing (very) fast, in ways that the previous generation of Indian robotics companies did not get to benefit from.

This document is our attempt to lay out what has changed, why the next ten years look genuinely different from the last ten, and which specific kinds of companies we want to back.

II. The ‘Prior’ Generation

We are using the word ‘prior’ very cautiously, and we will explain why. But first, a story.

In 2011, two BITS Pilani students named Akash Gupta and Samay Kohli founded a company called GreyOrange. They had built India’s first humanoid robot, AcYut, to compete in robo tournaments globally. They won a gold medal at the San Francisco RoboOlympics. The duo thought robotics had a future in India.

GreyOrange raised a $500,000 seed from Blume Ventures and a few angels and built its first product, Butler, an autonomous rack-moving robot for warehouses. The story often told is that the company struggled to find Indian customers and pivoted to the US as a result. The actual story, as Akash Gupta has recounted in Blume Ventures’ portfolio article, is more interesting and more useful for thinking about Indian robotics in 2026.

GreyOrange’s first major customer was Flipkart, introduced through a college senior in 2012-2013. The Indian e-commerce boom that followed pulled GreyOrange along with it. As Akash described the trajectory in that commentary: “I remember at the peak, we had to install more than 40 systems. And before that, we only had 4 systems. So, we scaled rapidly from 4 to 40+. By 2015-2016, we had installed sortation systems in any place that you could in India. You could name an e-commerce company or a courier company; we had a sortation system installed there.”

The customer list extended to Delhivery, Jabong, and the broader Indian e-commerce and logistics ecosystem. Indian industrial robotics outside automotive had a real customer base by 2015. GreyOrange served it.

The constraint that GreyOrange ran into was not the absence of customers. It was the size of the Indian fulfilment market at that point on its growth curve. After GreyOrange had captured most of the major buyers in India, the question became how to scale further. India in 2016-2017 did not yet have the warehouse density or operating tempo to support a unicorn-scale fulfilment robotics company. The team turned to international markets to find the scale that India did not yet offer.

The first international stops were not the US. They tried Japan, Europe, and Chile. The Japanese experience produced one of the most instructive episodes in Indian deep-tech export history.

As Akash describes it in the same commentary: “I remember we shipped the first batch of 70 robots to one of the customers in Japan. And in Japan they do a white room analysis where you keep the bot in a white light and look at it. Soon we received a long email with a picture of every bot and things like, ‘here is a scratch,’ ‘this bot head seems to be a little bit damaged,’ and more. That’s absolutely next level of attention to detail.”

The team then concluded that the only international market with both the scale and the customer-development culture to support the business at the size they needed was the United States.

In 2018, after a $140 million Series C, GreyOrange moved its operational headquarters to Atlanta. The company sells today to H&M, Adidas, Walmart, IKEA, and Apple. Per the same Blume commentary, GreyOrange had then crossed roughly $100 million in revenue and 15,000 deployed robotic agents. It is, as of date, without a shred of doubt, the most successful Indian-origin industrial robotics company.

The deeper lesson of the GreyOrange journey, and the one most easily missed, is that the bigger pivot was not merely geographic; it was rather operational. The first US deployment, a 300-robot project with GXO Logistics, taught the founders that having robots up and running is one thing, and ensuring they actually improve warehouse operations is another beast altogether. This insight produced the pivot from hardware-led deployment to software-led orchestration. The company’s central product today is GreyMatter, an AI-driven warehouse orchestration platform. The robots are still there, but the bigger moat is the software that runs them.

Two things from this history matter for the thesis of this document. First, India in 2013-2016 had real customers for industrial robotics in fulfilment but did not have enough of them, at sufficient volumes, to support a unicorn-scale company.

Second, Indian manufacturing quality in that era was still ‘learning’, for lack of a better word, which constrained the international markets a hardware-led Indian company could credibly serve. Both of those constraints have eased materially over the last decade, and the easing is what makes the current Indian robotics opportunity different.

Worth a separate clarification before moving on. The ‘lazy’ version of the GreyOrange story sometimes gets extended to claim that Indian manufacturing in 2017 was not buying industrial robotics at all. That version is certainly wrong.

Indian automotive was the country’s single largest robotics customer that year, accounting for roughly 45 percent of national robot installations in 2024 per IFR data and a higher share in earlier years. Maruti Suzuki’s Manesar plant, commissioned as a greenfield site in 2007, was running roughly 2,500 industrial robots by 2017. Hyundai Chennai, Renault-Nissan Oragadam, Volkswagen Chakan, Toyota Bidadi, Tata Steel Kalinganagar, and Samsung Noida were all greenfield facilities running automation, supplied by FANUC, ABB, KUKA, and Yaskawa for two decades. The robots were going somewhere. They were going to greenfield Indian factories.

III. What Actually Changed

Four things have shifted since GreyOrange made its pivot decision. They are a) dated, meaning you need to move in quickly, else you will be late to the party, b) specific, and c) reinforce each other.

The economics of robot deployment have transformed: McKinsey’s industrial-robotics practice wrote in mid-2024 that historical robot ROI cycles of five to eight years through the 1980s-to-2020s period have compressed to one to three years today. This stat, though, is for ‘American Manufacturing’. I am taking the liberty to say that, if not the exact number of years, the ‘compression in years’ trend must hold true for India and the rest of the world, for that matter.

A five-year payback is a board-level capex decision that competes with every other long-cycle investment in a manufacturing company. A 1-3-year payback period is a procurement decision a plant manager can make. The compression has been driven by falling hardware costs, the embedding of AI in robot control, and lower integration costs as the deployment stack matures.

The greenfield manufacturing wave is bringing new sectors into India at unprecedented scale and specification: India did not start building greenfield manufacturing in 2020. Maruti’s Manesar plant was commissioned in 2007, while Suzuki Motor Gujarat (Hansalpur) opened in 2017 at a million-unit annual capacity, Hyundai Chennai, Tata Steel Kalinganagar, Reliance Jamnagar, and Samsung Noida (the world’s largest mobile-phone manufacturing plant at its 2018 expansion) are all greenfield Indian facilities running automation, built well before PLI. What changed in 2020-2026 is not ‘greenfield’ as a phenomenon, but it’s the sectoral composition, scale, automation specification, and geographic spread of the new wave.

As of March 2025, committed investments under the PLI scheme had reached Rs 1.61 lakh crore across 14 sectors. The scheme has had its share of misses, and India’s manufacturing share of GDP has slipped to 14 percent from 15 percent at scheme launch.

However, all this activity has introduced entirely new categories of manufacturing to India at scale: semiconductor packaging (Micron’s $2.75 billion ATMP at Sanand, inaugurated by the Prime Minister in February 2026; Tata’s first fab at Dholera due in 2028), EV cell manufacturing, and advanced electronics packaging at iPhone Pro Max quality at Foxconn Devanahalli. These are sectors that did not exist in India at a meaningful scale before 2020, and their automation specifications are substantially higher than Indian automotive’s historical bar.

This matters for who wins the procurement. The traditional industrial robotics product- FANUC welding robots, ABB paint cells, KUKA material-handling arms- was optimized for automotive body-in-white and powertrain work, where the supplier relationships at Indian OEMs have been locked in for two decades. The new wave needs different categories of robotics: vision-led manipulation for electronics assembly at sub-millimetre tolerance, autonomous mobile robots for semiconductor cleanrooms, AI-driven defect inspection at pharma blister-pack resolution, force-controlled handling for fragile components, among others.

These are the categories where Indian startups can aim to win big. The new opportunity is that India is buying new categories of robotics, in greenfield procurement environments where incumbent supplier relationships from the auto era do not transfer.

Physical AI has crossed the deployability threshold for unstructured environments: Indian factories outside automotive were the wrong customer for the last wave (now behemoths in their respective target sectors) of companies. What changed, between 2022 and 2025, is that foundation models for robotic manipulation began to work.

Covariant, the UC Berkeley-linked robotics AI company behind RFM-1, one of the first commercially deployed robotics foundation models for warehouse manipulation, was functionally pulled into Amazon’s robotics orbit in August 2024 when Amazon hired its founders and part of its team and signed a non-exclusive license to Covariant’s robotic foundation models. Google DeepMind’s RT-2 showed in 2023 that vision-language-action models could transfer web-scale visual-language knowledge into robot control. NVIDIA’s Project GR00T, announced in 2024, evolved into a broader humanoid robot foundation-model platform, with Isaac GR00T N1 released in 2025 as an open, customizable foundation model for downstream humanoid robotics developers.

None of these existed when GreyOrange was making its India-versus-US call. All of them exist now, and the result is that a robot deployed into a Tirupur garment unit or a Sanand cell-assembly line does not have to be told what each object is. It has to be told what to do with the objects it sees. The chaos of an Indian factory floor outside the FANUC-served automotive segment is no longer a barrier to robotic deployment. It is the natural deployment environment for the new generation of AI-native robots.

The specific constraints of the 2010s era have eased materially: This shift has two components.

The first is that the Indian fulfilment and e-commerce market in 2015-2016 has grown by an order of magnitude. Quick-commerce companies that did not exist then- BlinkIt, Zepto, Swiggy Instamart- now operate at breakneck tempos that materially change warehouse-automation economics.

Meesho, a small social-commerce experiment in 2016, is now one of the largest e-commerce platforms in the country. Reliance Retail’s acquisition of a controlling stake in Addverb in 2022 for $132 million is a direct measure of what changed: a single Indian retailer made a single warehouse-automation acquisition for a sizeable chunk of money.

Addverb has since reported Rs 334 crore in revenue for the year ending March 2025, with customers including Reliance, Unilever, Flipkart, Coca-Cola, and PepsiCo. The Indian fulfilment buyer base in 2026 is broader, better-capitalised, and operating at volumes that justify automation in ways the 2015 ecosystem did not.

The second is that the manufacturing quality bar inside India has risen meaningfully. The current PLI wave is changing this directly. Foxconn’s Devanahalli facility near Bengaluru has begun assembling iPhones for Apple as part of Apple’s India manufacturing ramp-up. Micron’s Sanand facility is a semiconductor ATMP plant for DRAM and NAND, operating assembly, packaging and testing processes rather than conventional electronics assembly. Tata Electronics’ Hosur facility and its acquired Karnataka iPhone assembly operations are now part of Apple’s India supply chain, covering precision components, enclosures and iPhone assembly.

The Indian manufacturing base of 2026 is producing at quality thresholds the Indian manufacturing base of 2017 could not consistently achieve. For Indian robotics companies that intend to manufacture hardware in India and export globally, this means the path to Japanese, German, or Korean OEM customers is open.

Run alongside these four shifts, two more developments deserve mention. The first is that India’s deep-tech funding has roughly doubled in 18 months. Tracxn data reported via the CFA Institute shows Indian deep-tech companies raised $1.06 billion in equity across 137 rounds in the first seven months of 2025, double the same period in 2024, and on pace to exceed the $1.6 billion full-year 2024 total.

Within that, Indian robotics startups raised approximately $117 million in 2024, versus $54 million in 2023 and $28.8 million in 2022. The second is that the Government of India released a draft National Strategy for Robotics in 2023, naming MeitY as nodal agency and aiming to position India as a global robotics leader by 2030. Government strategy documents are easy to overrate, but the existence of one creates institutional sponsorship and capital flows that did not exist even half a decade ago.

IV. The Core Argument

Leg one: Indian factory chaos is the moat, and the export market spans the entire emerging-market industrial footprint

Most Indian robotics theses treat the Indian factory’s unstructured nature as a problem to solve. The honest reading is that a robot designed to work in a German automotive plant assumes jig-fixed parts, structured lighting, climate-controlled air, stable voltage, and a workforce trained to ISO standards. A robot designed for a Tirupur unit or a Sanand cell-assembly line has none of those. It has to handle variable parts, monsoon humidity, voltage dips, mixed SKUs, and operators reading the interface in ‘not their native’ language. Solving for those conditions produces a more general robot than solving the German problem.

That robot has a market beyond India. The traditional industrial robot leaders continue to dominate global industrial robot shipments and have spent decades calibrating their R&D to structured, precision-mechanical environments. However, across the emerging-market manufacturing footprint that spans Southeast Asia, Sub-Saharan Africa, and Latin America, in cities like Lagos, Jakarta, Dhaka, and Hanoi, the established robotics OEMs have yet to build something very substantial for ‘chaos’. The companies that get this right in India will inherit those markets by default. The Indian winners are all set to potentially win markets the behemoths have never seriously tried to enter.

CynLr’s stated goal of what they call a ‘Universal Factory’ is an expression of this thesis at one extreme: a production line that can profitably handle custom, variable orders without retooling. Genrobotics’ Bandicoot, a robot designed for the uniquely Indian problem of manual sewer cleaning, is now deployed across 21 Indian states with 300 units in active operation per the company, and has signed a deployment agreement with Dubai Municipality. Both companies are expressions of the same underlying observation: what works for India works globally.

Leg two: The vertical data moat beats the horizontal model moat

It will be hard for Indian startups to out-compete Covariant, GR00T, or RT-2 on horizontal robotic foundation models (at least for now, never say never). They presently do not have the compute, the talent density, or the frontier-lab culture.

The defensible Indian play is vertical. Foundation models trained on data that only Indian companies can plausibly own. Say, textile-thread defect imagery from Tirupur and Bhilwara, PCB solder-joint failure data from Sriperumbudur and Noida, pharma blister-pack anomaly libraries from Hyderabad and Ahmedabad, or even Agri-spraying target data from Punjab and Maharashtra. The model architecture is commoditizing, with weights and architectures published openly by DeepMind, Meta, NVIDIA, and Physical Intelligence. The proprietary manufacturing dataset behind any deployed model is not commoditized, and the Indian manufacturing footprint generates that dataset every working day.

The diligence question for any AI-vision robotics startup pitching us reduces to one filter. Do you own a dataset that frontier labs cannot reproduce? If yes, the moat is real, and the company is investable. If no, the company is a fine-tuning wrapper on someone else’s model.

Leg three: Cobots are the wrong form factor for the Indian MSME wedge

We believe that this is the contrarian position in this thesis, and we want to state it as such. We are either proven right or wrong.

The Indian robotics consensus rests on a single assumption: that affordable collaborative arms are the wedge into the MSME market. The pitch is logical on its face. Indian SMEs cannot afford a Rs 40 lakh KUKA cell, but they can afford a Rs 10 lakh cobot. Per MarketsandMarkets data published in May 2025, the global cobot market is projected to grow from $1.42 billion in 2025 to $3.38 billion by 2030 at a CAGR of 18.9 percent.

We think this consensus is wrong for India, and we are willing to say so before the evidence is complete.

A cobot’s defining feature is safe collaboration with humans in a shared workspace. That feature solves a problem MSMEs do not actually have. A Coimbatore auto-ancillary unit does not need a human and a robot sharing a two-meter work envelope. It needs a cheap, simple, perception-equipped machine doing one task in a corner of the shop floor. The collaboration premium is engineering paid for capability the customer does not use.

The MSME wedge into Indian manufacturing is not a Rs 10 lakh cobot. It is closer to a Rs 3 lakh smart fixed cell with embedded vision and a single-task control loop. The product that wins the Indian MSME market in 2030 will look less like a Universal Robots UR5e and more like a smart sewing machine, a smart inspection station, or a smart screwdriver: dedicated machines that happen to have AI vision inside them.

If we are wrong about this, we will have screened out one of the larger categories in Indian robotics. If we are right, Peer will be the fund that took the position before the consensus moved.

V. Two Paths, Two Funds

The most important argument in this thesis is structural. There are at least two kinds of robotics companies that will emerge from India in the next decade, and they require completely different fund strategies. Trying to back them with a single set of diligence criteria is not what we want to do.

Path A: The deep-tech full-stack company

The archetypal Path A company is CynLr. It builds its own hardware, its own perception stack, and runs a lab with custom robotic test rigs. It also sources 400-plus parts across 14 countries. It has opened a design and research centre at the Unlimitrust Campus in Prilly, Switzerland, partnered with EPFL LASA and CSEM. It has a research collaboration with the IISc Centre for Neuroscience that funds PhDs and builds a researcher pipeline.

Ati Motors is the second archetype. Founded in 2017 at IISc Bangalore by Saurabh Chandra, the company builds autonomous mobile robots for industrial material movement. In January 2025, it raised a $20 million Series B led by Walden Catalyst Ventures and NGP Capital, bringing total funding to roughly $37 million. Its Sherpa robots have travelled hundreds of thousands of kilometers across 50-plus factories globally. Customers include Bosch, Forvia, and Hyundai. The company operates in India, the US, Mexico, Thailand, Singapore, and Vietnam. Chandra has said directly: “We do the full stack ourselves.”

A Path A company has these prerequisites, none of them optional.

Full-stack ownership of hardware, perception, and integration. The chaos problem cannot be solved by software bolted onto someone else’s robot. The hardware tolerances, the actuation profile, and the sensor placement co-evolve with the perception model. The arms that ABB and FANUC built for automotive jigs cannot be made to behave intelligently in a Tirupur unit by adding an API on top.

Institutional R&D partnerships with leading academic centers. Industrial robotics needs years of foundational research per usable engineer, and the usual “drop in for a hackathon” partnership does not work here. A Path A company without a serious academic anchor is operating with a structural disadvantage.

International presence early, primarily for talent. The Indian-domiciled supply of senior robotics engineers is insufficient. The Switzerland office, the Texas office, the small Lausanne research outpost are how Indian companies access engineers who left for global frontier labs and aren’t moving back to a pure-Bengaluru employer. Gokul NA, co-founder and head of product, design, and brand at CynLr calls it “casting a net to catch them all.”

International design-in customers from year one. Denso and GM at CynLr. Bosch, Forvia, and Hyundai at Ati Motors. These are not ‘export markets’ in the conventional sense. They are co-development partners who validate the technology at global manufacturing standards while the Indian market matures. A Path A company that pitches “win India first, then go global” is misreading the market. The correct sequence in 2026 is global design-in, Indian deployment-scale, then commercial expansion to emerging-market manufacturing.

Long-horizon capital structure is a given. Path A is a 7-to-10-year capital cycle, and Series A check sizes in the $2-5 million range. Follow-on reserves of 2-3x. The deep-tech robotics company that delivers an exit in five years is the exception, not the model.

The Path A exit profile is either strategic acquisition at a $300-million-plus valuation in the Covariant mould, or domestic IPO in the GreyOrange mould. CynLr has publicly stated targeting break-even before 2030 and a $75 million fundraise toward IPO. GreyOrange is on a similar trajectory at a much larger scale, obviously.

For a fund of Peer’s stage, the Path A allocation is the highest-conviction, longest-hold portion of the robotics portfolio.

Path B: The vertical AI software company

The Path B archetype is a company that does not build hardware at all. It builds a vision model fine-tuned on a specific Indian manufacturing dataset, deploys it as software on existing line cameras, and sells a subscription.

The customer is a pharma exporter who needs blister-pack inspection at audit-trail standards, or a textile exporter in Tirupur who needs thread-defect detection across hundreds of SKUs and shift patterns, or a PCB assembly line in Sriperumbudur that needs solder-joint defect classification at sub-millimetre resolution.

The product is software, and the hardware is already in the customer’s factory. The integration is API-level, and the unit economics look like SaaS rather than industrial automation. A vertical AI vision company can reach $2-3 million ARR with thirty contracts at $5,000-10,000 per month, in eighteen months, without raising a Series B.

A Path B company has different prerequisites.

Deep vertical domain expertise from day one: Where generic computer vision is commoditized. A team that has spent, say, a decade or so in Indian pharma manufacturing or in Indian textile QA has access to defect taxonomies, regulatory contexts, and operational quirks that an outsider cannot replicate by hiring an ML engineer.

Proprietary dataset ownership: The same diligence question as Path A, applied at the SaaS layer. Do you own image data that frontier labs cannot reproduce? If yes, this is a real moat. If no, the company is a wrapper.

Fast revenue path: The Path B founder should be able to articulate a path to $1 million ARR in 12-18 months without raising more capital. The economics of vertical AI software in Indian manufacturing favour the company that can sell into tens of factories quickly, not the one that spends 18 months tuning a model on synthetic data.

Smaller capital requirements, faster exit timelines: Path B seed checks are in the $300K–$1.5 million range, and Series A in the $2-4 million range. Exit profile is strategic acquisition incumbents, unless there is a rare breakout company.

The Path B company will not produce a GreyOrange-scale outcome. It will produce a steady series of millions exits, with shorter holds and lower capital intensity. For a seed-stage fund with disciplined check-size discipline, Path B fills the part of the portfolio that produces faster DPI.

VI. What Peer Looks For

We will write Path A and Path B differently, with different check sizes, and different time horizons. The honest framing is that we are running two parallel theses in one sector, and we will tell founders which path we think they are on at the first meeting.

For Path A, we look for:

A founding team with deep robotics systems experience, not generic ML credentials: The founder of a Path A company should have spent meaningful time on factory floors. These are people who have been working on this problem for ten-plus years and have the patience to keep working on it.

A working demonstration in a real factory, not a lab: A robot that has run for 200-plus hours of continuous operation in a customer environment is investable. A robot that has done a successful 90-minute demo at an investor’s office is not.

An institutional R&D anchor: A research collaboration with IISc, IIT Madras, IIT Bombay, IIT Hyderabad, or equivalent. The partnership should fund students, build a researcher pipeline, and produce defensible IP. If the company has no academic anchor, we will ask why, and we will treat the absence as a significant negative signal.

A globally credible first-customer relationship: One Fortune-500 manufacturing customer who is paying for a pilot is worth ten SME letters of intent.

For Path B, we look for:

A vertical domain expertise that produces a defensible dataset: We are interested in founders who can answer the data moat question without flinching. We are not interested in founders pitching us a “horizontal AI platform” that happens to be applicable to manufacturing.

A revenue model that does not depend on raising the next round: A Path B company that needs $20 million to reach commercial scale is misreading the category. The right Path B company gets to break-even on Series A capital.

Customer evidence within 12 months of seed: We will not back a Path B vision SaaS company that has not converted a paying customer within twelve months of our investment. The bar here is faster than it would be for a horizontal SaaS company because the manufacturing buyer cycle, while real, is not the binding constraint at the deployment scale these companies target.

For both paths, we look for one common thing. Founders who can articulate why their company could not be built in Germany or the United States. The thesis of this entire document is that there is something specific about the Indian manufacturing context that creates structural advantage for the right kind of Indian robotics company. A founder who cannot answer this question is, at best, building a derivative product.

VII. What Has Not Changed

The pieces I have written for Peer in the past have always tried to be honest about constraints. This one is no different. There are things about the Indian robotics opportunity that have not changed and will not change in the near future, and they deserve to be named.

The Indian labour arbitrage has not closed. Per Apollo Academy’s September 2024 wage comparison, Indian manufacturing wages sit at roughly 3 percent of US manufacturing wages, versus China at 20 percent and Germany at full Western levels. The Apollo data drew on ILO and China NBS sources, giving the figure institutional grounding.

Blue-collar wages in India are growing at 5-6 percent annually per Deloitte’s 2025 workforce report, which is meaningful but not transformative. The Devanahalli paradox is a paradox because the economics of manual assembly still pencil out at scale. Founders pitching us a ‘wages-are-rising’ argument as their primary demand driver are misreading the data. The case for Indian robotics in 2026 is not that labour is becoming expensive. It is that automation is becoming radically cheaper, that greenfield specification windows in new sectors are open, and that AI has eliminated the technical barriers that constrained the previous generation in unstructured environments.

The component import dependency is real. Gokul’s line in an earlier interview, “we are trying to build a car, but what’s available are cycle wheels,” captures the structural problem. CynLr sources 400-plus parts across 14 countries. Indian actuator, encoder, force-torque sensor, and LiDAR supply is thin to non-existent at industrial grade. Any robotics company building in India in 2026 is operating with a supply chain handicap that does not appear on the pitch deck but will appear in every BOM negotiation. There may be a separate thesis here on Indian component manufacturing, and we are watching that space, but we are not pretending it is solved.

The senior talent supply is genuinely scarce. India has perhaps a few thousand engineers with hands-on industrial robotics systems experience. We expect every Path A company to report this constraint. The Indian-origin engineers at NVIDIA, Tesla, Boston Dynamics, and Covariant who are starting to come back are the actual founder pipeline for the next decade. We expect this constraint to ease over five to seven years as the ecosystem matures.

The sales cycle remains long. Manufacturing procurement involves plant managers, engineering teams, finance approval, and corporate procurement at larger organizations. Six to eighteen months from first conversation to signed contract is normal. We will not back companies that have planned for ninety-day enterprise sales cycles and run out of patience before they reach revenue.

The incumbent response is coming. ABB, FANUC, KUKA, and Yaskawa are not ignoring India. As the new-sector greenfield wave builds, they will localise, acquire, and price-compete in categories outside their historical automotive base. The Indian companies that build on product differentiation alone, without customer relationships, proprietary data, or switching costs, will face a harder competitive environment in five years than they do today. This is a feature of the opportunity, not a bug. A market without an incumbent response is a market that has not arrived yet.

VIII. The Invitation

The companies we want to back over the next five years already exist, in one form or another. Some of them are founded. Some of them are still inside ABB India, FANUC India, NVIDIA Bengaluru, and the IIT-IISc research labs. Some of them are engineers who left India a decade ago for global frontier labs and are starting to ask whether the next thing they build should be back home.

If you are building in industrial robotics or physical AI in India, in either the Path A or Path B mould, we want to hear from you. We are especially interested in founders who can articulate clearly why their company is built for India’s specific manufacturing context and why that context produces a globally competitive product. We are interested in founders who can answer the data moat question. We are interested in founders who treat the Indian factory’s chaos as input rather than constraint.

We are not interested in funding the next affordable cobot, the next horizontal robotic foundation model, or the next Indian-priced version of a Western product. We are not the fund for those companies, and we will say so on the first call.

GreyOrange built a real company that found real Indian customers, ran into the TAM limits of the Indian fulfilment market of 2015-2017, and pivoted internationally to find scale, eventually landing in the US.

CynLr, Ati Motors, Genrobotics, Addverb, and the next generation of companies that have not yet been named are building into a markedly different environment: more Indian customers across more sectors at higher operating tempos, manufacturing quality bars that can clear export inspection, and physical-AI infrastructure that did not exist even half a decade ago. The question is which of them are building Path A and which are building Path B, and which of them are right about both the technology and the timing this time around.

We are betting that more of them are right than wrong, and that being early to this category will look in retrospect like being early to Indian SaaS in 2014. We are willing to take that bet.

Subscribe now

Building India’s Manufacturing Backbone, One Design, One Factory at a Time: The Bidso Story

When Vivek Singhal, then the GM for sports, fitness, and auto accessories at Flipkart, walked through factory floors in Sri Lanka’s cycle manufacturing belt during the early part of this decade, something clicked. He was there on behalf of a brand he had helped build, right from sourcing components, vetting suppliers, to negotiating specs.

The factory owner pulled out a catalogue. Five frame options, half a dozen wheel variants and about a dozen colour choices. Component by component, the product was assembled. It was all so seamless.

While at Flipkart, Vivek helped launch multiple private-label products from scratch. To do this, he travelled extensively across China and Sri Lanka to understand how sourcing really worked at the ground level.

But the question that kept nagging him was: why was manufacturing in India so broken, and why did every brand have to independently solve the same problem?

The numbers tell a rather stark story. China commands an astounding 65% of the global consumer goods supply. India has less than 2%. Mind you, this isn’t because of a lack of demand, capital, or talent; it is rather a lack of manufacturing infrastructure that actually works.

The answer, Vivek realised, was that nobody was building the capability spine that Indian manufacturing desperately needed. So he decided to build it himself.

In March 2026, Bidso announced a ₹63 crore Series A round. Blume Ventures joined us in the round along with Sadev Capital and venture debt from Alteria Capital. It is the latest milestone in a journey that began, as most good ones do, with a founder who could not stop asking why.

The Flipkart Crucible

Vivek’s entrepreneurial instincts preceded Flipkart. He had already run a business that didn’t fly. He then joined Flipkart with a clear-eyed understanding that this was a pitstop, not a destination. “I knew I would get out again to do something on my own,” he says.

But Flipkart turned out to be more than just a ‘pitstop’. It was a masterclass. Managing categories between 2018 and 2021 meant Vivek sat at the intersection of brands, consumers, and supply chains. He could see, with unusual clarity, a pattern that most people in the e-commerce world were too busy to notice: the brands listed on Flipkart and Amazon were struggling with supply and not demand. Finding reliable manufacturers who could deliver consistent quality, at competitive prices, on time was the real bottleneck choking India’s consumer goods ecosystem.

When a brand owner he knew casually told him, “You know what, why don’t you manufacture this? I buy it from China today. If you make it, I will buy from you.”

That conversation stayed with him. For Vivek, this was a clear market signal, and eventually, the one he could not ignore. By the time he left Flipkart, the question had shifted from whether to act on what he had seen to how far he could take it.

Vivek figured that the real opportunity lay beyond just ‘running’ factories. It was in aggregating supply and demand across an entire ecosystem.

The Band Gets Back Together

Rahul Agarwal didn’t need convincing. He and Vivek went back to XLRI Jamshedpur, same batch, same hostel corridors. They shared the kind of history that compresses a pitch deck into a single conversation over chai.

At Udaan, leading sourcing for hardware and kitchenware. During the course of his role, he saw firsthand how unreliable manufacturing throttled even the most aggressive distribution engines.

Where Vivek saw the problem from the brand side, Rahul had lived it from the demand aggregation side. The same dysfunction, viewed from two ends of the pipe. When Vivek laid out the idea, a platform layer that sits between brands and factories, Rahul saw what he had been experiencing, reframed as a business.

“For me, the unlock wasn’t just fixing supply,” Rahul says. “It was about making demand predictable for manufacturers and supply reliable for brands at the same time. Nobody was solving both sides together.”

Aditya Krishnakumar, another XLRI batchmate, joined as the third co-founder, bringing a deep focus on product and design. Where Vivek and Rahul had identified the supply-demand gap, Aditya pushed the ambition further: what if Indian manufacturing didn’t just supply products, but created them?

His lens expanded Bidso’s trajectory, from a manufacturing platform to a design-led product company.

The founding team bootstrapped for the first year. Bidso was formally incorporated in January 2023, but the intellectual work had been done months earlier. By September-October 2022, the team had a document that laid out how the RFQ system would work, how brands would interact with the platform, and how the backend would match manufacturing capability with incoming demand.

The business plan was a blueprint drawn from years of watching the system from the inside. From the outset, the founding team’s roles were complementary: Vivek on supply and manufacturing, Rahul on demand and commercial strategy, and Aditya on product and design.

“The core has always been the same,” Vivek says. “Supply product lines - white-label product lines—to brands. How we have gone about it, the GTM, that’s changed. But the core business has remained the same since we started.”

(L to R): Vivek Singhal, Aditya Krishnakumar, Rahul Agarwal

Two Relationships, Two Languages

Running Bidso means managing two fundamentally different relationships simultaneously. On one side, there are brands - D2C startups, Flipkart private labels, companies like FirstCry and Hercules Cycles—that need reliable manufacturing partners. On the other hand, there are factory owners - proud, opinionated, often people who have poured their life savings into injection moulding machines and assembly lines.

These, as you can imagine, aren’t the same conversations. They don’t use the same language. The founding team figured this out early.

The co-founders divided the world cleanly. Vivek owns the supply side: identifying manufacturing partners, negotiating terms, getting them to invest in new capacity, and commissioning new sites. If, say, a factory needs five additional machines, it’s Vivek’s job to make that happen.

Rahul owns demand, but more importantly, he architects it. From identifying emerging D2C brands to structuring long-term partnerships with large incumbents, he is responsible for turning Bidso’s manufacturing capacity into predictable, compounding revenue streams.

His role sits at the intersection of sales, category strategy, and customer insight, ensuring that what Bidso builds is always tightly aligned with what the market actually wants.

Aditya runs what is effectively Bidso’s product and design engine - translating market demand into manufacturable, scalable products. His team operates at the intersection of industrial design, engineering, and production, ensuring that every product is not just differentiated in form, but optimised for cost, tooling, and large-scale manufacturing.

In a business where small design decisions can significantly impact unit economics, his role is as much about precision as it is about creativity.

Nayan Goswami, the operations head, ensures the internal machine runs - production schedules, dispatch, quality, and the daily grind of making things work.

“There is a huge difference in the profiles we are dealing with,” Vivek explains. “My job is to build a pipeline of who can be the next factory partner for us. Rahul’s job is to identify who can be the next big brand. Aditya’s job is figuring out what product lines these factories should be manufacturing.”

This structural clarity, uncommon in early-stage startups, where founders often blur into each other’s roles, has allowed Bidso to scale without the coordination chaos that typically accompanies growth in operationally heavy businesses.

The FOCO Engine

Bidso didn’t arrive at its current business model on day one. It got there by going through two others first.

The earliest version of Bidso was a marketplace. Simply connecting brands with manufacturers, while taking a cut for the introduction. On paper, it was the classic asset-light platform play. In practice, it was a nightmare. There were no exclusive lock-ins on either side. Brands could go directly to the manufacturers Bidso had introduced them to. Manufacturers had no obligation to prioritise Bidso’s orders.

Also, the fundamental problems that plagued Indian manufacturing: inconsistent quality, missed delivery timelines, zero operational transparency, persisted because a marketplace, by definition, doesn’t control operations. It only just facilitates introductions.

The second iteration was exclusive supply capacity. Bidso locked in dedicated manufacturing capacity with partner factories through guaranteed volumes and predictable output. This was better. There was some assurance on total capacity, some predictability in what the factories could deliver.

But two problems refused to go away. Scaling was slow as each new factory relationship required negotiation, alignment on processes, and a whole new learning curve. The cost advantages were hard to capture because Bidso still didn’t control the factory floor. Product development was sluggish; spinning up a new model in a facility someone else operated meant working around their priorities and limitations.

The questions that eventually broke the model open were simple: how do we get a factory operationalised faster? How do we control quality, delivery, and customer promises directly? How do we develop new product lines rapidly within the same premises?

The answer was to control the operations without owning the factory. That was the birth of the FOCO model: Franchise Owned, Company Operated.

Here is how it works. A local manufacturing partner invests in the factory infrastructure: land, building, machinery, utilities. Bidso takes exclusive operational control of the facility under a long-term lease-cum-operations agreement. The partner earns a fixed annual return on their investment. Bidso bears no capital expenditure on the factory itself, converting what would traditionally be a heavy CapEx outlay into predictable operating costs.

What Bidso does invest in is the capability layer on top in terms of proprietary tooling and moulds, modular product designs, the entire NPD stack — new product development, rapid prototyping, licensed IP integration, and also centralised quality control protocols and compliance infrastructure (BIS, SEDEX, EN 71, ASTM) so that every unit is audit-ready from day one. They also do the category management intelligence that determines which products get built where, based on capacity availability, capability mapping, and demand signals from the brand portfolio.

Each FOCO unit is essentially a self-contained manufacturing node: partner-funded infrastructure, Bidso-led operations, shared quality and compliance standards, and a product engine that feeds new model launches into the facility on a quarterly cadence.

The interplay of machine throughput, assembly capacity, and storage infrastructure defines the total output. As utilisation climbs, Bidso layers in additional product lines to exhaust the unit’s capacity. Each new launch unlocks incremental monthly revenue while leveraging shared components and modular design to keep per-model investment low.

The model solves two problems simultaneously. For Bidso, it enables rapid, non-linear capacity expansion with full operational control and an asset-light balance sheet. For the factory partner, it offers a predictable return backed by Bidso’s demand pipeline, without the operational complexity of running a multi-brand, multi-SKU manufacturing operation. Neither side carries the full burden alone. Both have skin in the game.

The Numbers

Bidso’s growth curve tells the story of a business that found its footing and then accelerated hard.

The company works with 25 brand partners, with repeat orders accounting for ~85% of total business. The product portfolio spans 40 models across baby walkers, ride-ons, scooters, tricycles, and electric ride-ons, all focused on children’s outdoor mobility.

The Flipkart relationship is perhaps the clearest illustration of how the model compounds. In April 2024, Bidso supplied 4 SKUs across a single product line. By September 2024, that had grown to 20 SKUs across two product lines. By September 2025: 50 SKUs across four product lines. When a manufacturing partner consistently delivers quality, competitive pricing, and design flexibility, it is only natural that brands will want to expand.

As the relationship scaled, Rahul worked closely with the team to expand across categories and SKUs, turning what began as a transactional supply relationship into a deeply embedded manufacturing partnership.

The licensing partnerships tell a similar story. Multi-year deals with Hasbro (Transformers, G.I. Joe, My Little Pony), Hello Kitty, Peppa Pig, Harry Potter, and Chhota Bheem & Friends have opened a new dimension: taking globally loved IPs and manufacturing them in India for the Indian market and, eventually, for export.

Behind this growth is a tightly run demand engine led by Rahul, where repeat orders and deepening brand relationships form the backbone of revenue.

The China Problem (and Why It’s Really an Ecosystem Problem)

Ask Vivek about India’s manufacturing ecosystem, and you won’t get the standard ‘Make in India’ talking points. What you’ll get instead is a founder’s-eye view of what actually needs to be built, and how far there is to go.

On a recent trip to China, Vivek visited a factory and decided to check on bearings for one of Bidso’s products. Within thirty minutes, three different manufacturers walked into the office. Each one carried the exact same bearing, available in six different quality grades and price points. Ready to ship.

“This isn’t about individual companies,” Vivek says. “It’s about ecosystems.”

A toy, he explains, needs plastic moulding, electrical components, spare parts, packaging, and assembly. That’s four or five industries coming together for a single product. In China, each of these tiny components has an entire industry behind it, thousands of specialised suppliers, all competing, all innovating, all driving down cost and driving up quality simultaneously.

Vendors evolve to suit the market. They are already innovating on things like reducing packaging weight while increasing strength. The cost structures are broken down to operation-task-level atomic economics.

In India, the contrast is brutal. Take polymer compounds: you can buy commodity-grade polypropylene off the shelf. But the moment you need additives to increase tensile strength or flexibility, you enter a world of pain. Indian compounders are few, expensive, and opaque. A 5% additive can push the cost up 20-30%. In China, you describe your product to a compounder, and they hand you four ready-made samples to try.

“In India, you are scouting for months to find reliable vendors,” Vivek says. “And when you do, they rarely invest in R&D. You have to think through every innovation across the entire value chain yourself.”

Then there’s the talent problem. Skilled workmanship in India is vanishingly thin. Vivek describes struggling for months to find a single operator for a robotic welding machine. Four people who claimed to be experts turned out to be inadequate. “Even if I go to China and buy a hundred laser-cutting and welding machines, all robotic—I won’t find a hundred people in India in a year to staff them.”

The ecosystem that feeds into product manufacturing—spares, design, packaging, electrical parts—barely exists in India at scale. This, Vivek has come to believe, is the real problem to solve. Not just building Bidso’s own capacity, but helping build the second-order supply chain: the components, the vendors, and the ecosystem that makes manufacturing actually work.

“Building a world-class ecosystem that feeds into OEM/ODM growth,” he says, “is the real unlock for Make in India.”

The Aircraft Carrier Strategy

When Vivek describes Bidso’s long-term strategy, the metaphor he reaches for is an aircraft carrier.

The carrier itself is built on two capabilities: production excellence and design capability. Vivek has been instrumental in building the manufacturing backbone, while Aditya is assembling the design engine that determines what gets built on top of it.

“Anything that can be manufactured, we should be able to manufacture better than anybody else in India,” he says. “That’s production capability. And on design - every product we launch should be the best, most intuitive, best value-for-money offering in its segment.”

Today, the carrier is deployed in one theatre: children’s outdoor mobility. Tricycles, scooters, ride-ons, walkers. But the strategy is designed to be category-agnostic. Balance bikes and hoverboards are next. Small home appliances are on the horizon. When the time and resources are right, the same engine can attack any product category where Indian manufacturing can compete.

The design play runs deeper than most people realise. Under Aditya’s leadership, Bidso’s design approach operates on a 70-20-10 framework.

Seventy percent of design work is market-driven - responding to what customers are asking for, modifying existing products based on sales data. Twenty percent is form-factor innovation—same product, fundamentally reimagined in how it looks and feels. And ten percent is the frontier: open briefs sent to design agencies and independent designers, asking them to rethink how a child interacts with movement and play. No constraints, it could be a bucket you sit in and slide across the floor, as long as it develops motor skills.

“In toys, the difference between a product that sells and one that doesn’t is often invisible,” Aditya says. “A slightly better riding posture, a more intuitive steering angle, a smarter fold—these are small decisions, but they compound.”

That ten percent is where Bidso Labs comes in—an emerging experiment in crowdsourced product design. The idea is that external designers can upload product concepts to Bidso’s platform. If a concept generates enough interest, Bidso co-patents it, manufactures it, sells it to brands, and shares the revenue.

This push toward an IP-led pipeline is being driven by Aditya, who is building Bidso Labs as a structured way to tap into external creativity while retaining manufacturing leverage.

In a country where patents are rare, manufacturing bandwidth is scarce, and distribution is a mystery to most inventors, this could unlock a pipeline of product innovation that no single in-house team could match.

Peer Capital: The Patient Bet

When Bidso started raising capital in early 2023, the company was barely a month old. The term sheet from Peer Capital came in February 2023, one month after incorporation.

“It was largely the team that they bet on,” Vivek says. “They stuck by us because they believed in the team, not exactly the numbers.”

For a manufacturing startup, the choice of investor matters in ways that don’t apply to most software businesses. Manufacturing is slow, and it’s capital-intensive. It grinds. The feedback loops are measured in quarters, not sprint cycles. A VC that optimises for hockey-stick growth metrics will lose patience with a business where the primary activity involves negotiating lease terms with factory owners in Rajkot.

What Vivek found in Ankur Pahwa and the Peer Capital team was something rare: genuine patience married to operational engagement.

“For us, patience has been the most required trait, and Peer Capital has delivered on it,” said Vivek.

The trust was tested early. Through 2023 and into early 2024, Bidso was short of its projections. The team was, by Vivek’s own admission, “kind of lost at certain moments.” A different investor might have escalated, demanded course corrections, created pressure. Ankur’s approach was different: trust the team’s capability, ask the right questions, and stay out of the way until asked to step in.

The biggest test came when Bidso pivoted from a multi-category, tech-platform positioning to a focused, operations-heavy manufacturing play centred on toys. This wasn’t a minor tweak. It meant abandoning the narrative that had attracted the original investment—the platform play, the tech multiple, the scalable model.

The conversation with Ankur, when it came, wasn’t about valuations or multiples. It was operational. Why do you want to do this? Will it give you an edge? What will the financials look like? Have you considered alternative models?

“It was never about imposing a view,” Vivek recalls. “It was about making us introspect—why are we thinking this way? And after that conversation, we were very relaxed. We knew we had support.”

But the pivot also surfaced a more specific problem. The FOCO model—shifting from a marketplace approach to underwriting manufacturing capacity through franchise partners—was a drastic change in how Bidso would deploy its capital. This wasn’t something the founders had pitched when they had first met Peer Capital. They needed more than investor consent; they needed external validation that the model was sound, compliant, and would hold up to scrutiny.

Ankur stepped in with something most early-stage VCs can’t offer: a two-decade deep network in corporate India. Through his connections, he put Vivek in front of a veteran who had understood the books of companies operating asset-light manufacturing models at scale. The feedback was specific and actionable - the model structure was sound, the margin projections were reasonable, and the balance sheet treatment was compliant and reflected the commercial reality.

“If it was just Ankur telling me it was okay, I might have gone with confirmation bias—we are on the same team, right?” Vivek says. “But getting that external perspective at such a critical juncture, from someone who had seen how similar models work at scale, that changed everything. Today, all our growth is on that model.”

Ankur, for his part, points to something less obvious than operational support when he talks about why the relationship works. “The biggest thing I find attractive about Vivek as a founder is his intellectual honesty,” he says. “He is not faking it till he makes it. If he doesn’t know something, he says it. If something’s not working, he proactively has that conversation.”

He reaches for the old Narayana Murthy line, ‘bad news takes the elevator, good news takes the stairs’. “That is actually very true with Vivek. But more importantly, he doesn’t just come and throw a problem at you. He has thought through a few things he wants to bounce off you. He comes with ideas on what he should do, and he is very open to evolving his thinking. That requires a lot of honesty with yourself as a founder.”

The result, Ankur says, is that the trust equation simplifies. “I know he is saying what he means. And that raises my bar of acceptance for anything that can go wrong by 10x.”

That trust was pressure-tested most intensely during the fundraise. What followed was three months of relentless outreach. For a founder, the worst part of a fundraise isn’t rejection. It’s silence, the dead air between pitches when you have exhausted your pipeline and haven’t heard back from anyone.

That dead air never came. Ankur co-built the investor pipeline, made introductions, and after every pitch, successful or not, debriefed with Vivek. The feedback was candid: this is what you should build into the next deck, this is what you should discard, these are not the right investors for you. When the temptation arose to bend terms for a lukewarm investor, Ankur held the line.

“As a founder, it’s very tough to think that an investor is not right for you because they are offering money,” Vivek says. “There is a temptation to sign off. But having that backing from Ankur - ’don’t go for it, let’s wait it out, we will find the right people -and then actually connecting me to those right people, that was huge.”

By the end of it, they had closed with Blume Ventures. But Vivek indexes more on what the process did to the relationship than the outcome itself. “Post that, things have changed between how Ankur and I chat. Those three months really helped in how comfortable we are in this journey together and his belief in us and Bidso’s success shines through in every discussion. The relationship is nothing like a typical investor founder relationship I heard about before; it’s embedded in a deep personal bond of shared values and belief.”

“I can see very clearly when Ankur is wearing his Bidso hat versus his Peer Capital hat,” he adds. “And I have the comfort of knowing that even when it’s at Peer Capital’s disadvantage, he is giving me the right advice.”

(L to R); Aditya, Rahul and Vivek ; the band with the Peer Capital term sheet

What Comes Next

If you track Bidso’s evolution in fast-forward, the pattern is unmistakable: each phase has been about tightening control over a deeper layer of the value chain. In the earliest days, Vivek was buying products off the market and reselling them to brands—pure aggregation. Then the team started locking in supply relationships with existing manufacturers. Then they moved to exclusive capacity tie-ups. And now, with the FOCO model, they underwrite and operate the manufacturing capacity itself.

“We were not here two years back,” Vivek says. “We were not designing our own toys. We were not underwriting capacity. The evolution has been from going to multiple suppliers, to locking in supply, to now owning those supply capacities. And we will double down heavily on this.”

The next leap, and the one Vivek is most animated about, is becoming a design-led manufacturing company.

Much of this next phase hinges on the design engine Aditya is building. The idea is to actively shape demand through original products, modular platforms, and IP-led innovation.

The design engine is being assembled on multiple fronts. Bidso has already brought on dedicated product and toy design leads who are reshaping the existing portfolio—introducing modularity, improving customisation, and filing for patents and trademarks. The goal is an IP-led product pipeline where Bidso doesn’t just manufacture what brands ask for but creates products that brands can’t get anywhere else.

The insight underpinning the entire strategy is that Bidso’s manufacturing capabilities—injection moulding, metal fabrication, assembly—are not inherently tied to any single product category. The same factory floor that produces a toddler’s ride-on can, with different moulds and tooling, produce a small kitchen appliance or a household product. Today, all of this capability is pointed at outdoor mobility toys. Tomorrow, it opens new fronts: adjacent babycare products like prams, strollers, and high chairs. After that, small home appliances and household products in plastic and metal.

“The idea is to bring design-led manufacturing dominance into multiple consumer product industries,” Vivek says. “The capabilities we are building today will be replicable across these categories. That’s the playbook.”

The revenue targets reflect this ambition: ₹1,300-1,500 crores within five to seven years. With 20% of revenue targeted from exports within three years, Bidso is building toward a position where it isn’t just India’s largest outdoor toy manufacturer, but one of the country’s defining consumer goods manufacturing platforms.

Multiple estimates suggest 7-8% of global consumer goods manufacturing could shift to India in the next five to seven years, a move from $8 billion to over $40 billion. That opportunity won’t be captured by companies that can only manufacture. It will go to platforms that can design, engineer, license, comply, and deliver, all at once.

The ecosystem that doesn’t yet exist won’t build itself. Someone has to start. Given the momentum: the FOCO model validated, the design engine assembling, the capacity pipeline filling, and now Blume Ventures joining Peer Capital on the cap table, Bidso is no longer just drawing up blueprints; the build is underway.

What does it actually cost to build a company, and when do you start being able to see it?

I have been thinking quite a bit about what it ‘actually’ costs to build a company. No, not the money bit. That one we model in spreadsheets and walk our families through. I am talking about the ‘other’ costs, the kind that nobody ‘itemizes’, but certainly feels it or maybe even talks about it behind closed doors.

These are costs that arrive without a bill. You don’t notice you are being charged for it. You find out only when the account is empty. Founders are trained to price market risk, execution risk, and financing risk. They build models for all of it. What they systematically fail to price is the personal cost; because no VC term sheet has a line item for it, no board deck tracks it, and no post-mortem names it as the cause So most of them only see it in hindsight, and hindsight is the one vantage point from which you can’t do anything about it.

That’s the conversation I want to have here. What you actually trade to build a company, when you start being able to see it, and what it takes to stop paying more than you meant to.

The three sacrifices nobody names

There are three sacrifices that compound over a founder’s life, and only one of them shows up in a spreadsheet.

The monetary cost: This is not just the salary delta. It is the full opportunity stack: the promotions, the vesting schedules, the compounding returns on capital you never deployed because you were burning it instead. Most founders look at this like a one-line item.

Silent erosion: The slow bleed of time, health, friendships, relationships. Not being there when your kid had their first stage performance, pushing the visit to your parents by a week, then a month, then six. The friendships that ended because the invites stopped getting accepted. This is the cost nobody sends you a bill for, and the thing about not getting a bill is that you don’t notice you are being charged.

Depletion of self: This is the identity layer. Who you were before the company started. The version of you that had hobbies, a creative life, a particular kind of curiosity, a discipline about your own body and time, the chances you used to take. The version that existed before you decided you were a founder above all else.

If you ask me which one founders most underestimate going in, it’s silent erosion. We all model the monetary risk, but we rarely model the tsunami coming for our health and relationships. We know it exists, but knowing isn’t anticipating, and anticipating isn’t preparing.

If you ask me which one founders most regret coming out, it’s depletion of self. You can remake money, you can maybe try to repair relationships. But looking back at a version of yourself you don’t recognize, and realizing you don’t even remember who you were before, now, that’s a sacrifice with no return path. What people regret most isn’t that they didn’t make enough money, it is rather that they lost a self that’s never coming back.

Both of those last two are hard to act on because they are hard to see. So it’s worth being specific about what each one looks like from the inside. Let’s start with erosion first.

What silent erosion actually looks like

Most professional jobs have recovery cycles built into them: weekends, evenings, holidays. The keyword is recovery; you go in depleted, you come out restored. Founders don’t get that, and it’s not because they work harder than everyone else. It’s because the context switch never completes. The thread stays open. They are always clued in to work, even when the calendar says otherwise.

So the signs of silent erosion are signs that the recovery function has stopped working.

On health: Sleep becomes purely functional. You sleep because the body has to, and the restoration part has gone missing. Exercise drops out, you say that you will do it later, you will do it next week, and it never happens. You are getting sick more often, and when you get sick, it takes longer to come back.

On relationships: You are physically present and mentally somewhere else. Your spouse starts briefing you on your own life. ‘Remember, we have to go to that dinner’, ‘Remember, our child has a PTM tomorrow’. You used to know these things.

Friends keep inviting you, but you don’t go. Slowly, the invitations stop. Your parents, who might be in another city, you say you will go next Diwali, you will go after this fundraise, but doesn’t really happen.

On time: Leisure starts to feel like guilt. You can’t remember the last time you were bored or did something just because it was non-engaging. You wake up and check your phone before you check whether you slept well.

There is one metric that captures all of this: the recovery. One bad week, in a normal professional life, used to take a weekend to fix. Now it takes longer. The lag is lengthening, that’s the leading indicator. By the time it shows up as a performance problem inside the company, you are already six to twelve months into the degradation curve. That’s the alarm going off in a frequency you have trained yourself not to hear.

If erosion at least trips the occasional alarm, the third sacrifice is all the more ‘behind the scenes’.

What depletion of self actually looks like

Depletion of self announces itself in absences.

You can hear it in how founders reminisce. ‘I used to play guitar’, ‘I used to write’, ‘I used to go on solo trips’. The phrase has no return date attached. There’s no when-I-pick-it-back-up. Just ‘used to.’

You can see it in how the personality flattens. The conversation radius shrinks.

You talk about the company, the business, the industry, the round, the hire, or the competitor. Your curiosity about the world outside the company narrows. You are hyper-curious inside the company, and you stop looking outside.

You can see it in the irony that founders, who are professional builders, who spend their lives convincing teams and investors that they can build things, they lose the ability to build anything for themselves. The discipline that should be the easiest thing for a builder to maintain becomes the hardest. You can’t hold a workout regime, you can’t keep a sleep schedule, you can’t finish a book. The irony is that the ‘builder’ cannot build himself/herself.

Then there is the identity fusion. If someone asks how you are, you answer with how the company is doing. ‘We just closed a great quarter’, ‘We are hiring fast’. The question was about you, but the answer was about the company. You didn’t notice. Sadly, nobody else did either.

The most dangerous version of all this is that, from the outside, it looks like drive, it looks like hustle. From the inside, a lot of it is mental numbness wearing the costume of ambition. This matters operationally: a depleted founder making decisions at high velocity looks identical to a healthy founder doing the same, until the quality of those decisions starts to diverge. By then, the organization has already been shaped by months of low-quality inputs.

Knowing all this, you would expect founders to ration it. Most don’t, and the reason is a single bargain they keep making with themselves.

The ‘just one more milestone’ bargain

Every founder I have worked with has, at some point, made the bargain. ‘Just this one more milestone, and then I will take my foot off’. Then they hit the milestone, and a new milestone is already standing in the doorway.

When I am sitting across from a founder making that bargain again, my job is to name the pattern out loud. ‘You said this before the Series A’. ‘You said it before the Series B’, you are making the same decision with different vocabulary. The loop is invisible to the person inside it because the milestone label changes every time. Pre-seed becomes Series A, becomes scale, becomes profitability. The justification refreshes, and the ‘cost’ compounds.

The harder ask is to get the founder to write down what would have to be true for them to feel they have earned the rest. Write it down, date it, and then look at it twelve months later. What you almost always find is that the criteria were met, and the rest wasn’t taken.

The question underneath all the milestones is whether you re deferring life, or whether you are afraid of what you will find when the company is no longer the excuse.

Here is what happens to the bargain over time. Each deferral lowers the threshold for the next one. Your reference point for what counts as sacrifice shifts with each repetition. The first time you skip a friend’s wedding for a board meeting, it’s a hard call. By the tenth time, it’s just the calendar. This is the normalization of sacrifice, and it’s the most insidious part; because you lose the ability to feel the cost even as the cost keeps compounding. Sacrifice stops being a decision and becomes the operating system.

When the personal leaks into the company

Up to here, this reads like a private matter: the founder’s health, relationships, or sense of self. Sadly, it doesn’t remain ‘private’, it soon leaks. The question is whether you are watching for the leak before it hits the P&L.

Four signals I watch for as an investor, well before any of these show up in the numbers. These are leading indicators. They show up in the operating rhythm of the company months before they hit revenue, retention, or team stability.

Meeting behavior changes: Founders stop facilitating and start dominating. The room used to be a place where curiosity surfaced. Now it’s a place where decisions get announced. This pattern has a specific signature; meeting prep declines, but meeting control increases. The founder is less prepared and more directive. What looks like decisiveness is often the collapse of deliberation.

The team reads the room, and soon, strong performers go quiet. People who used to lead conversations, push back, and ask the hard questions figure out that the founder is in a different gear now, and their opinions are not going to land the way they used to. So they stop offering them, and the silence in those rooms is one of the loudest signals you will ever get.

Latency increases: Now the same decisions take days. Emails sit, Slack threads stretch across a week. The company hasn’t gotten more deliberate; it has gotten more avoidant. Decision latency is one of the most measurable operational indicators I track. When it doubles without a corresponding improvement in decision quality, that’s depletion, not process maturity.

Investor updates get vague: They get over-polished, and they round up the positives and round down the rough patches. The overall picture migrates from what is actually happening to ‘here is the best version of what is happening’. I have never seen a healthy, clear-headed founder become less transparent. Opacity tracks depletion.

The single thread running through all four: the founder’s inner circle has shrunk. They are ‘talking’ to a lot of people but confiding in very few. Fewer inputs, fewer disagreements, before decisions get made. That’s when things start to go wrong in ways that don’t look like a single bad call.

The distinction matters, both for investors trying to read a situation and for founders trying to read themselves. Not every bad call is a depletion signal. The difference is worth getting precise about.

Being wrong is a context problem. You read the market wrong, you hired wrong, you priced wrong, or maybe you picked the wrong segment. The information you had was either incomplete or you read it incorrectly.

Depletion is a process problem. The information was available, the right people were in the room, but the decision still turned out badly. The pattern isn’t in the what of the decision; it’s in the how. No prior deliberation, no challenge. No one in the room felt safe enough to push back.

The signs that say depletion rather than wrong: You make a decision and reverse it three days later when slightly new data shows up. That’s often seen as adaptability, but that’s usually low conviction made on no deliberation.

Your emotional response to a small operational failure is wildly disproportionate. A minor miss gets a major reaction. You start taking short-term fixes that you know are creating long-term problems. You cut people to fix burn; you discount to chase revenue. You are not solving, but rather sedating.

The framework I use: is this person making bad calls, or are they making calls badly? The first is content, the second is process. Process is depletion.

Where the ‘heroism’ pressure actually comes from

The performing, the not asking for help, the inability to be straight at home, I believe that none of it is a personality defect. It is manufactured by the incentive structures that surround every founder from day one. It’s worth naming exactly where it comes from, because you can’t push back against a force you haven’t located.

I think about it as four sources.

Pitch rooms: Investors, by selection, are looking for confidence. They are looking for people who present strongly, hold conviction, and can defend a thesis under pressure. Vulnerability, in that room, is systematically selected against. So founders learn very early that one mode is what gets funded. A behavior that gets rewarded in pitch rooms gets generalized to every room.

Boardrooms: Boards have a structural information asymmetry problem. The founder controls what gets surfaced. The board sees what the founder chooses to show them. Most boards, even well-functioning ones, are operating on a curated version of reality. Founders who have internalized the heroism script present the curated version not out of dishonesty, but because that’s the mode they are in. The cost is that the board can’t help with what it can’t see.

Founder Twitter: Nobody posts about the failure. Everyone calibrates toward the win. Your reference set is artificially skewed, and your private struggle starts to feel like a personal defect rather than a structural part of the journey.

The Press: How many profiles get written of founders who are struggling? How many cover stories feature the company that didn’t make it? The mythology is constructed in retrospect. Founders internalise the retrospective mythology going forward. The press is reporting from a destination. The founder is still on the road. Until they get there, every honest moment feels like a betrayal of the script.

What does this cost the founder, you ask? It costs them the ability to ask for help.

Asking for help reads as a weakness in this culture, which is precisely backwards from what it actually is. In every high-performing system I have seen: great companies, great athletes, great investors, the ability to accurately assess one’s own limitations and recruit help for them is a core competency, and certainly not a vulnerability. The mythology has it exactly backwards.

It costs them honesty, with investors, team, partner, and friends. The performance becomes the default mode, and at some point, you can no longer easily switch out of it, even with the people who knew you before the performance started.

It costs them the ability to separate the company’s struggle from their own. When the company is having a hard quarter, they personally feel like a failure. The strategy failed, the market shifted, the competitor moved, none of that; they personally failed. That fusion is what breaks people, more than the actual struggle of the business does.

The bigger critique here is of the culture itself. Hustle culture, as the industry has built it over the last decade, has made founders less resilient. Or at the very least, it’s convinced founders that resilience means not talking about the breaking.

From hero to architect

If the culture won’t hand a founder an exit from all this, the founder has to build one. I think of that shift as moving from hero to architect, and the first thing to say about it is that there’s no automatic stage where it happens.

I want to be clear about that, because the comforting answer would be that you get to Series B and the load lightens, or you hit some ARR threshold and the equation changes. It doesn’t work like that. Every stage produces a fresh justification for not stopping.

Pre-seed: You have to do everything because there is no one else.

Series A: the team is new, and you can’t yet trust them.

Series B: The scale is breaking everything.

Series C: The stakes are too high to delegate.

The justifications are always available. They are always plausible, and they are almost always wrong.

What actually has to be true for the load to lighten, and this is specific:

You have built a layer of leadership below you that you can trust to make the calls you would make, not similar calls, not adjacent calls, the actual calls. This requires that you have hired for judgment, not just execution, and that you have given those people enough runway to develop it. And your definition of success has expanded to include your own sustainability, not as a nice-to-have, not as a soft people thing, but as a hard operational input, the same way runway is an input.

You have come to believe, all the way down, that your depletion was a bug and not a feature. Because the culture taught you it was a feature, but remember, the culture is wrong.

When all three are true, the hero model is a choice and not a necessity. Which is also when it gets most dangerous, because plenty of founders who hit that point still refuse to switch modes. They want to keep being the hero. They have built a company that no longer needs them to be the hero, and they cannot let themselves become anything else. That’s the curse, and that’s the moment that breaks otherwise excellent companies.

The reframe I’d offer: stop asking ‘when can I stop?’ Start asking, ‘what am I protecting by not stopping?’ The honest answer, most of the time, is self-importance. The belief that without you, the thing falls apart. Sometimes that belief is true, and sometimes it’s the last thing standing between you and irrelevance, and you are afraid to find out which.

Choice or collapse

The founders who do make that shift, how do they get there? It is rarely the way they later say they did.

If you ask me whether the move from hero to architect comes by choice or by collapse, the honest answer is mostly collapse. Founders then narrate it as a choice afterward, because narrating it as a choice keeps the hero arc intact. That narrative is almost always reverse-engineered. The clean version of the story gets constructed after the breakdown, because the breakdown itself doesn’t fit the hero arc.

Real clean transitions are rare, and they almost always have a trigger somewhere. A health scare, a relationship ultimatum, or a therapist who said the thing nobody else was willing to say, or maybe an investor who delivered a hard message that landed.

Voluntary transitions in the pure sense — a founder waking up one morning and choosing differently with no external prompt — those are very rare. Believe me, most are forced.

Both can work, but forced transitions carry more wreckage in every dimension. Teams get unstable because the transition is reactive, not planned. Relationships are already damaged by the time it starts; the partner has been absorbing anxiety for years, the kids have adjusted to absence. The health costs have already been paid and continues compounding. A voluntary transition gives you the chance to make the shift while the collateral damage is still containable.

The founders who make it as a real choice usually have one thing in common. Someone in their life: a partner, a friend, an advisor, a mentor, who saw the pattern early and named it before the crisis. That’s why these conversations matter. The pattern needs to be named out loud, by someone the founder trusts, before the body, the relationship, or the company forces it.

What I had to let go of

I have been through a version of this myself. So rather than just recommend the switch, let me be specific about what it costs to make it.

The things I had to actively let go of were not what I expected.

The identity of being the hardest worker in the room. The one with the most to lose, the one who was supposed to know everything, the control that came with being the bottleneck — which is a strange thing to call control, but it is one. As long as you are the bottleneck, nothing moves without you, which means nothing exists without you, which means you exist as long as it exists. That’s addictive feedback loop, and letting it go felt like letting go of a self.

The hardest specific habit to release was wanting to be the smartest person in the room on every topic. Delegation isn’t really about time management. It’s an identity problem. You are handing over not just the task but the right to have the best answer on it. And for founders who have built their self-concept around knowing the most, that transfer is genuinely painful. Founders who think they are bad at delegation are almost always fine at the mechanics of it. What they are bad at is the loss of primacy that comes with it.

The thing that surprised me, and surprises most founders who go through this, is how much of the sacrifice was voluntary, and self-imposed. A function of your own assessment of what your identity required, not of what the company actually needed. You were choosing martyrdom because it gave you status and meaning. The company wasn’t asking for all of it; you were.

The other surprise, and this one is harder to admit: the company doesn’t fall apart when you step back. If you have built well, it works. In fact, in most cases I have seen, the company’s decision quality improves when the depleted hero steps back and the architect steps in. The most surprising thing, for a founder who has spent years being the load-bearing wall, is realizing the building was actually being held up by everything around them all along. The wall was load-bearing by choice, not by necessity.

The sacrifice report card

If you wanted to check where you actually stand on all of this, not vaguely, but on paper, here is the exercise I would put in front of you.

Who did I say I was doing this for, and are they actually better off than when I started? This is the beneficiary question. Whoever you named at the beginning, your family, your team, your customers, yourself, go check on them. Look at them now, not the company they are connected to. Them. Have they actually benefited from the version of you that this company produced? The honest answer is the report card’s first grade.

What did I stop doing that I used to do for myself? The identity-depletion question. Make the list, say, the number books you used to finish, the friends you used to call, the trip you used to take alone every year, the instrument you used to play, the fitness discipline you had, the hobby, whatever was on the list before. Look at the gap between then and now. That gap is the second grade.

Who has quietly adjusted their life around my absence, and have I acknowledged it? The silent-erosion question: your partner, your parents, your kids, your closest friends, people you didn’t even notice were adjusting. Have you named what they have absorbed? Have you said thank you in a way they could feel? That’s the third grade, and most founders fail it without knowing.

If everything goes exactly to plan from here, will I like who I am when I get there? The forward question. The first three are about looking back, and the problem with looking back is you can’t change what you find. This one looks forward. If the next two years go perfectly, say, if the round closes, the launch lands, the growth compounds, etc, will the version of you who arrives at that destination be someone you would want to be? If the answer is no, you have a bigger problem than any of the milestones can fix.

That’s the report card. Most founders won’t sit with it long enough to fill it out honestly. Partly because the answers are uncomfortable, and partly because filling it out makes the costs real in a way that knowing them doesn’t. The ones who do, who write down the answers, look at what they wrote, and let themselves feel it; are the ones who give themselves a real shot at being the architect rather than the hero.

The rest just keep paying the bill nobody sends them.

Inside Fabriclore, an end-to-end demand generation and fulfilment platform for fabrics.

Yeh jo des hai tera swades hai tera

tujhe hai pukaara

ye wo bandhan hai jo kabhi toot nahi sakta

mitti ki hai jo khushboo tu kaise bhulaayega

tu chaahe kahin jaaye tu laut ke aayega

nayi nayi raahon mein dabi dabi aahon mein

khoye khoye dil se tere koi ye kahega

In a theatre in Abu Dhabi in late 2004, the then 31-year-old Vijay Sharma sat with tears rolling down his face as A.R. Rahman crooned these Javed Akhtar-penned lines. On screen, Bollywood superstar Shah Rukh Khan played a NASA scientist who had returned to India on a short visit, gone to a village in Uttar Pradesh, and was realizing, slowly, that he would not be leaving again. The film was Swades.

Vijay had gone in for a Bollywood film. He came out with what he would later call a “defining moment”. Slightly cliché (yes), but it was a feeling that was impossible to shake for Vijay.

He was several years into an 11-year stretch with one of the largest oil and gas EPC companies in the Middle East. The job paid well, and the work was structured. The path forward was visible enough to coast on. None of that was in his mind before he sat down in the theatre.

It would take him another four years to actually make the move. Then another six to start Fabriclore. Then seven more to walk away from a version of Fabriclore that was working, to build the version that wasn’t yet working but should be.

The decisions that came later were made the same way as the first one was: deliberately, and against the apparent rationality of staying put.

A marketplace for women, a textile shop in Sikar House, and a cafe in Delhi

What he came back to build first was, however, not Fabriclore.

Vijay returned to India in August 2008 and, within two years, had registered an engineering consultancy in Gurgaon. By the time he shut it down in 2021, it employed 100 people and clocked $3.5–4 million in annual revenue, 90% of it international. Water technology company Aquatech in Pittsburgh was the anchor client. German and Singapore firms added the rest.

Anupam Arya joined first, in 2011. Vijay had met him at Luminous Engineering and Technology Services in 2008, where Anupam, fresh out of Amity Business School, was already restless. Hired into international sales, he was moved within two months to patent research for Luminous’s new ventures in batteries and electric two-wheelers. Sandeep Sharma, Vijay’s first cousin who had grown up in his Delhi parental home, joined in 2012.

Sandeep then moved to Jaipur in 2014 to work with his father’s textile business, which operated out of a shop in Sikar House, one of Jaipur’s busiest fabric areas.

While running point at the shop, he had also started a women-targeted marketplace called naaari.com (yes, with three a’s). It sold everything from garments and books to crafts and fabric. He uploaded his father’s fabric inventory to the site. Within a few months, the data told him something he wasn’t expecting: nine out of every ten orders were fabric. The crafts or kurtas barely moved. People came to naaari.com for fabric because nobody else was selling it online.

In 2015, on one of his monthly Delhi trips, Sandeep took the data to Vijay and Anupam. The three of them sat at a cafe near the Sector 12 metro station in Dwarka, Delhi, and worked through it. Anupam suggested the name. Sandeep bought the domain, and by April 2016, they had 120 orders.

The first few months of Fabriclore were operationally tiny. Sandeep was still running the shop in Sikar House. He and his wife did everything: photographed the fabrics, took the orders, printed the labels, cut the meterage, packaged it, dropped it with the courier, and answered customer support calls when packages didn’t reach on time. Five to 10 orders a day, 40 to 60 cuts of fabric, every day, by hand, on the floor of a Jaipur shop.

(L to R: Anupam, Vijay, Sandeep)

What seven years of D2C taught them

Between FY17 and FY20, Fabriclore went from INR 1 crore to INR 7 crore. The team grew. Anupam moved to Jaipur in 2019 to be closer to operations. 8,000 orders shipping out every month, all through their own website.

Sandeep travelled far and deep across India’s textile craft clusters in those years. Name a craft, Fabriclore had a collection ready—Indigo, Dabu, Bagh, Ajrak, Ikat, Kalamkari, Bandhani, etc. By the time the catalogue was settled, Fabriclore had built a supply chain spanning over 30+ production techniques across 14 states. They had brought India’s craft clusters online, one at a time, and each cluster taught them something about producing in small quantities, with custom design, on lead times the formal mills wouldn’t entertain.

Three of the biggest names in the global fiber business: Liva (Birla’s viscose), Lenzing (the Austrian maker of Tencel and Modal), and Bemberg (Japanese premium cupro) had come on board as marketing partners. These are fiber houses that supply the mills upstream. Their willingness to attach their names to Fabriclore’s catalogue was, more than anything, a credibility signal in a category where authenticity is hard to verify.

In June 2020, after four months of due diligence that began the day before the COVID lockdown, the Rajasthan Venture Capital Fund put INR 4 crore in. Vijay shifted to Jaipur in August 2021 and shut down the engineering consultancy. He was a full-time co-founder, finally.

Then the numbers stopped flattering them.

From FY21 to FY23, revenue stagnated between INR 10-11 crore. The team had grown, the marketing spend had grown, and the order volume had grown. The economics hadn’t, however, moved.

At one point, the Facebook ad spend was 45 lakh a month against monthly revenue of 1.5 crore. Almost a third of every rupee earned went straight to Meta. The rest was getting shared between the courier, the payment gateway, the photographers, and the warehouse. Vijay started saying, half in frustration, half in clarity, that they weren’t running a fabric business; they were running an unpaid subsidiary of the digital infrastructure economy.

What they were learning, though, was bigger than the numbers showed.

Six years of running a D2C fabric brand had taught them what almost no founder of a B2B sourcing platform could learn from a deck. They knew which mills in Surat made viscose worth carrying and which ones cut corners. They had learned to search in Coimbatore and Erode for cotton spun, Banaras for silk, Punjab for knits, Jaipur and Jodhpur for processing. They knew which fiber brand reps actually returned calls and which didn’t. They knew what it cost, in days and rupees, to source 500 meters of a custom blend versus 5,000. And, finally, they knew what D2C founders complained about over WhatsApp at 2 AM, because they had been those D2C founders for six years.

The fragmentation was the thing that had been eating their margins, and they could see it from the inside.

The pivot of a third kind

In the famous sitcom Friends, Ross Geller’s most quoted line is shouted from a staircase in season five, mid-couch-move.

Pivot. Pivot. PIVOT, he screams.

The joke works because the couch isn’t actually pivoting. It is clearly stuck, ‘forcing’ Ross to scream.

Pivots happen, in most cases, because the founders’ hand is ‘forced’. The business has hit a hard ceiling, so to say. The unit economics breaks, the runway shortens, or the market moves, something pushes the change of direction.

Fabriclore’s pivot in 2023 had no ‘very apparent’ trigger, none of the obvious things that usually ‘force’ a hand. The D2C fabric business revenue had stagnated at a decent INR 10-11 crore for three years, but the business wasn’t bleeding. It generated good cash, it paid salaries, and it could have sustained doing exactly that for as long as needed.

In short, it was working decently well by every measure that wasn’t growth, and was the kind of operation most founders spend years trying to build and never reach.

Vijay, Anupam, and Sandeep walked away from it anyway.

The B2B realisation wasn’t exactly an ‘arrived-as-a-flash’ moment; it was more through conversations the three founders kept having through late 2022 and early 2023. Brand after brand told them the same four things: they couldn’t find suppliers, they couldn’t get low Minimum Order Quantity (MOQs), they couldn’t get sampling and bulk from the same place, and they couldn’t trust what was on the invoice. Fabriclore had spent years solving these exact problems for itself. The question was whether to sell that capability instead of competing with it.

The conversations between the three cofounders before they decided were as expected: Should they keep the D2C and add B2B? Should they pivot to garmenting, the obvious adjacency? Should they shut the consumer brand entirely and bet on B2B as a single-line business? Discussions, more discussions, and obviously differences.

The decision in mid-2023 was the hardest of the available options. They finally decided to discard the consumer business while it was still going decently well, and become the supplier to the next thousand consumer brands instead of being one of them.

Vijay calls it a “desi bitter kadha” (an ayurvedic home-made decoction used to treat colds, coughs, and fever) they had to drink. They drank it.

What followed was definitely harder, painstakingly slower, and far less certain. The first six to nine months were brutal.

Retail margins were traded for trade-grade ones, workforce had to be downsized from 94 to 28 within six months to stay afloat, inventory had to be liquidated, suppliers needed to be paid for stock the company hadn’t yet earned from, and salaries went unpaid for seven months.

Vijay took a personal loan against his house and put it into the company. Even more, he, Anupam and Sandeep stayed without drawing any salary for a year. The core team stayed intact because they truly believed the next version of the company would be worth all the trouble.

Revenue dropped to INR 5.5 crore in FY24 from INR 11 crore the year before, with a monthly run rate of around 50 lakh. Marketing spend went from 45 lakh a month to zero, overnight.

Why has B2B Fashion Supply Chain market stayed open despite the funding

The bet they had made on B2B sourcing wasn’t a category nobody had tried. India’s B2B space hasn’t been short of capital. Industrial supplies, packaging, raw materials, agri inputs, etc, have received hundreds of millions of dollars across the last decade. So, why is fabric sourcing still operationally feudal?

The answer is rather straightforward: organising the financial layer compounds faster than organising the production layer, and, naturally, capital tends to follow what compounds.

Bill-discounting (financing a seller’s unpaid invoices) and supply chain credit are powerful wedges. They scale through working capital cycles, the unit economics work quickly, and a platform can move from one category to another without necessarily rebuilding its core. Several B2B players over the years have tried to go deeper in their respective categories: taking inventory positions, building procurement workflows, digitising supplier networks, et al. Some of that work has been substantive, but financing layers got built faster, scaled faster, and absorbed the headline capital because that’s where the venture model has historically had the most patience.

Reorganising the production layer is a different beast. Fabric sourcing is a physical, technical, multi-stage workflow with thousands of mills, hundreds of processing units, dozens of techniques, and a highly subjective notion of what counts as ‘the right fabric.’

No amount of credit changes that workflow. It just makes paying for it slightly easier. The textile-specific problem: high fragmentation, low MOQs, custom processing, variable quality, fashion-driven SKU churn, doesn’t yield to a financial wedge. It needs someone to actually take on the discovery, the sampling, the BOM, the QC, the warehousing, the studio, and the salesperson on the floor of a Surat or Tirupur mill. Then it needs someone to turn all of that into productized, repeatable SKUs that a D2C brand can reorder 3-6 months later.

The current generation of emerging players in textiles is closer to the production layer but still operates one layer above it. One model aggregates a verified supplier network and standardises procurement workflows for speed and transparency. The moat is demand aggregation. The dependency is on existing supply structures, which means margins compress as volume grows.

Another model works with the deadstock and surplus inventory layer to enable low MOQs and quick fulfilment. Liquidity is its only real moat. Consistency and repeat supply aren’t part of the design. A third digitises traditional wholesalers and garment manufacturers into an online discovery layer, focused mainly on running garment categories. This is the typical catalogue visibility, but not supply creation.

What sits below all three of them, and what Fabriclore is building, is the actual production layer.

The framing Vijay uses for it is “demand-driven”, and the distinction matters. A discovery platform shows you what already exists, while a demand-driven platform decides what gets made. Fabriclore moves control upstream, into greige sourcing, custom weaving, dyeing, printing, finishing, and absorbs the QC, testing, and delivery on the other end.

The result is that an inherently variable supply chain becomes a set of predictable, reorderable SKUs. A buyer who liked a 60% linen/30% cotton/10% viscose blend at 220 GSM in dusty rose colour can come back six months later and order it again, at the same specification, without the original mill having to remember anything about them. This very repeatability is what compounds into a moat: supplier dependency, process ownership, and reliability that an aggregator cannot replicate by adding more suppliers to its network.

International parallels exist. China’s Baibu, a B2B textile platform running a self-operated supply chain, has raised ~465mn across multiple rounds from investors including Tiger Global and Source Code Capital. South Korea’s Swatchon did a smaller, designer-grade version of the same thing. Both companies operate at the layer Fabriclore is building toward.

That layer, in India, has stayed open because, frankly, it’s really hard to solve. But once solved, it brings a paradigm shift in the Indian textile sourcing structure, not just for India but for exports, more importantly. Seven years of building a D2C brand, working with thousands of artisans across India, getting small collections of fabric created and then marketing and selling them digitally across the globe, somehow gave the trio the exact resume to solve this problem.

Shark Tank, a 12-hour wait, and 800 leads

The Sony Pictures call came in mid-2023, asking them to participate in Shark Tank India Season 3. Anupam thought it was a PR scam. Vijay, when the call eventually came directly to him, was skeptical for a different reason: the show seemed built for D2C founders with food brands and skincare lines, not B2B textile companies. They said no.

Their team, internally, mutinied. The young people in the office wanted them to do it. Sony’s producers kept calling back. Eventually, the founders relented.

The shoot at Sony Studios in Mumbai’s Film City on 9 October 2023 was, in Vijay’s recall, “a maaaaad house.” A real, working television studio with three days of rehearsals across three levels of assistants before the shoot itself. Day five began with makeup at 9 AM. They were told the shoot would start at 3 PM. It started at 9 PM. Twelve hours of waiting. By the time they walked into the tank, the burden of expectation had been wrung out of them.

What they remember most clearly is Lenskart’s Peyush Bansal saying “mere hisaab se yeh ek shaandaar business ban sakta hai” (I think this can become a brilliant business) and Shaadi.com’s Anupam Mittal cutting in with “dum toh hai” (there is merit). The deal itself didn’t happen. They were already in serious conversations with Peer Capital and couldn’t offer the kind of terms a Shark Tank deal usually requires. The validation, on national television, from two operators who had built real businesses, was the part that mattered.

Off-stage, one of the producers told them theirs was the most “bare bones, truthful pitch” the show had recorded that season. The episode aired in February 2024, and 800 leads landed in their inbox in a week (link to the episode below).

That’s when they realised the pivot was actually working.

(L to R: Vijay, Anupam, Sandeep)

A Starbucks at Phoenix Mall

Ankur Pahwa, Managing Partner at Peer Capital, first met Vijay at a Starbucks in Phoenix Mall, Whitefield, Bangalore, in mid-2023. They had spoken on a Zoom call a few months earlier and had been circling each other for a while, with Peer Capital trying to build a point of view on the opportunity and Vijay persisting on his end.

Ankur’s first impression of Vijay is worth quoting on its own terms, because it explains why the round took as long as it did and why it was eventually the right one.

Vijay, in Ankur’s words, “doesn’t fit the conventional founder stereotype”. He doesn’t dominate the room; he doesn’t oversell. He, in fact, listens carefully, chooses words deliberately, and tends to favour substance over volume. In a world where volume means conviction, charisma means capability, this is the kind of founder who can easily be underestimated.

Ankur uses two phrases that capture his eventual read: “silently formidable, and antifragile operator”. Someone whose modesty masks competence, whose grit gets stronger under adversity, who doesn’t impress in the first 20 minutes of a meeting but separates massively across 20 months.

What Ankur saw in Anupam was different in shape. Younger energy, curiosity and thirst to keep learning that doesn’t seem to settle. Anupam is the founder who, given any operational problem, will relentlessly look for a “systemic productised” solution. Anupam also brings modern design thinking and a systemic, UX-driven approach to every unit of the business, essential to turning any POC into a scalable model.

Sandeep is a different grain. Heads down, operationally obsessed, the founder who knows the manufacturers and the processors at the level of the people inside them. Customer pain, vendor relationships, the unglamorous middle of the supply chain, that is where Sandeep lives. Ankur calls him a “son-of-the-soil” founder, a Bharat founder, if you will. Every conversation with Sandeep eventually circles back to a customer problem he is in the middle of solving.

Three different operators, and one unusually complementary team.

The colour of money

The first call between Vijay and Peer Capital had happened in July 2023. By the time Ankur flew to Jaipur in October 2023 to spend a full day with the team, the conversations had moved past the deck. The Investment Committee discussion happened late October, and the term sheet came on Dhanteras (fate does play its magical role, eh?).

What was meant to be an INR 6 crore raise turned into an INR 12 to 14 crore plan once the business model was rebuilt with proper unit economics. Then came due diligence and co-investor onboarding, which Vijay describes as the toughest part of the entire fundraising journey.

Mid January 2024, on a long drive from Delhi to Jaipur, Ankur and Vijay had the conversation that shaped the round. By then, it had become clear that Fabriclore needed a larger round than originally planned, and the co-investor piece was harder to put together than either of them had expected. Some of the terms in play weren’t the kind that build a business well over the long run.

Ankur’s posture in those conversations stayed consistent. Peer was comfortable with terms that were fair for everyone in the room: the founders, the company, the existing investors, the incoming investors, and Peer itself. Anything short of that wasn’t a deal Peer wanted to lead, and anything beyond that wasn’t a deal Peer wanted to block.

He said as much to Vijay directly on that drive. “Look, Vijay, if you feel that you are getting money from somewhere else, and are comfortable with the terms, you can have it. We can back away and enable you to go ahead without us, if it helps you to continue building out your visions.” It wasn’t an ultimatum. It was how Peer thought about the round: the right terms for everyone who mattered. The founders had to feel comfortable with the offer and the partners they chose to take on as part of their journey.

Vijay stayed.

The reason, in his own framing, was something he had started calling “the colour of money”. “This game is not a T20 game,” he says. “It’s a test match game. Every session will have its own challenges and opportunities.” He needed a ‘peer’ who would be useful in the seventh hour of the third day, not just on the day the cheque cleared.

The round closed in May 2024, almost a year after the first call.

The relationship since then has not been about board governance. It has been about coaching. “The good thing about Ankur is that he’s a coach,” Vijay says. “A coach pushes you, but a coach also protects you.” The instinct to chase revenue at any margin, which is easy in textile, where the industry is large enough that volume is always available, was the first thing Ankur pushed back on. “Vijay, if you want to do business on a margin of 2 to 4 per cent, it doesn’t matter,” he had told him at one point. “We will not create value. We will create an NBFC then.”

The conversations between the two of them, two to three times a week now, are about the things that don’t usually show up on a deck: whether revenue is being earned at a quality that compounds, whether the team is being built deep enough to scale, and whether the right things are being measured at the right cadence. The kind of detail that requires Ankur, more than once, to fly to Jaipur and sit with the team for entire days, going from the bottom of the funnel up instead of the other way around.

It was one of these visits that prompted Sandeep to move to Delhi to open the market there. It was another that brought a new team member into the Delhi office. Ankur’s effect on the company is rarely a single decision. It is an accumulating push toward boldness in a founder team that, after seven months of unpaid salaries and a mortgaged house, has every reason to be cautious.

Vijay calls these the “Tom and Jerry” moments. The kind of working relationship where the disagreements are useful precisely because both sides care about the same long-term outcome.

In Vijay’s own framing, the dynamic looks like this: “It wasn’t just reviewing outcomes. It was constantly questioning, discussing, and challenging. It wasn’t always comfortable. Founders, by nature, often operate with strong belief, sometimes even overconfidence, that they are right. That mindset was no different here. There were many moments of agreement, and just as many disagreements.

But those questions, often persistent and deeply probing, began to trigger a different kind of thinking. They pushed conversations beyond surface-level answers, forcing a deeper dive into the why and how. Over time, this led to a more structured, methodical approach to solving problems, rather than relying only on instinct.

Vijay goes on to add “Sometimes, after long hours of work, I’d start to feel like there was always something more being questioned by Ankur. No matter how much was done, another question would come up. But over time, that feeling shifted. What initially felt like constant questioning was actually a push toward higher standards and sharper thinking. Each question wasn’t about finding faults. It was about finding gaps, and those gaps, once addressed, were what moved things forward.”

What Fabriclore actually is now

The business that emerged from the pivot is not a marketplace. It is an important distinction to make.

Fabriclore is a ‘1P platform of its own kind’. One that is reliable, consistent, and without being bloated with assets and inventory. A single point of contact for the buyer, with the entire underlying complexity of mills, processing units, traders, logistics, and quality control absorbed into the company itself. A D2C founder placing an order doesn’t see who wove the fabric, where it was processed, how the BOM was generated, or which of the 300+ weaving mills and processing units the order was distributed across. They see one quote and one accountable counterparty.

What makes that possible is the platform Anupam has spent the last two years building.

The ERP is built on a bespoke open-source framework, customised heavily, with a proprietary business intelligence layer on top. Every meter of fabric is tracked, every customer order automatically generates the corresponding purchase orders to the right supplier, the right BOM accounting for shrinkage and wastage at that specific volume, and the right production schedule based on the printing technique being used.

Default supplier allocation is automated, and so is stock replenishment. Pricing is governed by a configurator that lets the salesperson close on the spot without an internal approval cycle, because the margin and payment-term guardrails are baked into the system. Anupam calls this an “approval-free supply chain”, and it is being engineered to run that way internally and, over time, externally.

The point of all of this is compression. Legacy fabric sourcing: discover new fabrics, new suppliers, buy sampling, generate inquiries, source bulk, RFQ, PO, runs to roughly 50 days from first conversation to first delivery, much of it spent in the back-and-forth between buyer and a fragmented set of mills, processors and traders. Fabriclore’s productized stack does the same workflow in 0 to 7 days.

The customer-facing layer is built as a phygital system: physical and digital working as one continuous experience. On the physical side, the Experience Studio in Jaipur. A 10,000 square-foot space in Malviya Nagar’s industrial area, away from the city’s traditional fabric markets. There is no retail signage, no walk-in fabric shop nearby. It still gets around 80 visits a month from buyers who come specifically to touch and feel fabric they discovered online.

Every swatch is pre-cut, QR-coded, and labelled with a full technical specification. A buyer can walk in, find a base fabric, request a digital print mockup, place a customised order, and leave with the order already routed through the ERP to the right mill and processor. On the digital side, Fabriclore has built a cross-platform webapp that handles fabric discovery, repeat orders, and the full order lifecycle for buyers anywhere in the world. The goal is to make this webapp a sourcing operating system for fashion business, right from fabric discovery to mood boarding to design management to production. Think of it as a backbone for fashion businesses.

The Jaipur studio is the prototype. The intent is to replicate it across every major garment manufacturing cluster in India and eventually abroad. Wherever garments are being made, Fabriclore wants to be there in person, with the digital layer running underneath.

The strategic point of this setup is what the studio plus the webapp lets Fabriclore become. Most fabric sourcing in India still happens through trade fairs and sourcing expos, events that happen once or twice a year, where a buyer flies in, sources what they can in a few days, and leaves. Fabriclore is building the version of that experience that doesn’t ‘end’ when the fair closes. It becomes a permanent, always-on sourcing platform that sits inside a fashion brand’s year-round operations.

Sales offices in Delhi-NCR and Jaipur anchor the offline sales motion, with Mumbai under setup. Surat is the warehouse and fulfilment hub.

This year, Fabriclore moved 5 lakh meters of fabric a month at peak. Several processing mills are now on capacity buyout, which means they are working majorly for Fabriclore.

The triangle that makes the unit economics work

Most B2B fashion sourcing companies build for one customer profile. Fabriclore built for three.

The scale of demand across the three is what makes the triangle work. India has roughly 20,000 readymade garment manufacturers and exporters as registered with Clothing Manufacturers Association of India (CMAI). The country has 50,000+ fashion businesses across offline and online, per ET’s count.

There are also 200,000+ boutiques and tailoring units listed on JustDial. Three customer profiles, three different orders of magnitude. A single-segment B2B platform plays in one of those pools. Fabriclore plays in all three, with each pool feeding the others over time. This is critical to understand.

Manufacturers and exporters bring volume, which gives Fabriclore command over its supplier base. Five lakh meters a month becomes 10–15% of a tier-A processing mill’s monthly capacity, which is the difference between being a price-taker and being treated as a partner.

D2C brands pay in advance, reorder consistently, and grow into bigger volumes, which makes them the segment that compounds; this is where Fabriclore is deliberately scaling.

SME businesses, the local boutiques and tailoring units, are the discovery layer, as in, the smallest SME of today might very well become the D2C brand of tomorrow, and Fabriclore wants to be in the relationship before that transition happens. The cash cycle, built deliberately around the mix, lets the business grow without leaning on debt-funded inventory.

What comes next

In FY 26, the company grew 3X over FY 25 and also become Ebidat positive in the fourth quarter. FY27 is targeting 100 crore.

The cleanest read of the trajectory is the one Vijay uses internally: FY24-25’s full year of crore is now being clocked in a single month. FY25-26’s full year revenue is targeted to land as a monthly figure by March 2028.

Margin headroom is structural. Today’s 17% gross margin moves to a planned 22% through three deliberate shifts: D2C share moving from 42% to 70% of revenue, make-to-order and direct sourcing staying intact, and international revenue scaling from 10% to 20% of the mix.

Dubai is the planned international gateway, both because the Gulf is itself a sizable apparel market and because it is the cleanest stepping stone to Europe and the US.

Also, the Gulf isn’t a cold start. The same round that brought in Peer Capital also brought in Raju Shroff, chairman of Regal Group of Companies, a Dubai-based business his father founded in 1952 as a textile trading firm, which has since grown into a diversified group spanning textiles, technology, real estate, and investments.

Regal operates a distribution network across the Gulf. Raju himself has spent nearly three decades inside the family business and holds an MBA from London Business School. That’s an unusually deep operator profile to have on a startup’s cap table, and Ankur is clear about why it matters.

“Raju is not a passive backer for us. He treats the company the way an experienced operator does: engaged on the actual problems, generous with his time, sharp on the calls that matter. For Vijay, Anupam, and Sandeep, having that kind of voice in the room as they think through international expansion is the kind of help that doesn’t really show up in any term sheet,” said Ankur.

What makes the timing right isn’t only the cap table, though. Two shifts are happening in global fashion sourcing right now, and both are moving the underlying economics in Fabriclore’s direction.

The first is on the demand side. D2C brands are buying smaller with more frequent batches. The era of fashion businesses placing large fixed-volume orders for the season is winding down. Inventory is expensive, trends move fast, and the brand that ties up working capital in fabric is getting caught flat-footed by the next cycle. Smaller batches, more sourcing events per year, more customisation per batch is the mantra. That shift is exactly the workflow Fabriclore is built to absorb: low MOQ, fast turnaround, productised SKUs that can be reordered. The growing D2C market needs the kind of backbone Fabriclore is building.

The second is on the export side. India is the fifth-largest textile and apparel exporter globally, behind China, Bangladesh, Vietnam, and Turkey. Three of those four countries are smaller than India by every meaningful measure: population, total capacity, fiber availability, or domestic demand.

The reason India hasn’t climbed the rankings is the same reason fabric sourcing is operationally feudal at home: the supply chain isn’t vertically integrated. Mills, processors, traders, and exporters operate as separate islands. The work of stitching those islands into a system is the work Fabriclore is doing at home. If India climbs the export rankings over the next decade, companies that have already done the integration work at home will be the ones positioned to absorb the global volume.

The headline number Vijay uses, when pressed for a vision, is conservative by his own description. India’s textile and apparel market is roughly $40 billion in annual size as of 2025. Fabriclore today is around 0.0075% of that. Even 1% of the market in five years would be $400 million in annual revenue.

He calls this his “choti si aasha” (a small hope).

“Fabriclore isn’t a marketplace and it isn’t a financing layer. It’s a platform that takes responsibility for what gets made, which is the only part of the textile supply chain that compounds into a real moat. Combine that with the tailwinds, D2C brands sourcing smaller and more often, India’s exports needing real vertical integration to climb the global rankings, and you’re looking at a once-in-a-decade opportunity in a category that doesn’t get talked about enough in venture,” said Ankur.

18 years after Vijay first met Anupam at Luminous, 14 years after Sandeep joined the engineering business, and 10 years after the three of them named a company at a cafe near the Sector 12 Dwarka metro station, this second act is the one they want to be remembered for.

Test match, third day, second session. A long way to go. They are playing it on the front foot now.

Fabriclore Shark Tank episode:

The global gaming market is projected to reach approximately $130 billion by 2029 from the current $100 billion by 2025. The path to success for new studios has never been more challenging. While subgenres like match-3, word, and puzzle games are well understood and have scaled, they suffer from a “cloning race” where mechanics are easily imitated, and user acquisition (UA) costs are skyrocketing.

In such an environment, the timelines for launching a game, experimenting with gameplay mechanics have crashed as competitors can rapidly flood the market with competing gaming titles. There is rapid alpha decay as studios vie for attention, engagement and revenues. There is incredible pressure

At PeerCapital, we have long held the view that success belongs to studios who view the process of developing and running games akin to running a quantitative trading firm. In a world where data around gaming themes, revenues, engagement metrics, monetization metrics are easily available, success increasingly lies with teams that have superior systems, and decision frameworks to parse information to pick the signal from the noise.

At the heart of a gaming studio lies the combination of the capability of a team to build the infrastructure that can parse signals from data and the decision making frameworks to interpret and make decisions.

The addressable market continues expanding in developing economies driven by multiple infrastructure catalysts - affordable smartphone proliferation, maturing digital payment ecosystems and near-ubiquitous and economical internet connectivity.

Casual games achieve broad demographic penetration through minimal onboarding complexity, making them accessible to new gaming audiences including older demographics and casual entertainment consumers.

The modest time investment required aligns with fragmented consumption patterns - commutes, queues, brief downtime where users increasingly favor rapid engagement loops over extended sessions. This preference for quick dopamine driven experiences persists across shifting entertainment landscapes, evidenced by parallel trends in content consumption.

The Team: Product DNA and Shared History

The common denominator in the founding team is a strong background in building systems to build, scale and maintain games across leading gaming studios viz. PlaySimple Games, Zynga.

Om Misra (CEO): A former Senior PM at PlaySimple Games, Om has previously scaled titles at PlaySimple Games working across, UA, analytics, and feature design across retention and monetization metrics.

Tapan Ranjan: A Senior Software Engineer at PlaySimple Games in his last role, Tapan has extensive experience in building and scaling infrastructure to support rapid ideation, creation and running liveops on games.

Harsh Garg: As Senior Software Engineer at PlaySimple Games during his past stint, Harsh led end-to-end development and release of multiple successful titles including Word Trip, Daily Themed Crossword, Zen Files, and Destination Solitaire

(From L to R): Tapan, Om and Harsh

Looking Ahead

The “micro-moment” of gaming—the 5-minute commute or the waiting line—is a resilient consumer habit. As global ad budgets rise and smartphone penetration deepens in emerging markets, the infrastructure Spill Games is building will allow them to capture this “dopamine loop” more efficiently than almost anyone else in the space.

We are excited to partner with Om, Tapan, and Harsh as they build the future of high-velocity, institution grade game development.

How three founders with very different lives arrived at the same problem, and built Redacto to solve it

In the early days of Ola Money, the ride-hailing app’s wallet offering, the platform faced a problem that required a war footing. Drivers would accept a ride request, mark it as started, and the money would get deducted from the customer’s wallet, without the customer ever stepping into the car.

Amit Kumar, who was then heading product at Ola Money, was tasked with solving this. His team’s solution was an OTP shared at the point of boarding to confirm that the right driver had arrived for the right person. They were worried that it would kill the experience. After all, they wanted the customer to “feel like a King. They just had to get in and get out.” Was this OTP one step too many? Security and user experience, as Amit puts it, have “a very tilted balance.”

Customers, however, loved it.

That instinct to find the friction and build the fix is what led him to Redacto a decade later.

The Unlikely Team Behind an Inevitable Problem

A payments engineer, a cybersecurity consultant, and a fintech sales head walk into a startup. The punchline is that it works, and the reason it works is that none of them arrived at this problem from the outside. They had all lived inside the infrastructure.

Amit Kumar spent two decades in product and technology. He ran Ola Money, one of India’s first digital wallets, before heading engineering for payments at Grab. He then built Octifi, an offline buy-now-pay-later startup out of Singapore that got into Y Combinator (W21), running BNPL not on apps and e-commerce sites but inside mom-and-pop stores. That offline-first bet gave them an average contract value of ~$1,200 against competition doing ~$150. They scaled to three countries and got acquired by LatitudePay, an Australian fintech looking to expand into Southeast Asia. Amit exited fully in 2023.

Shashank Karincheti is a rare hybrid: a security and compliance practitioner who became a product leader. He holds a Master’s in Cybersecurity and is a licensed CPA from the US, certified across information security, technology risk, and technology audit. He spent the better part of a decade as a consultant, sitting across from CFOs and CTOs at multinationals, auditing everything from Bill of materials at engineering companies to data centers in Chennai. (As an aside, he speaks eight languages — Telugu, Tamil, Kannada, English, Hindi, Odia, Bengali, Marathi — a quirk that came from growing up in a family that moved with L&T projects across the country.)

“Consulting is all about understanding the problem,” he says, “understanding exactly where it sits in the framework, figuring out the gap, and then pushing the client past that benchmark.” When fintech came calling during the pandemic, he joined Razorpay, initially as an individual contributor, eventually growing into the role of Deputy CISO and DPO, with a front-row seat to security and regulatory compliance at scale.

Vaibhav Sharma is the third strand. If Amit is the builder and Shashank is the frameworks-to-product architect, Vaibhav is the one who has lived inside the problem from the commercial side. He started in financial services in 2003, selling loans, selling insurance, and even running collections.

He moved into the back-end technology solutioning of payments, eventually landing at Chillr, a mobile banking fintech working with HDFC Bank in the pre-UPI days. He was one of six people on the founding team. Then UPI arrived and changed the game entirely. He went on to lead Zeta’s banking partnerships and launched LendingKart’s unsecured overdraft product. Over the years, he accumulated deep BFSI relationships across NBFCs, insurance companies, and banks that you cannot manufacture, only accumulate.

Amit and Vaibhav found each other through a mutual network in mid-2024. Shashank came in through Amit’s network, but not as a late addition. He had been wrestling with the same problem from inside Razorpay.

It started with a question Amit put to him: do you know how many companies have accessed your personal data? Shashank didn’t. In fact, no one really does.

That conversation quickly revealed the scale of what building a real solution would actually require: real-time deletion at national scale, across every system, every vendor, every downstream dependency. He knew better than most how hard it was, and he knew there was no technology that could solve it. That was the moment he stopped being a consultant to the problem and decided to become the one building the solution.

What Amit had been deliberate about from the start was a founding team with genuinely complementary skill sets: Amit on product and technology, Shashank on product architecture, compliance design, and regulatory interpretation, Vaibhav on sales, partnerships, and the specific kind of consultative selling that enterprise financial services demands. “Skills which are complementary in nature,” Amit says. “Very, very hard to find.”

The best way to understand why that combination matters is to understand the problem they set out to solve. Vaibhav has a story that captures it better than any market sizing slide.

Years ago, he applied for a co-branded credit card. Filled out the form, submitted the documents, waited. The card never arrived. Life moved on, and he forgot about it entirely.

Three years later, a letter showed up. At an address he no longer lived at, from a financial institution, about a card he had no memory of applying for. Somewhere in a system (or several systems) his name was attached to an incomplete transaction. Nobody had closed the loop, nobody knew it was still sitting there. It had just floated in the dark.

He laughs about it now. But what happened to him wasn’t an exception (sadly). It is how enterprise data infrastructure works. A person’s information enters a system and then travels across databases, vendors, and departments with no map and no expiry date. The person it belongs to has no idea. The company holding it often doesn’t either.

That is the privacy engineering problem Redacto exists to solve.

For Amit, the formal moment of clarity came after his exit from LatitudePay. His instincts pulled him toward legal tech, an area where technology was clearly being under-applied. Somewhere in that exploration, he came across India’s Digital Personal Data Protection Act.

He read it the way seasoned engineers read things: looking for the load-bearing walls.

He found one. ‘Right to forget.’

Three words buried in a compliance law, but words that could mean a nightmare for companies. When a person applies for a loan at a bank, their Aadhaar number, PAN card, address, phone number, all of it, spreads across multiple systems, databases, and vendor APIs before the application is even decided. It goes sideways into marketing systems, downward into third-party processors, outward into supply chains. Pinpointing every location where that data lives and then having the ability to anonymize or delete it is, as Amit describes it, “a behemoth problem.”

Boards were beginning to ask their CEOs: ‘Are we compliant?’ No CEO could answer with confidence because no one actually knew where all their sensitive data was. Legal teams could produce policy documents. IT teams could gesture at their systems. But the automated, auditable, provable trail of where every piece of personal information had travelled, that was missing entirely.

“Data privacy is solved by tech, not lawyers. But at its core, it’s a business problem. Every CEO and CMO sitting on years of customer data is now asking the same question: can I legally use this? Can I cross-sell? Redacto doesn’t just make you compliant, it helps you monetize the data you’re already collecting, legally.”

Shashank had arrived at the same conclusion, but from the other direction. If Amit saw the problem through an engineer’s eyes, Shashank saw it through the lens of someone who had spent years inside compliance itself — and knew exactly what it would take to turn regulatory requirements into a working product at scale. The DPDPA is not a checklist. It is a framework with significant legal nuance, and translating that nuance into technology choices requires someone who has lived on both sides of that wall, which he had.

“People always confuse protection with governance,” he says. “Protection is a subset of governance. Personal data was never governed technically.” The rails existed for collection. They existed, partially, for storage. What didn’t exist was the connective tissue: consent flows, data lineage, vendor accountability, deletion propagation, breach notification, the full journey from the moment a customer’s data enters a system to the moment, legally required and technically verified, that it leaves.

He also brings a conviction about what India’s digital future could look like when data is properly governed. Consider this: instead of 10 banks cold-calling a borrower who never asked for a loan, an intermediary, with explicit consent, goes to those 10 banks with anonymized creditworthiness signals and returns offers without ever exposing the individual’s personal details. The consumer gets competitive offers. The banks get warm, qualified leads. Nobody’s number gets shared without permission. Nope, it isn’t some utopian vision; it is a genuine downstream use case of consent infrastructure done right. “How beautiful is that,” he says.

Between Amit’s technical clarity, Shashank’s regulatory and product depth, and Vaibhav’s deep understanding of sales, the shape of the company became obvious. The next question was how this should be built.

Building for the Problem, Not the Slide

Redacto’s first product decision was to resist the temptation to build another point solution. Consent management, data discovery, or vendor risk assessment, any one of those alone is a serviceable product with a definable market.

But the founding team made a different bet. They decided to build a platform, something that covered the entire data lifecycle from the moment of consent through to breach response and data subject access requests.

The logic was simple: when a board asks the CEO if they are DPDPA compliant, the CEO doesn’t want five different software systems communicating a partial answer. They want one. They want to be able to say: yes, provably, here is the documentation.

Building the full platform was the vision. But every platform needs a door. They chose third-party risk management (TPRM)/vendor governance as theirs. The choice wasn’t certainly arbitrary. Shashank had personally conducted hundreds of vendor risk assessments across his consulting years. He knew exactly how they were done: an email, a questionnaire, an Excel sheet returned weeks later, filed somewhere, never verified. He had sat on both sides of that process. That pattern recognition, built over hundreds of engagements, is what told the team this was the right door.

“Excel is the solution for every problem in India,” Amit says wryly. Redacto’s TPRM module replaced those spreadsheets with automated AI-powered assessments, vendor trust centers, and questionnaire workflows. The door, once opened, became a cross-sell into every other module.

Their first paying customer was Razorpay, followed by customers like Pinelabs, Viva Money, Indifi, Oxyzo, Motilal Oswal, and a host of others. These are not small names, and landing them as a company less than a year old required a clarity of pitch that is rare. “We knew we’d built the right platform,” Amit says. “The market was just beginning to prove us right.”

But having the right platform and the right early customers solved only half the problem. The harder half was getting into the room.

The Distribution Insight

The partnership strategy deserves its own chapter because it represents the kind of thinking that separates companies that survive enterprise sales from those that don’t.

About six months into building, the Redacto team faced a question that every enterprise software startup faces eventually: you have built good technology, no doubt there, but the customers you want to win won’t take the meeting because you are too small to be on their approved vendor list. You need a minimum revenue threshold just to fill out the RFP. You are basically invisible.

The answer was to become visible through a proxy. They built structured partnerships with the right alliance partners, like the Big Four, as well as with system integrators, which build core banking solutions and sit inside the banks already. The logic was simple: if a firm that the banks already trust is recommending Redacto, the company doesn’t need to get past the front door because the front door is already open.

There were two moments that gave the market proof it couldn’t ignore. The first came from the government itself. MEITY’s ‘Code for Consent’ initiative, a challenge to build the foundational consent management infrastructure that the DPDPA demands, shortlisted six companies from a pool of 46 applicants. Redacto was one of them, alongside Jio Platforms and four others. Being on that list, in that context, was proof of technology that no sales pitch could have manufactured.

The second came from the insurance sector. The Insurance Information Bureau, the regulatory body for India’s entire insurance industry, put out a public RFP for a privacy technology platform. Every major competitor entered; companies older, better-funded, and better-known than Redacto. However, Redacto won.

“No regulated industry anywhere in the world has a standardized API protocol for propagating data rights requests across an ecosystem. Redacto is working closely with the insurance industry to help make the entire ecosystem DPDPA-compliant,” Amit says.

Vaibhav simultaneously built a parallel partnership track, focused on companies that could become resellers, organizations with their own large customer bases who needed a compliant consent layer on top. For example, a marketing technology firm has come on board specifically because the intersection of consent and marketing was where the regulatory rubber would meet the road. If a promotion cannot go out because consent has not been verified, the marketing team suddenly cares very deeply about the compliance platform.

India’s DPDPA published its final rules in November 2025. The enforcement deadline is May 2027. The penalty for non-compliance is ₹250 crore per violation. RBI, SEBI, and IRDAI have each issued their own sectoral mandates, several of which are already live. The majority of India’s largest banks and businesses have yet to select a privacy technology vendor, and active RFPs are already running across the sector.

By the numbers, the market is large enough to be serious. India alone represents a $5.9 billion annual total addressable market. The global data privacy market sits at $32 billion, with APAC growing at 41% CAGR.

Amit’s goal for the near term is direct: win India. “Ensure that as far as anybody in India wants to be privacy-compliant, they think of Redacto first.”

The PeerCapital Partnership

A market with a live regulatory deadline, a founding team that had spent decades inside the problem, and a platform being built at exactly the right moment. That was the picture when PeerCapital came in.

“We needed a partner,” Amit says. “Not just a VC. Because we knew we would make mistakes. We wanted someone who would help us navigate them, not someone who would push us toward the US market or treat India as a second option.”

What drew PeerCapital to Redacto was a team thesis that is harder to find than it sounds. Shashank brings deep domain expertise in security, compliance and regulation as someone who has spent years in the exact corridors that enterprise privacy sales requires you to navigate. Amit is the repeat entrepreneur who understands both the engineering challenges of building a platform and the business challenges of scaling one, and Vaibhav is the commercial engine, having the kind of enterprise sales instinct that only comes from having been inside the system long enough to know how it actually moves. Three very different profiles, pulling in the same direction.

“It’s not just a young bunch of technically savvy AI-first guys,” as Karthik Prabhakar puts it. “These are people who understand the rules of the game, understand building a business, and understand being in the regulated corridors.” That combination, rare in any founding team, is what got PeerCapital across the line.

On the partnership front, Peer has been actively connecting Redacto to channel partners. The logic is the same one Redacto itself identified: distribution through trusted intermediaries opens doors that a young company cannot open alone.

Shashank captures what the relationship means in practice: “I’ve heard from many founders that VCs say, ‘Show me revenue immediately’. That has never been the case with PeerCapital.”

What Gets Built Next

The platform today covers consent management, AI-powered data discovery across hundreds of system types purpose-built for Indian PII: Aadhaar, PAN, UPI IDs, automated vendor risk management, compliance and governance infrastructure, and an agentic AI services wing that replaces the gap assessments, vendor risk assessments, and privacy impact assessments.

A gap assessment that used to cost tens of lakhs and take months now costs much less and runs in hours. A vendor risk assessment that took weeks takes minutes. The implementation cost drop, against a fragmented stack of 4–5 point solutions plus annual consulting, is approximately 90%.

The sectors Redacto is actively expanding into include BFSI, NBFCs, media companies, consumer-facing brands, and enterprises inside loyalty or membership programs sharing customer data with retail brands.

The longer horizon is what Shashank calls the UPI moment for privacy. Paytm succeeded not when people in Tier 1 cities started using it, but when a person in a small town in Jharkhand learned to scan a QR code and understood that money had been transferred to the merchant.

The analogy he draws for Redacto’s potential is the same: not when a CISO in Mumbai understands consent infrastructure, but when an ordinary person, anywhere in India, understands that they have a right to consent to any digital use of their personal data, and can exercise it.

“That’s the dream,” says Shashank.

This is a question that comes up more often than you would think, maybe not in pitch meetings or investor updates, but in the quieter moments, say, late at night, after a bad week, when you are running on fumes, and you look across at your co-founder or your early team and wonder: do they care about this as much as I do? Are they carrying the same weight? Or am I doing this alone?

It is one of the most human questions in the founder journey. It is also one of the most dangerous, because how you answer it, and more importantly, how you act on that answer, can either strengthen your founding team or completely destroy it.

The structural reality

Let me start with something that needs to be said as plainly as possible: as a founder, you will always feel more pain than anyone else. Always.

This isn’t some ‘special’ flaw in your team, as in, it isn’t only happening with ‘you’. It is structural as you carry the financial risk, the existential risk, the reputational risk, and the long term responsibility. The company is your baby. For good or bad, your identity is fused with it in ways that even a deeply committed co-founder or early employee will never fully match. And, truth be told, that’s okay. It is how it will be.

It is also quite possible that within a founding team, one founder feels the burden more than the other. That probably reflects something about who they are in the founding relationship, say their role, temperament, or stage of life. But, mind you, it doesn’t automatically mean the other person isn’t pulling their weight.

The mistake I see founders make, over and over (sadly), is treating this as a betrayal. They feel the weight, they look around, they don’t see the same level of ‘visible’ suffering in their co-founder, and they conclude: this person doesn’t care enough.

That conclusion is almost always an emotional response, not an evidence based one. The gap between those two things is where founding teams fracture.

The paranoia cycle

Let me dig in a bit deeper. Here is what usually happens. It follows an almost predictable pattern.

A founder is overloaded. By virtue of being overloaded, they start to feel that others are not as overloaded, and therefore are not as committed. They start judging the effort others are making. Slowly, the trust they have placed in their co-founder or key team members starts to pull back.

The team notices, but they may not know exactly what has shifted. They can feel the disconnect. As a result, their performance drops, not because they have stopped caring, but because the dynamics have changed. Trust has eroded, and people perform differently when they sense they are being doubted.

This is the unintentional consequence of founder paranoia. The very thing you feared - that your team isn’t committed enough - becomes a self fulfilling prophecy. Not because it was true, but because your response to the fear made it true.

The only way to break this cycle is to rebuild trust. I mean, trust in both directions: structural trust through clear roles and expectations, and emotional trust through honest conversation.

But there is another thing I would ask any founder sitting with this feeling. Before you judge the other person, look at yourself first. When does this doubt tend to creep in? In my experience, it’s almost always when you’re mentally exhausted, personally stretched, or in a bad state. Mental fatigue creates distrust the way a mirage creates water. It looks completely real, but it’s not actually there. Your rationality has left the building.

Sometimes the weight of leadership means the problem isn’t the other person at all. Sometimes it’s just you, running on empty, projecting your exhaustion onto everyone around you. That introspection am I feeling this because it’s real, or because I’m burnt out? is just as important as any assessment of your co-founder.

The trust equation

A founder friend recently brought up something that I think gives useful structure to all of this. There is a concept called the Trust Equation, originally developed by David Maister, Charles Green, and Robert Galford in their book ‘The Trusted Advisor’. It was written for professional relationships, and, as you can guess, it maps well onto co-founder dynamics.

The equation is simple: Trustworthiness equals Credibility plus Reliability plus Intimacy, divided by Self Orientation.

Let me translate that into founding team language.

Credibility is competence: does your co-founder actually know what they are doing in their domain? Can they be believed when they say they will handle something?

Reliability is consistency: do they follow through? Do their actions match their words over time?

Intimacy is the safety to be honest: can you have real conversations with this person? Can you say “I don’t know” or “I am struggling” without it being used against you?

Self Orientation, the denominator, is the one that changes everything. The more someone is optimizing for themselves, as in, their title, their ego, their personal outcomes, the more it divides and erodes every other element. You can be credible, reliable, and emotionally safe, but if your co-founder senses you are fundamentally in it for yourself, trust collapses.

What I find powerful about this framing is that it makes trust diagnosable. When trust is eroding between co-founders, it’s usually not some vague “we don’t get along.” It’s that one of these four components has broken. Maybe competence was never there, or maybe reliability has slipped, or maybe you have stopped being honest with each other, etc. Identifying which component is broken gives you something specific to work on, rather than just sitting with a general sense of unease.

Measure ownership, not pain

With that said, I truly believe the reframing every founder needs is to stop measuring ‘pain’ and start measuring ownership.

Pain is emotional. It’s invisible, subjective, and impossible to ‘compare’. You have no idea what someone else is actually feeling, and even if you did, ‘matching’ pain levels isn’t what builds companies.

Ownership, on the other hand, is observable. It shows up in specific behaviour. Does this person take initiative and solve problems without waiting for permission or direction? Do they demonstrate instincts to protect the company? When things break, do they step up and defend? Are they responsible for outcomes, not just tasks? Do they think long term, or are they optimizing for themselves and their role?

These are the things that actually matter, and they are actually measurable. You can point to them, discuss them, and build around them.

The question founders should be asking isn’t “does my co-founder feel the same pain I do?” It’s “is this person acting like the owner of what they need to deliver?” Because startups are not built on emotional ownership but are rather built on behavioral ownership. That distinction is everything.

There is a simple thought experiment that helps clarify this: if your co-founder or key team member left tomorrow, would the company be meaningfully weaker? Would you really, viscerally feel the impact? If the answer is yes, then regardless of whether their pain looks like yours, they are clearly contributing something essential.

If the answer is no, as in, if the person leaving wouldn’t fundamentally change anything, then you might have a genuine ownership misfit. But that’s a very different problem from “they don’t seem as stressed as I am.”

The three personas in every founding team

I want to spend a moment on something that I think explains a lot of the tension founders feel but rarely name.

In any founding team, people tend to fall into one of three roles. Not formally, not by title, but by temperament.

The visionary: This is the person who carries the narrative. Think of them as the big picture thinker, out there inspiring people, setting the most ambitious goals, always pushing for the next leap. They are relentless, impatient with slow execution, and always building the story of what the company can become. Their burden is the pressure of existence: does this thing deserve to exist, and can I make it real?

The builder: This is the operator. They are the person who turns the visionary’s thinking into real systems and processes. Product development, hiring, execution rigour, that’s their domain. They thrive on making things work, and they get frustrated when they can’t marry the vision to execution. Their burden is execution pressure: can I actually deliver what we have promised?

The anchor: This is the person who protects the company. Think financial discipline, risk management, team stability, they are the backstop. They make sure the visionary’s ambition and the builder’s velocity don’t tip the company into reckless territory. They are more risk averse by instinct, and their flashpoints are always about pushing back and being more measured. Their burden is the weight of protection: can I keep this thing from breaking?

So, you have three very different people, carrying three very different burdens, showing up in three very different ways. The question is: how do they hold together? I have been thinking about an analogy that I believe captures this well.

The ‘Vitruvian’ founder

Think about Da Vinci’s Vitruvian Man. The image is famous, but the philosophy behind it is what matters here. It’s about proportion, harmony, and structural balance.

A founding team works the same way. The circle is vision, creativity, ambition, the raw energy that pulls a startup forward. The square is structure, discipline, financial control, the systems that keep it from coming apart. The best teams don’t perfectly balance the two. They operate in the tension between them. Too much circle, and you get motion without direction, or, too much square, and you get control without momentum.

But here’s the insight that I think matters most: the Vitruvian Man was never about all dimensions being identical. It was about all dimensions being in proportion. And that’s exactly how founding teams should think about the balance between co-founders. The pain, the commitment, the sacrifice, none of it has to be equal. It has to be proportional. Proportional to the role you play and proportional to where you are in the journey.

A visionary’s burden looks like existential anxiety about whether this thing deserves to exist. A builder’s burden looks like execution pressure — can I actually deliver what we have promised? An anchor’s burden looks like constant vigilance — can I keep this from breaking? All three are real. All three are heavy. They are just different. And the founding team holds together not when everyone carries the same weight, but when the proportions are in harmony.

Note: This is a AI generated. It isn’t perfect, but it gets the point across.

If you strip it down: the circle represents what is natural and instinctive — the vision, the creativity, the ambition that pulls a company forward. The square represents what is deliberate and constructed — the discipline, the financial rigour, the operational systems.

The founding teams that work best are not the ones where everyone is the same. They are the ones where different personas are in the right proportion, and where each person understands that their co-founder’s burden is just as real, even if it doesn’t look like theirs.

Remember: Not equal. In proportion.

Where things actually go wrong

Based on what I’ve seen, founding relationships tend to break down in a few specific places. They are almost always avoidable.

Unstated expectations: This is without doubt the biggest one. It’s never the stated expectations that cause problems. It’s the things you assumed but never said. “I’m all in” sounds like alignment, but all in to you could mean something completely different from all in to your co-founder. It could mean different levels of sacrifice, different time commitments, different financial expectations. If you haven’t scratched beneath the surface of what “all in” actually means: how long, what sacrifices, what you expect from each other, you are building on ambiguity, and ambiguity in a founding relationship is a ticking clock.

Blurred ownership: A lot of the emotional reaction to feeling like others are not pulling their weight comes from unclear responsibilities. “You were supposed to do it.” “I thought you were doing it.” “You never told me I was supposed to.” When ownership zones are undefined, everything feels like a failure of commitment. But it’s actually a failure of clarity. Each co-founder needs a defined domain, clear authority, and clear outcomes they are responsible for. The ambiguity is what creates the relationship hurdles.

Equity that doesn’t match sacrifice: This is uncomfortable, but it has to be said. Most co-founders default to equal splits because it’s the easiest, least disruptive option. But over time, if one person feels they are sacrificing more, doing more, carrying more, an equal split starts to feel deeply unfair. That resentment compounds quietly. This is exactly why reverse vesting exists. It ties equity to continued commitment and performance rather than treating it as a one time allocation. Having these conversations about rebalancing early, before resentment sets in, is far better than avoiding them.

Sitting in silence: This is the killer. A founder starts to feel doubt about their co-founder’s commitment, yet they don’t say anything. The doubt creeps in further and trust erodes. By the time they actually address it, the relationship is already damaged beyond easy repair. A lot of startup failure, honestly, comes from co-founder failures and not failures of product or market. Also, most co-founder failures don’t come from a dramatic conflict. They come from unspoken issues and doubt that were never surfaced. Once you have concluded, in your own head, that your co-founder doesn’t care enough, trust completely erodes. From that point on, there is very little interest in trying to solve anything.

What actually helps

If you are feeling this tension, here is what I would suggest.

Have the hard conversation: But not when you are angry or exhausted, but deliberately. Revisit whether you are still on the same mission. Ask openly if the excitement has changed. Ask what “all in” means today, because it may have evolved. Burnout changes people, life changes change people, and what was true when you started may not be true, say 18 months in. Regular, honest check ins between co-founders are not optional, it has to be built into the functioning infrastructure.

Define roles with painful clarity: Who owns what? What are the outcomes each person is responsible for? Where does one person’s authority start and another’s end? This is the thing that prevents the bleeding of responsibility that causes most of the resentment.

Look at the observable, not the emotional: When the doubt creeps in, come back to behaviour. Is this person taking initiative? Are they protecting the company? Are they delivering on their outcomes? That’s what matters. Not whether they look as stressed as you.

Reframe the question: The most productive version of this question isn’t “do they feel the same pain I do?” It’s “are they acting like the owner of what they need to deliver?” That reframe alone, from emotional intensity to behavioural ownership, tends to give founders a much clearer, more tangible answer.

What we look for as investors

I will close with what we observe from the other side of the table, because I think it’s useful for founders to know what investors are actually assessing when they meet a founding team.

We want to know: will these people be there for the long haul? Can they build a company together? These are late outcomes in early businesses, so we are looking for leading indicators.

Stress resilience: How do these people deal with things when they go wrong? Are they a team that discusses openly, or one that avoids difficult conversations? A fragile team avoids hard topics and runs on emotion. A strong team addresses things directly.

Decision velocity: How quickly does this founding team make hard decisions collectively, say, firing an early employee, going out to raise in a bad market, cutting a product line, and how quickly do they execute those decisions without second guessing? Startups require decisions made with incomplete information. The teams that deliberate endlessly or relitigate after committing are the ones that worry me.

Aligned ambition: Are they talking about the same goal? “We want to build a large company” means nothing if large means 50 million to one co-founder and a billion to the other. Just like “all in,” “large” needs to be defined. Misalignment in ambition is one of the quietest and most destructive forces in a founding team.

The loyalty test: What do they say about each other when the other person isn’t in the room? Do they defend each other? Do they frame decisions as collective? A great founding team, even if the individuals are very different personas, will say: “We debated it together, we discussed it, we came to a joint decision.” It’s the subtle signals: how much they truly value each other that tell you the most about whether this team will last.

The question behind the question

I want to end by going back to where this started. “Is my co-founder feeling the same pain I am?” is the question founders ask. But it’s not actually the question they need answered.

The real question is: can I trust this person to build this with me?

Trust, as I have highlighted, isn’t about matching emotional intensity. It’s about credibility, reliability, honest communication, and a genuine orientation toward the company over oneself. It’s about proportional commitment, not identical commitment.

Vitruvian balance - different dimensions, different strengths, different burdens, held together by structure and proportion rather than by everyone suffering in the same way.

The founding teams that last are the ones where the pain is acknowledged, the roles are clear, the conversations are honest, and each person’s contribution is understood for what it is and not measured against someone else’s.

If you are sitting with this question right now, the most important thing you can do is not to let it fester in silence. You must surface it, diagnose it, and, more importantly, build the necessary structures that let trust compound rather than erode.

Because the real risk was never that your co-founder feels less pain than you. The real risk is that the doubt you carry about it, unspoken, slowly becomes the thing that breaks a team that could have actually gone the distance.

The Gap

Consumer goods is roughly a trillion-dollar global market. China commands 65% of global supply, while India has less than 2%.

Make no mistake, this isn’t because India lacks demand, capital, or talent. It has never had what we call the ‘capability spine’, the integrated layer of product design, quality infrastructure, and manufacturing depth that makes it genuinely easy for brands to produce here.

India has factories. What it hasn’t had is a trustworthy full stack Original Design Manufacturer that can go from a brand’s brief to a finished, compliant, shelf-ready product.

That’s the gap Bidso was built to close.

Our thesis: Complexity+1 (https://peercapitalvc.substack.com/p/indias-manufacturing-opportunity)

At Peer Capital, we have spent the past two years developing a thesis around Indian manufacturing that departs from the consensus ‘China+1’ narrative. Everyone is saying China+1, but Vietnam beats India on cost, and China’s scale is unbeatable for commodities. India’s edge lies somewhere else entirely: in managing complexity.

We call it Complexity+1. India’s manufacturing future belongs to companies that can productize variability: high-mix, varied-volume production, customisation at scale, engineering-intensive products where the complexity of the problem matters more than the cost of labour.

Bidso is a textbook expression of this thesis. The company manages 400-plus SKUs across 25 brand partners, handles seasonal variability, navigates safety compliance across multiple international standards, and delivers customisation, exactly the manufacturing profile where India can build a genuine, defensible advantage.

Bidso’s approach also maps to one of the two founder paths we believe will define Indian manufacturing: start with straightforward products, build operational muscle using domestic demand to reach scale and refine quality, and then move up the complexity curve as capability compounds.

Toys today, adjacent baby care, sports and fitness, and small home appliances tomorrow. The same manufacturing infrastructure: injection molding, metal fabrication, and assembly serves all of these categories. The playbook scales.

The Team

Vivek Singhal, Rahul Agarwal, and Aditya Krishnakumar have spent years inside Flipkart, Udaan, and ITC, watching brands struggle to source quality products from Indian manufacturers. They’ve lived the problem statement.

What convinced us to lead the round, more than the very apparent pedigree, was the depth of operational thinking they brought to the table from day one. This is a team that understands manufacturing economics at a granular level. They have literally run factories, managed production lines, and debated with polymer compounders over material specs.

The Evolution

Bidso today looks very different from the company we backed in 2023. The original pitch was a tech-enabled sourcing marketplace. Over two and a half years, the team pivoted — deliberately, methodically — into a full-stack ODM platform with its own product designs, exclusive manufacturing capacity through a FOCO (Franchise Owned, Company Operated) model, and multi-year licensing deals with IPs like Hasbro, Hello Kitty, Peppa Pig, Harry Potter, and NASA.

That pivot wasn’t easy for anyone, including us. We backed it because the operational logic was overwhelming, and because the team had done the work to prove it before putting it up for review.

Today, Bidso works with 25 brand partners across 400-plus SKUs, has more than doubled revenue in the past twelve months, and is India’s leading manufacturer of children’s outdoor mobility toys.

Why This Matters

Multiple estimates suggest 7-8% of global consumer goods manufacturing could shift to India within five to seven years, a move from $8 billion to over $40 billion. That opportunity won’t be captured by companies that can only manufacture. It will go to platforms that can design, engineer, license, comply, and deliver, all at once.

That’s what Bidso does.

The Round

The ₹63 crore Series A comprises ₹51 crore in equity and ₹12 crore in venture debt. The round brings together Blume Ventures and Peer Capital, along with additional participation from Sadev Capital, and venture debt firm Alteria Capital. More details in the link below.

We backed Bidso before there was anything concrete except pure vision and passion. We are still here because the team has earned that bet, and because we believe the best manufacturing companies are built over years, not quarters.

We are in it for the long build!

In this piece, I want to go into two more questions. The first is one that every founder carries but rarely says out loud. The second is about something that catches founders completely off guard, because nobody warns you about it.

Question 3: “Am I a visionary, or am I delusional?”

This question overlaps heavily with the first two, but it deserves its own space because the emotional weight of it is different. Conviction vs. stubbornness is something you can diagnose with signals and frameworks. But “am I delusional?” is an identity question. It cuts to the core of who you are.

The public advice is not helpful here. Most of what founders hear in the ecosystem is some version of: trust your vision blindly. The contrarians are the ones who change the world. If people don’t get it, that’s their problem.

Yes, sometimes that’s true. But sometimes, a founder who thinks they are being contrarian is actually just being wrong, and the agonizing thing is that in real time, those two situations look identical. From the inside, conviction feels the same whether you are right or wrong.

So how do you tell?

I don’t think there’s a perfect answer, but there are signals. People who are genuinely contrarian — founders who are right when the consensus is wrong — tend to have something to back it up. They have an opinion, yes, but they also have insight: they have noticed something in the market that others haven’t. They have data: something measurable that supports their thesis. Ideally, they have observable behavior: users or customers already doing the thing that the thesis predicts.

People who are delusional tend to have an opinion and nothing else. Or they have deep conviction but can’t point to any evidence that the market agrees, even partially, even in some small corner. They mistake the intensity of their belief for the quality of their insight.

The “what if I’m wrong” exercise I described earlier is critical here. The ability to articulate specific conditions under which your thesis would be wrong isn’t a sign of weakness. It’s the clearest signal I know that distinguishes a visionary from someone who’s just lost in their own narrative.

This is where the circle you build around yourself matters enormously. You need people who will sit with you in this question and help you see clearly, not cheerleaders who reinforce whatever you already believe, and not cynics who dismiss everything, but people who engage honestly. People who will tell you, with care, “I think you might be wrong about this part.” Building that circle, and actually listening to it, is maybe the most underrated founder skill there is

Question 4: “What does success actually do to you when it arrives?”

This one surprises people. Founders are conditioned to worry about failure. But the founders who have gotten past those early hurdles, who have actually tasted some version of success, carry a different kind of weight. Almost nobody talks about it.

The thing about success in startups is it doesn’t resolve uncertainty. It reshapes it. The emotional toll of that reshaping is something that catches founders completely off guard.

I wanted to spend real time on this question because the emotional side of success is profoundly under-discussed. No, I do not mean the strategic implications or the operational challenges; those get plenty of airtime. I’m talking about what success actually does to your inner world.

The feelings nobody warns you about

When success arrives — a funding round, strong early traction, whatever the milestone is — the emotional mix is more complicated than anyone prepares you for. Maybe 40-50% is genuine excitement. There’s a similar measure of pressure. Then there’s this strange 20% that’s hard to name: a feeling of having arrived, of disbelief, of “this is actually real.” Acceptance and vertigo at the same time.

But underneath all of that, there’s something that doesn’t get talked about: imposter syndrome doesn’t get quieter with success. It gets louder.

You would think raising a round would quiet the self-doubt. You have been validated, you think. But what actually happens is the opposite. You start hiring people who are smarter than you in specific domains. Slowly, a voice in the back of your head starts asking: what happens when they figure out I’m making this up as I go? You start talking to more experienced investors, bigger players, and you wonder if they can see through you. You enter new markets, and you feel the limits of what you actually know.

The echo of believing an imposter amplifies with success, not the other way around. And the cruelest part is that you can’t really talk about it. Not with your team, because you are supposed to be the anchor. Not always with your investors, because you are managing the relationship. Often not even with friends outside the ecosystem, because they don’t fully understand the specific pressure you are under.

The loneliness that success creates

Before success, you are an underdog. You can be vulnerable. You can say “I don’t know” freely, because nobody expects you to have all the answers. But once you have raised, once you have scaled, once you have a team of people looking to you, the emotional rules change completely. You become the stability anchor for the organization. You can’t vent to your employees and you can’t show the full extent of your doubt.

Also, depending on the relationship with your investors, founders often start filtering what they share. Not lying, just curating. Sharing the version of reality that maintains confidence rather than the full picture. I think that’s the wrong instinct, but I understand why it happens.

What this creates is a strange kind of isolation that’s invisible from the outside. The founder has more people around them than ever, and yet they are more alone in their emotional reality than they were when they were building in a room with two co-founders and no money. They become their own shock absorber. Nobody else can live it the way they are living it.

When your identity fuses with the company

There’s another thing that happens with success that is genuinely dangerous: founders stop being able to distinguish between themselves and their company.

It starts subtly. You take responsibility for everything, which feels right. After all, it is your company. But over time, “the company has a customer acquisition problem” becomes “I have a customer acquisition problem.” “The company’s growth is slowing” becomes “I’m slowing.” The company’s failures feel like personal failures. Its setbacks feel like your setbacks.

This is hyper-responsibility, and it creates an almost parental sense of duty that generates enormous anxiety. Every decision feels existential because it’s no longer just about the business - it’s about you. Your worth, your competence, your identity. The reframing that founders need is to separate the company’s performance from their personal identity. If the company isn’t acquiring customers, that’s a company problem to be solved. It’s not a referendum on who you are. That sounds simple. It’s psychologically one of the hardest things in the founder journey.

The comparison trap

The startup ecosystem runs on comparison. Who raised more? Who’s growing faster? Whose round was at a higher valuation? It’s like a song stuck in your head, it plays constantly in the background whether you want it to or not.

Even when you are doing well by any objective standard, your mind starts placing you in a hierarchy. Where am I relative to them? Am I falling behind? This distorts your sense of progress. You could be building exactly the right thing at exactly the right pace, but because someone else announced a bigger round, it doesn’t feel like enough. The benchmarking is unconscious, relentless, and it creates emotional pressure that has nothing to do with how your business is actually performing.

The swing between euphoria and despair

And then there are the emotional oscillations. Almost every founder I work with experiences this, but very few talk about it openly.

One day, things feel incredible. A great release, a key customer signed, a metric moving in the right direction. Two days later, something breaks, a hire doesn’t work out, a deal falls through, and everything feels fundamentally flawed. Then three days after that, another signal comes in, and maybe things are workable after all.

This swing between euphoria and despair happens on a cycle measured in days, sometimes hours. And what makes it so disorienting is that founders believe it shouldn’t be happening. You have raised money. You should have clarity by now. The oscillation feels like evidence that something is wrong.

But this is normal. It’s the nature of building at the early stage, where you are making decisions on incomplete information, and every signal feels like the most important one. The goal isn’t to eliminate the swings, you really can’t. The goal is to shorten your recovery time from each one. You’d think that it comes down to instinct. Nope, it’s a skill, and like any skill, it needs to be deliberately built.

The fear that evolves but never leaves

Here’s what I want founders to understand about the emotional arc: the fear doesn’t go away. It changes shape.

At the start, the fear is straightforward: fear of failure. What if nobody buys this?You are an underdog, you have nothing to lose except the venture itself, and in a strange way that makes the fear manageable. It’s clean.

But as success arrives, the fear morphs into something more complex: fear of responsibility. You now have employees whose livelihoods depend on you. Investors who have trusted you with their capital. A reputation, a public narrative, concrete milestones that can be measured and scrutinized, and anything concrete is breakable.

When you had nothing, failure was abstract. Now that you have something, failure is specific. You can picture exactly what it looks like, and that specificity makes the fear heavier, not lighter.

So what helps?

I don’t want to end this with platitudes. But I have watched founders navigate the emotional reality of success, and I have noticed what the ones who handle it well tend to do.

They practice detachment from daily metrics. Not indifference, they do care deeply about the numbers, but they don’t let a single day’s dashboard define their emotional state. The metrics are the North Star for the business, not for their self-worth.

They get comfortable saying “I don’t know.” Some of the best founders in our portfolio say this regularly. It’s not a sign of weakness. It’s a sign that they are secure enough to be honest about the limits of their knowledge, which means they are also likely to seek the best answers rather than performing with confidence they don’t feel.

They slow down their emotional reactions. A bad signal comes in, and instead of spiraling, they pause. A good signal comes in, and instead of celebrating prematurely, they pause. This is emotional regulation, and it compounds over time in ways that are hard to see in the moment but transformative in the aggregate.

More importantly, they find something outside the company that anchors them. It could be an obsessive fitness regime for one. For others, it might be family, a hobby, or travel. I know this sounds like wellness advice. But I have seen the difference it makes when things get turbulent, and they always get turbulent. A founder with a non-work identity has a floor that the company’s performance can’t break through. A founder without one is in freefall every time the metrics dip.

I will say something that might sound unusual coming from a VC: emotional stability is a skill. It’s not a personality trait you are born with. It needs to be built, deliberately, the same way you build a product or a team. The founders who invest in building it are, in my experience, the ones who go the distance.

The uncomfortable truth

These four questions — across both parts — don’t have clean answers. There is no framework that resolves the tension between holding on and letting go, no podcast that prepares you for the emotional complexity of getting what you wanted and discovering it comes with a whole new set of problems.

But they do have navigational strategies, and the one that connects all four is this: the founders who navigate them well don’t do it alone. They build a small, deliberate circle of people who will tell them the truth. People who will say “I think you might be wrong about this” and mean it as an act of care, not an attack.

Everything in the public domain ignores the emotional energy that founders need. Not motivational energy, but the kind that comes from people who help you hold on to your conviction when things look bleak, and who help you question it when things look too good. People who help you embrace the contradictions rather than pretending they don’t exist.

I think the biggest thing public founder discourse gets wrong is that it treats uncertainty as a problem to be solved. It is, in fact, the operating condition. The founders who build lasting companies are the ones who got comfortable building inside uncertainty — strategically, operationally, and emotionally.

If that sounds hard, that’s because it is. But at least you know you’re not the only one sitting with these questions at 3am.

As I close this piece, I realize both parts together make for a long read. But there is a lot left unsaid even now on each of these topics, and we chose depth over brevity. These questions deserve more than surface-level treatment, and if you have read this far, you probably agree.

What followed was one of the most revealing conversations I have had in years.

It wasn’t because anyone shared some breakthrough insight or a clever framework. It was because the things these founders were wrestling with - pre-funding, during funding, post-funding, relationship expansion - were almost identical. None of them had found answers online, in a podcast, or in any of the founder playbooks that circulate on the internet.

That evening crystallized something I have been thinking about for a while. There’s a category of questions that define the founder journey more than any business problem does. I want to go deep on them, starting with two in this piece and two more in Part 2.

The gap no one talks about

Before I get into the questions, some context on why they hit so hard.

Only about one in a hundred startups gets the limelight of funding. Out of that, maybe one in a hundred ends up being truly valuable: generating returns, getting to an IPO, whatever the milestone is. The funnel for glamour is impossibly narrow.

But the stories that get published are the heroic ones. The graveyard, which is at least ten times larger, barely gets a mention. That creates a massive gap between what the public discourse of building a company looks like and what the private reality actually feels like.

Every founder I speak to carries questions that they believe are unique to them. Questions that feel too personal, too existential, too embarrassing to bring up in ‘polished’ circles. The truth is, these questions are near-universal. They just don’t have the kind of clean answers that make for good LinkedIn posts.

Question 1: “Why don’t they see what I see?”

This is probably the most common ‘quiet’ question founders carry. It applies to investors (including us), potential customers, ecosystem players, basically anyone the founder needs to convince.

I share the following with every founder who tells me people don’t see what they see: the problem is almost never that your vision is wrong. It’s that somewhere between your head and their understanding, something is getting lost. Before you conclude that the world doesn’t get it, you need to figure out where the gap actually is.

The four gaps

In my experience, when people don’t get what a founder is building, it usually falls into one of four gaps.

The clarity gap: This is the most common and the most fixable. The idea might be brilliant in your head, but it’s not landing because of how you are presenting it. You are overcomplicating it, burying the insight in jargon, or leading with the wrong thing. I will come back to this with a specific test that I think every founder should apply.

The evidence gap: You have a compelling story, but no proof points. Nothing that demonstrates that what you are saying actually works. People hear a narrative, but there’s nothing to anchor it in reality. A story without evidence is just an opinion, and opinions are easy to dismiss.

The timing gap: This is the one that most founders believe is their problem, and sometimes they are right. The zeitgeist moment for your product doesn’t exist yet. You have genuine foresight and vision, but the market conditions that would make people intuitively understand your thesis simply aren’t there, yet you are early, not wrong, but the distinction doesn’t help you in a fundraising meeting.

The signal mismatch: This is the one that’s hardest to see. As a founder, you are talking about vision, you are talking about an aspirational future. But the person across the table - the investor, the potential hire, the customer - is processing risk. They are seeing both opportunity and risk, and they are weighing them differently than you are; you are transmitting on one frequency; they are receiving on another.

The critical thing is that these gaps require different responses. You can’t solve a clarity problem by showing more evidence. You can’t solve a timing problem by simplifying your pitch. And you definitely can’t solve any of them by concluding that the other person “just doesn’t get it” and moving on. Before anything else, you have to diagnose which gap you are dealing with.

The compression test

Here’s a practical test I use and recommend to every founder I work with: can you explain your problem in one sentence?

Mind you, not your solution, not your vision. The problem.

If you can simplify it into one clear line and people still don’t understand what you are doing, you likely have a fundamental clarity issue. If you simplify it and they immediately get the problem but question your approach, that’s actually a much better position to be in. Trust me, that’s a conversation worth having.

Let me give you an example. If you describe your company as “an AI-first platform for ABC,” that’s a fine line, but it doesn’t explain much. It doesn’t tell me why I should care. But if you reframe it to something like: “Indian households spend over a lakh on ABC, but the entire discovery experience is broken — our platform increases discovery and therefore consumption and commerce by X percent” — now I understand the problem, the scale, and the mechanism. Same company, but radically different level of clarity.

The compression test forces you to separate what matters from what’s decorative. It’s brutal, because founders naturally want to explain everything. But the ability to compress, to take a complex business and make it instantly graspable, is what differentiates the founders who get buy-in from the ones who keep wondering why no one understands them.

Move up the conviction ladder

Here’s another way to think about the communication challenge. When you are talking to someone who isn’t getting it, ask yourself: what am I actually communicating? An opinion? An insight? Data? Or observable behaviour?

These sit on a ladder, and they carry very different weights.

An opinion is just a belief. “This market is going to be huge.” That’s fine, but it’s easy to dismiss because it’s unfalsifiable. Truth be told, anyone can have an opinion.

An insight is a step up. It shows you have noticed something that others haven’t. “We have observed that consumers spend X but discover products through Y, which is fundamentally broken.” That’s more interesting because it suggests you are seeing something real.

Data is stronger still. It grounds your insight in something measurable. “Our users do X, they return Y times a month, they spend Z on average.” Now we are having a different conversation entirely.

The most powerful is demonstrated behaviour. Not what you claim will happen, but what is already happening. Users coming back, organic growth, customers co-creating use cases you never anticipated. That’s the signal that’s hardest to argue with.

When people aren’t getting what you are building, the move is almost always the same: go up the ladder. Less story, more evidence; less opinion, more data; less future projection, more present reality. Airbnb is the classic example of this. When the founders were raising, the skeptics were relentless: why would anyone stay in a stranger’s house? How do you know it’s safe? The Airbnb founders didn’t fight the question. They didn’t try to out-argue the skepticism. They instead showed the usage curve. They showed that people were already doing it, already comfortable, already coming back. The data did what the narrative alone couldn’t.

Stripe is another version of the same story. When they launched, payments were considered a solved problem. Everyone said there was no room. Stripe didn’t waste energy arguing about the market thesis. They showed what they’d built and let the product speak.

If you take one thing from this: when people don’t see what you see, don’t assume it’s their problem. Run the compression test and figure out which of the four gaps you are dealing with. Remember to move up the conviction ladder — less story, more evidence; less future, more present.

Find your believers, don’t try to convert the room

One more thing on this question, and it’s something I wish more founders internalized early: you don’t need everyone to believe in you. You need the right someone.

Founders burn enormous energy trying to convert skeptics. They treat every “no” as a challenge to overcome, every unconvinced investor as a mind to change. But this is actually counterproductive for two reasons.

First, if you over-convert - if you persuade someone who never intuitively got it — you may end up with an investor who backed you because of your sales ability, not because they believed in the thesis. When the chips are down, when things get hard (and they will), that investor won’t have the conviction to stay with you. They will second-guess, because they were never truly bought in.

Second, the energy spent converting non-believers is energy not spent finding your natural believers. They certainly exist. The best founders I have seen don’t try to convince the room, they read the room. They figure out who’s leaning in and who’s politely nodding, and they focus their energy accordingly.

Now, there’s an important caveat here. If you go out and genuinely nobody is believing — if after real effort, you have zero believers — then the question you need to ask yourself gets much harder. That leads directly to the next question.

The “what if I’m wrong” exercise

This is something we use extensively at Peer Capital, and I think every founder should do it. Write it down — don’t just think about it, because what isn’t written isn’t measurable.

Two columns. In one: what would prove me right in six months? What specific signals, metrics, or milestones would demonstrate that what I’m building is working? In the other: what would prove me wrong in six months? What signals would tell me that my thesis is off, my product isn’t landing, or my market isn’t there?

This exercise does something powerful. It shows that you are self-aware. That you have conviction, but you are not delusional. That you have thought about the downside without being paralyzed by it. Honestly, when a founder can articulate both what success and failure look like for them - clearly, specifically - that’s one of the strongest signals of intellectual honesty I can see across a table.

Every founder says they are convinced. The ones who can also tell me the specific conditions under which they’d be wrong, those are the ones I trust.

Question 2: “When is it conviction, and when is it stubbornness?”

This question lives right next to the first one, but it’s a fundamentally different beast. And it’s the one that, in my experience, most determines whether a founder builds something lasting or spectacularly flames out.

The ecosystem tells founders to have unwavering conviction. Never waver, never blink, trust your vision blindly, it is all that founders hear. Yes, conviction is essential because without it, nothing gets built. But what doesn’t get talked about nearly enough is how conviction, taken too far or applied to the wrong thing, becomes stubbornness. Stubbornness, dressed up in the language of vision, has destroyed more startups than any market downturn.

So what’s actually the difference?

Conviction is holding on to your vision while updating the path based on what you are learning. Stubbornness is holding on to the path even when reality contradicts it.

That distinction sounds simple. In practice, it’s agonizingly hard to see in real time.

Be stubborn about the problem, not the idea

Your problem is your North Star. You need to be fully, unshakably convinced about the problem you are solving. But the idea - the specific product, the specific approach, the specific go-to-market — that’s what evolves. That’s what has to evolve, because that’s what’s actually encountering reality on the ground.

Think about Slack. Most people don’t know this, but Slack started as a gaming company. The game itself never found real customer pull. But the internal communication tool they’d built for their own team had something the game didn’t -genuine user love. The founder, Stewart Butterfield, didn’t cling to the original idea. He followed where the pull was. That was conviction in the problem of how teams communicate, not stubbornness about the product he’d originally set out to build.

Shopify is another version of the same story. It started as a platform to buy and sell snowboards. The tools they built to power that store turned out to be far more valuable than the store itself. Today, Shopify powers a massive chunk of consumer internet. That only happened because the founders were willing to let go of the original idea while holding on to the deeper insight about what merchants needed.

Instagram was originally a complex location-sharing app called Burbn. The founders noticed that users were ignoring most of the features but obsessively using the photo-sharing piece. They had the clarity to strip everything else away and follow the signal. That’s high conviction in user behaviour, not in the original idea.

Netflix went from DVDs to streaming to original content. The idea kept evolving, radically, but the mission — changing how people experience entertainment — never wavered.

The pattern is consistent. As I often say to founders: be intense about your mission, but flexible about your approach. Be married to the problem, not the solution.

What stubbornness actually looks like

The tricky thing is that from the outside, conviction and stubbornness can look identical. Both founders are passionate, both are persistent. Both will tell you they believe deeply in what they are building. The difference shows up in how they respond to signals.

A conviction-led founder, when faced with evidence that something isn’t working, will engage with it. They will ask: what’s this data telling me? What’s changed since our last iteration? What am I learning? They separate their narrative from their numbers. They might start a conversation by acknowledging what isn’t working before explaining what they have done about it.

A stubborn founder does the opposite. They get defensive. They explain away the signals. Retention is low? “The market isn’t ready yet.” Engagement is falling? “Once we raise more money, we will fix that.” Revenue quality is poor? “At scale, this sorts itself out.” There’s always a future projection to compensate for a present weakness. The default explanation is external: the market, the timing, the investors who “don’t get it.”

I see this play out constantly in how founders talk about their metrics. The ones who worry me are the ones who don’t want to talk about the hard numbers — retention, engagement, revenue quality — and instead steer every conversation toward the vision. The vision is futuristic and therefore unfalsifiable. You can’t argue with a future that hasn’t happened yet, and that’s exactly why it becomes a hiding place.

Are you learning faster than the market is rejecting you?

This is probably the most useful lens I can offer on the conviction-stubbornness question.

In the early stages, things will not work. That’s a given. The question isn’t whether things are working perfectly, they won’t be. The question is whether each iteration is visibly improving the metrics. Can you point to what’s changed? Can you articulate what you have learned? Is the product meaningfully evolving, or is it the same product with cosmetic updates?

A conviction-led founder is experimenting regularly. The product evolution is visible. The narrative evolves with the evidence. A stubborn founder ships the same thing repeatedly and blames external factors when the results don’t change. Same product, same results, different excuses.

Then there’s the customer signal. Are you having frequent, honest customer conversations? Have you made pivots based on user feedback? Because that shows you are convinced about the problem and willing to let the customer teach you how to solve it. A stubborn founder says “customers don’t understand this yet” and keeps going back to the vision.

There’s also a subtler signal worth watching for. If only a few people are confused about what you do — and let’s say those few are investors, or a specific type of customer — that might be a communication problem or a timing issue. But if some users are genuinely loving your product, a few investors are immediately leaning in, and your early hires are energized by what’s being built, then you might be early, but you are probably not wrong. That’s conviction territory. The signal to hold on to is the pull, not the noise.

What I’m actually looking for across the table

A lot of founders think that when they are in a room with a VC, they are being judged on how confidently they speak. How commanding their presence is. How polished their articulation is. I want to be honest: that’s not what I’m looking at.

What I’m trying to see is this: when pushed - when given an alternative point of view, a piece of contradicting data, a skeptical question - is this founder reality-seeking, or reality-defending?

A stubborn, defensive founder treats skepticism as a personal attack. They get dismissive. You question the market size, and they hear “you don’t believe in my vision.” You push on a metric, and they redirect to the narrative. Confidence with defensiveness, that’s the profile that worries me.

A conviction-led founder does something different. They get curious. They engage with the challenge. They might say, “Yeah, that’s a fair question - here’s how we are thinking about it.” They acknowledge the uncertainty that exists in their building. Confidence with curiosity, that’s the profile that excites me.

The irony is that the founders who are most willing to say “I don’t know, but here’s how I’m figuring it out” are the ones who demonstrate the strongest conviction. Because they are secure enough in their mission that they don’t need every conversation to be a validation exercise.

What’s next

These two questions — why don’t they see what I see, and when is it conviction versus stubbornness — are the ones that tend to dominate the early stages of a founder’s journey. But there are two more that hit differently, and often harder.

In Part 2, I’ll go into the question every founder carries but rarely says out loud — am I actually a visionary, or am I just delusional? — and into something that surprises most people: what success actually does to you emotionally when it arrives. The imposter syndrome that gets louder, not quieter. The loneliness that scales with the company. The fear that doesn’t go away but changes shape.

As I close this piece, I realize it’s a long one. But there is a lot left unsaid even now on each of these topics, and we chose depth over brevity. These questions deserve more than surface-level treatment, and if you’ve read this far, you probably agree.

Part 2 drops soon.

Every one of these categories has produced at least a couple of breakout companies. Food delivery has Zomato and Swiggy; grocery has Blinkit, Zepto, and BigBasket (the OG in many ways, which reinvented itself); home services has Urban Company and an upcoming Snabbit. Fintech, which increasingly powers consumption inside the home, has Paytm, PhonePe and Razorpay.

There is a useful dichotomy here. Delivery to the home is well-funded, services at the home are well-funded. But the home itself?

Not the things that arrive at the home or the services consumed inside it, but the physical space: the décor, the lighting, the textiles, the objects and appliances that make a home feel ‘home’... like it belongs to someone.

India’s overall home economy market is projected to reach ~$80 billion by 2030. That makes it as large, if not larger, than several of the categories that produced India’s most celebrated startups.

And yet, this market has produced exactly zero breakout companies, no unicorn(s), no blockbuster IPO, no name that digitally native consumers reflexively turn to when they move into a new apartment.

We (VCs) have funded delivery to the home, we have funded services at the home, yet we haven’t (recently) funded the home itself. One could arguably say that VCs are still reeling from what happened when they previously funded home. However, we sincerely believe that the tide is set to change.

The Sociology of Consumption at Home

To understand why this matters now, it helps to step back and see the bigger picture.

The Indian home is undergoing a sociological transformation. It is no longer just ‘shelter’. For a growing segment of urban India, the home has become a workspace, a social venue, a wellness retreat, and a canvas for self-expression, often all at once.

The pandemic accelerated this, but the shift has outlasted it. Hybrid work is now structural, hosting at home has replaced going out for a significant share of social occasions, and for India’s digitally native young homeowners, how a home looks is as much a part of their identity as what they wear or where they eat.

This transformation has reshaped consumption patterns inside the home dramatically. Vaaree, one of our portfolio companies, recently put out its ‘Vaaree Home Index 2025’, a study built on aggregated data from over 500,000 customers across 48 home categories and 50+ million browsing sessions. It captures this shift in sharp detail.

Storage and organization products grew 270% year-on-year, reflecting a move from hidden utility to visible design. Wall décor adoption grew 64%, ambient lighting adoption rose 70%, and rugs and carpets saw over 102% GMV growth. The average selling prices of cushion covers, one of the most basic home textile categories, rose 42%, suggesting consumers are actively trading up, not just buying more.

The most striking finding is that consumers who browsed curated ‘Looks’, coordinated room-level styling rather than individual products, converted at 580% higher rates than those using standard product-led discovery. That’s by no means a marginal improvement; it suggests the consumer’s mental model has fundamentally changed. To put it in other words, people aren’t shopping for just a lamp; they’re shopping for a living room.

More importantly, this isn’t just a metro phenomenon. Emerging cities, including Indore, Lucknow, and Nagpur recorded some of the fastest growth in styling adoption. Jaipur climbed four positions to become India’s second most stylish city. The behavioural shift is spreading well beyond the obvious urban centres.

When every other form of consumption at home has been transformed by technology, and the consumer’s relationship with their physical space is evolving rapidly, the absence of a category-defining company in home furnishing is more than just an interesting gap; it’s a structural mispricing (dare we say).

Not All ‘Home’ Is Created Equal

Before we go further, a distinction that matters enormously: the ‘home’ category is not one market. It’s at least two fundamentally different businesses, and conflating them is one of the reasons this space confuses investors.

Furniture is one business. Everything else is another.

India’s home furniture market is roughly $25 billion, growing to $40 billion by 2031. It’s a massive category, no doubt. But for venture capital, it has deep structural problems. Furniture is high-ticket, high-touch, and physically demanding to deliver. Consumers overwhelmingly want to sit on a sofa before buying it, feel the mattress, and check the wood grain. Online-only furniture brands have struggled globally, and in India, the experience has been particularly brutal. The companies that have survived have done so by building expensive offline showroom networks, which defeats the capital-efficiency thesis that makes venture-backed businesses work.

Furniture also has long production cycles, high return rates on large items, and working capital intensity that bleeds early-stage companies dry. Building a furniture brand online is extraordinarily hard. It probably needs to be an offline-first business, and offline-first businesses, while potentially excellent companies, are rarely great venture outcomes.

Non-furniture home: textiles, décor, wall art, lighting, storage, rugs, kitchen accessories, and small appliances is a fundamentally different story. These are lower-ticket, higher-frequency, more shippable, and far more amenable to digital discovery and purchase. India’s home textile market alone is worth over $11 billion and projected to reach ~$16.7 billion by 2030. The home décor market is valued at USD 26.88 billion in 2025 and is projected to reach USD 42.36 billion by 2034. Crucially, this is where consumer behavior is shifting fastest.

So, when we talk about the venture opportunity in ‘home’, we mean non-furniture. That is where the economics, the consumer behavior, and the technology convergence all point.

The immediate question that one might ask is, ‘Isn’t horizontal e-commerce doing this already?’

We think that there is a deeper reason why horizontal e-commerce hasn’t won the category outright. You can standardize grocery; you can also algorithmically recommend fashion based on past purchases. Home aesthetics are fundamentally more contextual. What looks right depends on the room’s dimensions, the light, the existing furniture, the flooring, or the vibe someone saw on Instagram that morning.

This is primarily a discovery problem, not a search-and-filter problem, and it requires a fundamentally different kind of technology solution.

There’s a larger (and obvious) question for venture here: defensibility. So what does a winning company actually look like?

We think the answer lies less in any single product brand and more in the platform that sits between consumer intent and fragmented supply. The company that helps a consumer go from “I want a beautiful, functional kitchen” to a coordinated set of products from the right brands and manufacturers is building something structurally different (when compared to a D2C brand). A standalone product brand is one business. The platform that curates, aggregates, and distributes across brands and manufacturers, while owning a branded product line at the top of the catalogue, is a fundamentally different and, we believe, more defensible one. (We’ll expand on the operating model for this later.)

That platform thesis also opens up an angle most domestic-focused investors miss: exports. India is already the world’s second-largest exporter of home décor by volume. The manufacturing clusters — Panipat for textiles, Jodhpur for furniture, Moradabad for metalware, Karur for home textiles — produce world-class goods, but the supply chain for non-branded products is overwhelmingly export-oriented. These suppliers know how to meet global quality standards. What they lack is a technology layer for discovery, aggregation, and brand building. Today, these flows are largely served by trading houses with no technology. A platform that can organize this supply efficiently captures value in both directions: domestic marketplace and cross-border aggregation.

Emerging Hotspots Within Home

Beyond the core categories, several adjacent segments are showing early but unmistakable signs of venture-scale potential. These are worth watching closely.

The small appliances category: The market is already seeing a good uptick and is nearly worth $18 billion today and expected to reach ~28 billion by 2031. Kitchen appliances alone are an $11.4 billion market. What makes this category exciting is the convergence of design and AI.

Energy-efficient and green home devices: India’s residential sector accounts for 24% of the country’s total electricity consumption, making it the second-largest consuming segment.

As the Bureau of Energy Efficiency tightens star-rating standards (new benchmarks took effect in January 2026), consumers are actively trading up to inverter-grade, energy-efficient appliances. Smart fans alone have become a breakout sub-category: Atomberg, which built its brand around BLDC motor efficiency, has demonstrated that Indian consumers will pay a premium for products that save on electricity bills while looking good. The broader opportunity extends to smart lighting, energy-monitoring systems, and solar-integrated home products. The intersection of ‘saves money’ and ‘looks modern’ is a powerful wedge, given that electricity costs are a meaningful share of monthly ‘home’ spending.

Smart security and home automation: India’s smart home security market was valued at roughly $1.7 billion in 2025 and is projected to grow to 14.7 Billion by 2034 at 26.45% CAGR. Security cameras and smart locks are leading adoption, with particularly high consumer satisfaction scores.

What makes this interesting for venture is the shift from builder-installed legacy systems to consumer-purchased, app-controlled, DIY devices. The cost of a basic smart camera has dropped to a fraction of what a professionally installed system once cost. The trigger for mass adoption is the same one driving the broader home economy: millions of new homeowners who expect their home to be as connected and controllable as their phone. The winner here will likely not be a hardware company alone, but the platform that integrates security, lighting, climate, and energy into a single, coherent home operating system.

Home wellness: This is early but directionally clear. India’s water purifier market alone is worth $3.6 billion and growing at nearly 8.5% annually. Air purifiers, while still seasonal and concentrated in North India, are growing at 13.5 CAGR as pollution awareness spreads. But the real opportunity is in the convergence of air quality, water quality, ergonomic furniture for home offices, circadian-aware lighting, and sleep optimization into a ‘wellness layer’ for the home.

Post-pandemic, a meaningful segment of India1 consumers now treats the home as a wellness environment. The company that can curate and sell this wellness stack, rather than individual appliances, is building toward something with significantly higher lifetime value and cross-sell potential.

Each of these segments might look niche individually. But they share a common thread: they’re all categories where consumer willingness to pay is rising, where the product is becoming smarter and more connected, and where no incumbent platform owns the discovery or distribution layer. For an investor, the question is whether the platform that wins in core home (décor, textiles) can expand into these adjacencies and become the operating system for the entire home. That’s the larger play; while each of these, individually, might be significant, too.

The Operating Model That Wins

Understanding what to sell in the home category is only half the question. The other half, and arguably the more important half for investors, is how to sell it.

We think the answer requires splitting the product catalogue into three tiers: head, torso, and tail.

Head products: the top 10-20% SKUs that drive the majority of sales. This can support branded plays. These are the hero products: a signature cookware collection, a distinctive lamp range, a curated set of small appliances. Here, building a brand makes sense, and the margins can support the marketing spend; the repeat rates are higher because consumers come back for complementary pieces, and the brand itself becomes a moat.

Torso and tail products: the thousands of SKUs that fill out a catalogue and serve the long tail of consumer taste. These need a completely different model. Trying to build a brand across thousands of SKUs at home is a capital trap. The inventory costs will crush you, and the receivables will stretch you. The working capital cycle becomes a black hole that eats margins and fundraises alike.

For the torso and tail, the operating model has to be marketplace-led. Asset-light, negative to zero working capital. The platform curates, the supplier manufactures and ships, and the consumer gets quality and delivery reliability without the platform ever touching inventory for these products.

This isn’t a theoretical point we are making. It’s the lesson learned (expensively) by the first generation of home commerce companies that tried to be branded retailers across the entire catalogue. The ones that survived narrowed their owned-inventory footprint dramatically. The next generation of winners will start with this insight.

Home is not a pure D2C problem, and it’s not a pure marketplace problem. The businesses that tried to be either have struggled. The durable model is hybrid because in a space where trends and taste shift constantly, inventory is not an asset; it’s a liability.

Own the brand in the head in due course, not necessarily on day 1; run the marketplace in the torso and tail. This gives you the consumer relationship and margin upside from branded products, and the capital efficiency of a marketplace for everything else. If you can do this while maintaining quality control and delivery timelines across all three tiers, you build something that’s genuinely hard to replicate, an asset-light marketplace with high margins and strong customer trust.

The best analogy might be what Meesho did for fashion: built an enormous marketplace for unbranded and semi-branded long-tail products, maintained delivery quality through supplier management systems, and achieved economics that pure brand plays couldn’t match. The home category, especially non-furniture, has identical structural characteristics and is waiting for its version of this model.

Why Now: Six Forces Converging

We counted six forces that have converged in the last couple of years. Any one of them would be interesting. Together, they represent a genuine inflection point, a set of conditions that genuinely didn’t exist even two years ago.

One: Rising disposable income and the India1 consumer: Blume Ventures’ Indus Valley Report offers a useful framework that describes India not as one market but three. India1, the top ~10%, roughly 30 million households with ~$15,000 per capita income drives nearly two-thirds of the country’s discretionary consumption. India2, the aspiring majority of ~70 million households at ~$3,000 per capita, consumes heavily but is price-sensitive. India3 is not yet a market for most startups.

The home economy is one of the purest India1 categories. When someone upgrades from a rented 2BHK to an owned 2/3BHK, they want to have an ‘imprint on it’, and they don’t gradually furnish it. They furnish it all at once and keep upgrading it. And as India1 deepens, as incomes rise within this cohort rather than the cohort dramatically widening, the willingness to spend on home as identity increases meaningfully.

India’s residential property market achieved an 11-year high in sales volume in the first half of 2024. Real estate attracted $4.15 billion in private equity investments in 2024, a 32% year-on-year increase. Luxury home demand in India remained strong in 2025, with sales of Rs. 4 crore and above rising nearly 28% YoY across seven major cities. That downstream demand, all those new homes that need to be furnished, has to go somewhere.

Two: Quick commerce and horizontal e-commerce have built the behavioral bridge: Blinkit, Zepto, and Swiggy Instamart have done something revolutionary for the home category. They’ve trained millions of Indian consumers to buy home products: plates, storage containers, flower pots, kitchen accessories, impulsively for the first time.

Amazon and Flipkart have had home furnishing categories for years and have done the heavy lifting of proving consumers will buy décor and textiles online.

Again, just to reiterate, Amazon and Flipkart are fundamentally built for search, not discovery. They’re exceptional when a consumer knows what they want, say, “white cotton bedsheet, queen size, under ₹2,000.” They’re not built for when a consumer knows how they want to feel, say, “I want my bedroom to feel calm and Chettinadu-inspired.”

This is exactly what happened in fashion and beauty. Amazon has enormous fashion and beauty catalogues. But Myntra won fashion, and Nykaa won beauty because those categories required curation and brand storytelling that a horizontal marketplace structurally cannot provide. Home is following the same arc. The behavioral bridge has been built, now the category needs its destination.

Three: Generative AI has made the taste problem solvable: For years, the home category’s fundamental blocker was the visualization gap. Consumers couldn’t see how a product would look in their room, so they either didn’t buy or returned what they bought.

Generative AI has changed this. AI-powered tools can now let consumers visualize products in their own spaces, generate room designs from a mood board, and recommend complementary products based on what’s already in a room. AR-powered visualization will most likely reduce return rates.

The real constraint in home commerce was never supply; it was translating taste. Consumers know how they want to feel; they just don’t know how to assemble it. AI finally makes that bridge scalable.

If you remember the Vaaree Home Index stat quoted above, when consumers are presented with curated, room-level ‘Looks’ instead of individual product listings, conversion rates are 580% higher. The companies that deploy AI-powered discovery natively, not as a marketing feature but as the core product experience, will structurally break the return-rate ceiling that has held this category back.

This is perhaps the single most important shift. The taste problem was always the category’s defining challenge. For the first time, the technology to solve it actually exists.

Four: India’s manufacturing clusters have upgraded: PLI schemes and China+1 export demand have fundamentally improved the quality and scale of India’s home manufacturing base. IKEA itself increased its India sourcing from 30% to 50% in 2025, extending well beyond textiles into broader product categories.

The supply base is ready. What’s missing is the technology layer connecting these producers to Indian consumers efficiently.

Five: A generational shift in homeowners: Millions of young, digitally native, Instagram-influenced first-time homeowners are entering the market simultaneously. India’s urban population is expected to reach 600 million by 2031, up from 377 million in 2011. India’s housing demand is projected to reach 93 million units by 2036.

These consumers have a Pinterest board for their bedroom and have saved reels of themed living rooms. They don’t want to walk into a furniture store and choose from whatever’s on the floor. They want to discover a look, then buy the products that create it.

Six: Enabling technology for fragmented supply chains: This is perhaps the least glamorous driver, but one of the most critical. The home ecosystem runs on thousands of small suppliers and some noteworthy big players, each with different compliance requirements, catalogue formats, and fulfilment capabilities. Building for this market requires not just a consumer-facing product but serious back-end technology: supplier management platforms, compliance tools, inventory systems, and catalogue standardization.

Think of what Unicommerce became for the broader e-commerce ecosystem: the technology backbone that brands needed to sell across channels. The home category needs its own version of this: strong enabling technology that meets fragmented supplier needs on one side and powers consumer-facing experiences: shop the look, room visualization, personalized discovery, on the other. The tools to build this are now mature enough and affordable enough to make it viable.

Remove any one of these six forces, and the timing isn’t right. But right now, for the first time, all six are true at once.

Also, there’s a structural point worth making about what it costs to build a consumer platform today versus even five years ago. The infrastructure stack that a new home company inherits for free (or near-free) is extraordinary.

UPI processes billions of transactions at zero cost, logistics networks like Delhivery and Xpressbees offer plug-and-play fulfillment that would have required proprietary warehousing a decade ago. Cloud costs have cratered, social commerce and creator-led distribution have compressed customer acquisition costs relative to the TV-and-print playbook that the first wave relied on.

Given that generative AI is collapsing the cost of everything from product photography to catalogue management to customer service, the result is that a well-run D2C brand in 2025 can, according to our assessment (a quick back-of-the-envelope calculation) project ₹15-20 crore in first-year revenue on a ₹10-12 crore raise.

That ratio, revenue generated per rupee of capital deployed, would have been unthinkable for the companies that defined the first wave of Indian e-commerce. The cost of building a breakout consumer company hasn’t disappeared, but the cost of reaching the first proof points has dropped dramatically.

For a category like home, where the demand exists but the right model hasn’t been built yet, that changes the risk calculus for early-stage investors significantly.

Is the Home Ecosystem Underfunded? The Numbers Say ‘Yes’

India’s total startup funding in 2025 was $10.5 billion across 1,518 deals, a more selective year than 2024, with deal counts falling 39% even as India retained its position as the world’s third most-funded startup ecosystem. Even in this more disciplined environment, retail startups attracted $2.4 billion, fintech drew $2.2 billion, and enterprise applications secured $2.6 billion. Early-stage funding actually grew 7% year-on-year to $3.9 billion, suggesting investor conviction in scalable ventures remained strong.

Now compare: the entire home furnishing products sector in India has attracted $680 million in total venture funding across all time. In 2025, the sector saw just ~$75 million raised.

For context: in 2024, a single company, Zepto, raised $1.4 billion. That’s more than double what the entire home furnishing sector has raised in its lifetime.

A soon-to-be $80 billion market has attracted less total venture capital across its lifetime than what one quick commerce startup raised in a single year. That gap is striking.

A possible sentiment could well be, it didn’t work before, so it won’t work now, and it’s worth acknowledging that this instinct is understandable.

But it’s also worth noting that many of the categories we now celebrate as winners looked deeply problematic at similar stages. For example, Zepto is still burning significant capital, as is any other quick commerce company. The fundamental unit economics of quick commerce remain debated. And yet, nobody questions whether quick commerce is a real category because the behavioral shift is undeniable.

That’s the lens that matters here, not whether the current cohort of home companies has achieved profitability, but whether consumer behavior is shifting in a direction that makes a large outcome inevitable.

We believe the evidence, from rising home styling adoption to the move toward coordinated room-level purchasing, to the spread of digital home shopping beyond metros, says it is.

Where the Breakout Will Come From

Knowing where to look matters, and not all of the home value chain is equally positioned for a venture-scale outcome.

Furniture, as we’ve argued, is probably not the answer for venture. Interior design services are large but structurally hard to scale as they’re project-based, labor-intensive, and the value resides in the designer’s taste, which is difficult to productize.

Smart home and IoT is interesting enough to watch, but still comparatively early in India and largely dominated by established brands. So, we’d say we are excited but waiting for interesting ideas and plays to emerge.

The highest-potential space is a non-furniture home platform that combines three things: curated discovery powered by AI, a hybrid brand-plus-marketplace model with the right operating structure and deep supply chain integration with India’s manufacturing clusters.

The consumer starting point, “I’m moving into a new apartment and want to make it look like this”, is the highest-intent moment in the category. The company that captures that moment and guides the consumer through a full-stack experience, from visualization to purchase to delivery, will own the relationship. And if it runs the right operating model, branded where brand matters, marketplace where capital efficiency matters, it will have the margin structure to sustain it. The winner won’t be the company with the largest catalogue. It will be the company that becomes the default answer to “I’m doing up my home”, the one that makes the entire journey, from inspiration to installation, feel coherent.

And for the supply chain beyond brands, the vast world of unbranded and semi-branded products that Indian manufacturers produce to world-class standards, the opportunity is aggregation and exports. A technology-enabled platform that can organize fragmented supply, ensure quality, and connect these producers to both domestic marketplaces and global buyers would be building on real structural advantage.

India already manufactures world-class home products for the world. The missing layer is technology that organizes that supply and translates it into a coherent consumer experience.

Home is one of the last large consumer categories in India where behavior has shifted, but the platform hasn’t caught up yet. That gap rarely stays open for long.

For decades, India’s place in global technology was easy to describe. We were the services engine. The execution layer. The country that made software work at scale for the world.

That story is still true. But AI changes what “services” can become.

We are at crossroads today when it comes to AI - India is admittedly far behind US or China when it comes to fundamental research, foundational layer innovation and investment in research talent. We see reports of many founders building in AI moving to the US to be able to better access resources, talent and markets. The key question to ask is - while that is of course the headline opportunity, is that the ONLY opportunity?

We are entering a phase where India can move from being a services hub to becoming a global applied-AI powerhouse. Not because India will outspend the US or China in foundation model arms races, but because India has something structurally rare: a talent base that knows how to deploy technology in the real world, and a home market that forces solutions to handle complexity by default.

And that complexity is not a disadvantage. It is the edge.

India runs on constraints: many languages, diverse user needs, uneven infrastructure, high price sensitivity, and massive scale. If AI is going to create economic value, it will be through diffusion into workflows, into businesses, and into everyday life. In other words, through implementation.

This is where India’s opportunity sits.

Where we see the opening

The global AI conversation is still dominated by model size, compute concentration, and frontier capability.

India’s opening is different: AI that gets deployed.

The best articulation of this shift is visible in how the ecosystem is evolving. Google describes India’s AI startup story entering a decisive phase, moving “from prototypes to products” and “from early traction to sustainable businesses,” and reiterates a simple principle: if you solve for India, you build for the world.

That is the applied-AI thesis in one line.

Why India’s real advantage is complexity

Most countries build products for standardized environments. India is the opposite.

We live inside complexity every day:

1. language and cultural diversity

2. fragmented markets and supply chains

3. inconsistent infrastructure and last-mile delivery constraints

4. value-seeking customers who demand ROI fast

5. speed as a feature, not a luxury

If your product works in India, it has already been stress-tested under conditions that many markets never face.

That creates two advantages (the second one is well talked about everywhere, our focus in this article is on the first):

1. India becomes a domestic AI market that is massively underserved
   The problems are real. The scale is large. The willingness to adopt is high if ROI is visible.

2. India becomes an export engine for applied AI
   Many emerging markets resemble India more than they resemble Silicon Valley. Solve for India’s complexity and you build a template for the Global South.

Why now

This moment is not just about excitement. It is about the underlying economic pull.

1) The upside is macro, not incremental

NITI Aayog’s “AI for Viksit Bharat” estimates that accelerated adoption of AI across industries could add $500B–$600B over and above baseline GDP growth by 2035, driven by productivity and efficiency gains. It also highlights that financial services and manufacturing could see outsized impact, with meaningful portions of sectoral GDP attributable to AI by 2035. [Source]

That’s not a feature upgrade. That’s a national productivity wave.

2) Inclusion is now central to the AI agenda

NITI Aayog’s “Roadmap on AI for Inclusive Societal Development” frames AI and frontier tech as tools to break systemic barriers for India’s informal workforce, pointing to deeper barriers like lack of trust, poor access/usability, low awareness/skills, and outdated tools/processes. It proposes a national mission, Digital ShramSetu, to enable adoption at scale through partnerships and state-led implementation. [Source]

This matters because India’s AI opportunity is not only enterprise ROI. It is also mobility, skilling, and productivity uplift at the broad base of the economy.

3) The ecosystem is strengthening beyond “just startups”

Google’s January 2026 announcement emphasizes full-stack support to help startups bridge the hardest gap: moving from pilots to repeatable enterprise adoption. Access, trust, and go-to-market execution become decisive at this stage, which is exactly where India’s services DNA can be converted into product scale. [Source]

4) But the missing pieces are real

Carnegie’s analysis is blunt: if India wants to meet its ambition of becoming a global AI leader, it must urgently close gaps in talent, data, and R&D. The paper argues that compute and Indic models alone are not enough, and that without strengthening these enabling layers, India may fall short of its stated goals. [Source]

This is the core tension of the moment: India has momentum, but leadership requires fixing foundations.

The “services as software” unlock for India-scale challenges

The default playbook for decades was: build software, sell it to global markets. SaaS largely worked because the largest buyers sat outside India.

AI shifts the nature of software itself.

You can now deliver outcomes as a managed system, where AI becomes the execution layer inside workflows. This is the real bridge from services to product. It lets India do what it does best, deploy technology in messy real-world environments, and turn it into repeatable, scalable platforms.

One category has already started to prove this with reasonable scale in India: voice AI.

Voice works because it fits India:

• multi-lingual, multi-accent environments

• user cohorts that prefer speech over typing

• massive cost pressure on customer support and outreach

• immediate ROI through automation + better conversion

Voice is not the whole story. It is a preview.

The bigger opportunity is applying this “service as software” approach across sectors.

The India-scale AI opportunity areas

Below are arenas where applied AI can unlock multi-billion-dollar domestic markets and also build exportable playbooks.

1) Commerce: hyper-personalization for physical and virtual products

India’s commerce is fragmented and diverse. AI can become the layer that makes this tractable:

• hyper-local merchandising and discovery

• automated catalog + localization

• demand forecasting under noisy signals

• assisted buying experiences that drive conversion

The real value is not “recommendations.” It is conversion + ROI under constraints.

2) Entertainment: India as a consumer AI frontier

India is structurally built for consumer AI adoption: mobile-first, creator-led, and entertainment-heavy.
Opportunities include:

• AI-native content workflows (editing, localization, dubbing, summarization)

• gaming development acceleration and personalized content loops

• consumer agents built for Indian behaviors and languages

If India builds consumer AI that works at low cost and high diversity, it becomes exportable.

3) Healthcare: access under scarcity

India’s constraint is structural capacity. AI can multiply capacity through:

• triage and routing

• doctor workflow automation

• follow-up and adherence

• decision support in resource-constrained settings

This is not “AI replacing doctors.” It is AI increasing the reach of care delivery.

4) Edge computing: SLMs and local hardware as a feature

India will not always have abundant compute at the edge. This makes small models and hybrid deployment strategies strategically important:

• privacy-preserving inference

• offline/low-connectivity environments

• low-latency decisions in field operations

Edge-first AI is not niche. It is how India scales AI outside metro bubbles.

5) Skilling: adaptive coaching as national infrastructure

If inclusion is part of the thesis, skilling must be central. NITI’s inclusion roadmap explicitly highlights barriers like low awareness, skills, and outdated tools, and argues for systemic interventions at scale.

AI can enable:

• personalized learning paths

• coaching loops tied to outcomes

• simulated practice environments

• scalable, low-cost upskilling for the workforce

Skilling becomes the bridge between AI productivity and broad-based economic gains.

6) Fraud prevention and cybersecurity: trust as the economy’s foundation

As India’s digital rails deepen, fraud scales with them. AI can power:

• anomaly detection

• identity and risk scoring

• real-time monitoring for platforms and enterprises

• synthetic fraud detection

Trust is not a compliance category. It is a moat.

7) Robotics: pragmatic automation for productivity

Robotics in India should be approached as “ROI robotics,” not science experiments:

• healthcare support and assistive workflows

• logistics and warehousing automation

• manufacturing quality and throughput support

The wedge is narrow, measurable, and deployment-led.

What this means for investors in India

For a long time, the SaaS playbook was clear and rational: if customers were primarily in the US, companies often needed to be domiciled there too.

Right now, AI is largely following that same playbook.

But domestic conditions are changing. Capital is flowing into India’s VC ecosystem, and India-domiciled investors face a strategic question: do you stay out, do you find ways to invest outside India, or do you build conviction and find ways to invest inside India?

The applied-AI view suggests that “inside India” is not a compromise. It can be the highest-leverage path because:

• India forces defensible product building under constraints

• distribution is a core capability

• domestic scale can fund iteration

• export markets resemble India more than Silicon Valley
  

Why the bet is still worth making

Because the ingredients are aligning: macroeconomic pull (productivity), ecosystem support for scaling, and India’s structural advantage in deploying technology at scale.
The question is whether we can close the missing pieces fast enough.

If we do, India’s AI story won’t be “services upgraded by AI.”
It will be “India became the world’s proving ground for applied AI, then exported that playbook globally.”

Subscribe now

Every other VC deck has a China+1 slide (ours does, too) you know the one that talks about supply chains diversifying, PLI incentives coming in. It concludes by saying that India’s manufacturing moment has arrived, finally.

The narrative is directionally right. We want to just go a bit deeper, or at least our interpretation of it.

Thanks for reading! Subscribe for free to receive new posts and support my work.

We think that India won’t win by being a ‘cheaper China’ or a more ‘convenient Vietnam’. Vietnam has ~17 FTAs and $25 billion in annual FDI, while China has $4.66 trillion in manufacturing output and ecosystem depth built over four decades.

But, there’s a different game emerging, one where India has real advantages.

This is the ‘High-mix, low-volume manufacturing’ game. Think of it as customization at scale. These are products where managing variability matters more than minimizing cost. We call this ‘Complexity+1’.

Why do we believe this is India’s moment?

We see five things converging: the digital infrastructure for coordination finally exists (GST, UPI, smartphones on shop floors). Global supply chains are diversifying by necessity, not choice. Wage economics are shifting as Chinese labor costs rise. AI has crossed a capability threshold, and India has both the engineering talent to build these tools and the fragmented manufacturing base that desperately needs them. And critically: solutions proven in India’s chaotic environment can be exported to emerging manufacturing hubs worldwide.

And, we strongly believe that this is more than just about India capturing manufacturing share; it’s about building the software and AI infrastructure layer for complexity manufacturing globally, positioning India at the center of physical AI and the next industrial shift.

This piece lays out our investment thesis: where we see the opportunity, why the timing is right, and the paths available for founders building in this space.

Why Complexity? Why Now?

The old model was simple: design in the West, manufacture millions of identical units in China, ship globally. However, that model is now breaking in two places. Buyers want variants, not commodities: shorter cycles, more customization, and smaller batches. And companies now want diversification. Geopolitics has made single-country supply chains a liability. We believe that the opportunity is at the intersection: manufacturers who can handle complexity, located outside China.

This shift creates demand for manufacturers who can handle variability profitably. Not every country can serve this as it requires engineering depth, operational sophistication, and the ability to iterate quickly.

India has the raw ingredients. The keyword being ‘raw’.

• Over 1 million engineers graduate annually

• 1,700+ Global Capability Centers with deep product development expertise

• A $2.1 trillion domestic market that can absorb production runs and fund iteration

• Proven champions like Sona Comstar, Motherson, Dixon, who have shown complexity manufacturing works here

But here’s the honest question: those champions took decades to build. What’s the path for a new founder today?

We see two.

Path 1: The Coordination Layer

India’s manufacturing base is fragmented. Thousands of small suppliers, inconsistent quality, unreliable delivery. For decades, this has been framed as a weakness.

We see it as an asset ‘if’ you can coordinate it.

The playbook is that you don’t own factories, you aggregate them. Add quality systems, demand routing, and the software layer that turns fragmented supply into reliable delivery.

Why this works now (and didn’t before):

• GST (2017) unified interstate commerce: Before this, moving goods across state lines was a nightmare of checkpoints and paperwork. Now it’s seamless.

• Digital payments work at the factory level: UPI processes 16+ billion transactions monthly. You can pay a small supplier, say, in Tamil Nadu, instantly. Working capital cycles have compressed.

• Smartphones on every shop floor: Real-time coordination, quality documentation, delivery tracking, all of it is possible now through mobile-first software.

• AI makes coordination tractable: This is the unlock that didn’t exist even two years ago. When you have thousands of suppliers, matching orders to the right one, based on capability, capacity, quality history, location, was a human-intensive nightmare. Now, it’s a software problem. Demand forecasting across high-mix SKUs, quality prediction before shipment, anomaly detection in supplier performance, these are exactly the problems AI excels at. Coordination that required armies of people can now be software-first.

The founder playbook:

If we had to put the goal for path 1 in one line, it would be: build the orchestration layer. The play here is to connect the many fragmented suppliers to consolidated demand. Use software to quality-assure, route, and optimize. Capture the coordination premium that small suppliers can’t capture alone.

This is what we see in Fabriclore (thousands of fabric suppliers aggregated for fashion brands) and Tyreplex (aftermarket tyres with complex fitment specs).

These aren’t ‘manufacturing’ companies in the traditional sense. They’re coordination companies, using technology, building demand and distribution, that make India’s fragmented base accessible to buyers who need reliability.

Path 2: Start Simple, Build Into Complexity

The second path is actual manufacturing, but with a specific approach.

Sona Comstar and Motherson were built vertically over 40 years. That’s not a playbook a new founder can follow. Just build ‘world-class manufacturing’ isn’t actionable advice.

But there’s a different approach: start simple, prove capability, and then expand scope.

Begin with products that aren’t technically complex but have other forms of complexity—high SKU counts, seasonal variability, compliance requirements, or customization needs. Build operational muscle on these and then move up the curve.

Why this path is viable now:

• Domestic demand funds the learning curve: India’s consumption market lets you reach meaningful scale without export contracts. You can iterate, make mistakes, and improve without losing international customers.

• ‘Proven in India’ increasingly matters: Apple’s Indian production reaching 17-18% of global iPhone supply signals that quality certification in India is achievable. The ‘reputational’ barrier is lowering.

• Rising wages elsewhere shift economics: Chinese manufacturing wages have reached $6-8/hour. The gap with India has widened. Products that didn’t make sense to manufacture here five years ago now do.

• AI compresses the quality learning curve: This one is underappreciated. India’s quality consistency problem has historically required years of workforce training and process discipline. AI-powered visual inspection catches defects that human inspectors miss, and scales instantly. Process optimization algorithms identify patterns in production data to reduce waste and predict failures. A new manufacturer can now deploy quality systems that previously required decades of accumulated expertise.

• PLI creates runway: Production-linked incentives in 14 sectors provide margin buffer while you’re building capability. While it is not obviously a permanent subsidy, it provides a runway to reach competitiveness.

The founder playbook:

Don’t try to be Motherson on day one. Start with products where you can win: high-mix categories that larger players find too fragmented to focus on. Build quality systems, train workforce, and refine processes. Use domestic demand to scale and then expand into more complex products as capability compounds.

The key: revenue from the start, complexity over time. Not years of capability-building before the first sale.

Bidso is building this way—starting in toys(high SKU complexity, seasonal demand, safety compliance), with a path toward ‘second-order’ manufacturing as operational capability grows. The ambition isn’t to stay in simple products forever, it is rather to build the muscle that enables expanding up the complexity (including design and IP) curve over time.

What Makes India Different?

We think fragmentation is a feature, not a bug, for the right business model.

China’s manufacturing base is consolidated. That’s an advantage if you are a global OEM wanting a single supplier. However, it is a disadvantage if you are building a coordination platform as (well) there’s nothing to coordinate.

India has 63 million MSMEs, most of them small and underserved. That’s raw material for coordination businesses.

• Domestic demand provides a proving ground: Vietnam is 86% export-dependent. If export orders dry up, the economy shudders (as it did during COVID).

India’s domestic market, $2.1 trillion today, projected $4.3 trillion by 2030, provides a foundation. You can build scale domestically while refining quality, then export when ready. The domestic market is a training ground.

• Engineering talent exists, scattered but available: India doesn’t have a shortage of engineers, but it clearly has a deployment problem. Most engineering talent flows to services and software. But that’s a reallocation opportunity, not an ‘absence’. Yes, they need the right skillsets and companies that create compelling paths for engineers into manufacturing, whether through manufacturing-tech software or factory-floor roles, can tap a resource that Vietnam and Bangladesh simply don’t have. Vietnam, for example, has a 77% engineering talent shortage.

• Digital infrastructure actually works: This is easy to overlook because it’s fairly recent. GST, UPI, Aadhaar, India Stack, these create the rails for coordination at scale. You can verify suppliers, process payments, track shipments, and manage compliance in ways that were impossible a decade ago.

• The AI advantage is structural: India has something rare: the combination of engineering talent to build AI tools and a fragmented manufacturing base that needs them. China has manufacturing depth but faces AI development constraints from chip restrictions and geopolitics. Vietnam has neither the engineering base nor the manufacturing sophistication. The US and Europe have AI talent and advanced manufacturing, but their factories are already highly automated, with less need for the coordination and quality tools that India’s fragmented base demands.

India’s 1,700+ GCCs aren’t just doing IT services, they have actively started building AI/ML capabilities for global companies. That expertise can be redirected toward manufacturing problems: visual inspection, demand forecasting, process optimization, supplier matching, among others. The software layer that makes coordination and quality control tractable can be built here, for here.

And if it works here, it can work everywhere.

India’s fragmented base is a proving ground. AI tools that solve coordination and quality problems in India’s chaotic manufacturing environment are battle-tested for the hardest conditions. That same software can be deployed across emerging manufacturing hubs globally—Southeast Asia, Africa, Latin America, the Middle East—markets that are growing but not yet highly automated. The opportunity is to build the manufacturing-tech stack for the developing world.

What We’re Looking For

Founders who see one of three opportunities:

Coordination founders:

• Identify a fragmented category with complexity (SKUs, compliance, variability)

• Build the software and systems layer that aggregates supply

• Create reliability where it doesn’t currently exist

• Capture the coordination premium

Manufacturing founders:

• Start with products where you can win today: not commodities, not rocket science

• Build operational capability: quality systems, workforce training, process discipline

• Use domestic demand to fund learning and reach scale

• Have a credible path to expand into more complex products over time

Manufacturing-tech founders:

• Build AI-powered tools that solve specific manufacturing problems: visual inspection, demand forecasting, process optimization, design-to-manufacturing translation

• Target India’s fragmented base as the deployment ground—thousands of small manufacturers who can’t build this themselves

• Create the software layer that accelerates quality and capability building across the ecosystem

• Crucially: tools that can be deployed and generate revenue now, not R&D projects that need five years of laboratory work before commercial viability

In all cases, we are looking for founders who understand that this is a fairly long game (without it becoming a science project). Capability compounds, customer relationships deepen and the moat widens, but over years, not quarters.

That said, we are realists about fund cycles. We need to see a path to revenue within a reasonable timeframe, companies that can build toward complexity while generating cash along the way, not pure R&D bets that require five years of development before the first customer.

The Honest Version

We’re not claiming inevitability. This is a directional bet.

What has to go right:

• Founders execute on operations, not just fundraising

• Quality discipline becomes cultural, not cosmetic

• Engineering talent finds manufacturing and manufacturing-tech roles compelling

• AI/software tools actually get deployed on shop floors, not just sold

• Global buyers trust “Made in India” for more than just cost

What could go wrong:

• Other countries (Vietnam and Indonesia, for example) build capability faster than expected

• Domestic demand stays price-sensitive rather than quality-seeking

• Coordination businesses get disintermediated by platforms with deeper pockets

• Policy shifts away from manufacturing focus

Why we’re betting anyway:

The ingredients are in place in a way they weren’t before. GST and digital rails make coordination possible, AI tools compress the quality and capability learning curve, and rising global wages make manufacturing economics work. Also, supply chain diversification creates demand and domestic consumption provides scale. Finally, we have the engineering talent that can build the software layer.

The question is whether founders can convert these conditions into durable businesses.

We think some will. That’s the bet.

Love to hear your thoughts; do write in to ankur@peercapital.in

The finding that has caught attention is that peak performance is negatively correlated with early performance. The people who dominated as kids, the prodigies who specialized early and climbed fast, were consistently outperformed in adulthood by those who sampled widely, learned slowly, and specialized late.

Psychologist Robin Hogarth offered a framework that helps explain why. He distinguishes between ‘kind’ learning environments and ‘wicked’ ones. Kind environments have clear rules, tight feedback loops, and patterns that repeat. Say, chess is kind, so is golf, specialization works because the game stays the same.

‘Wicked’ environments, on the other hand, are different. The rules are unclear or shifting, feedback is delayed or noisy, or sometimes even absent. What worked yesterday may not work tomorrow.

In wicked environments, the specialist’s playbook becomes, for the lack of a better word, a ‘liability’. The advantage shifts to people who can reason by analogy, pull from unexpected domains, and adapt when the ground moves.

David Epstein built on this framework in his book ‘Range’, arguing that generalists thrive precisely because they have learned to navigate uncertainty rather than optimize for stable rules.

Early-stage startups are ‘wicked’ environments. This is well understood, but here’s what we have been thinking about: India itself is a ‘wicked’ environment for building companies. We mean this in the nicest way possible.

There’s India, and there’s ‘Bharat’; infrastructure that works brilliantly in one city might fail in another. Customer segments are so diverse, distribution economics defy the playbooks written in Palo Alto. Nothing works the ‘western’ way. The pattern-matching that got you from zero to one probably won’t get you from one to ten.

For years, this was framed as India’s handicap, the reason why building here was harder, and why unit economics took longer to prove.

We think that framing is backwards.

What if India’s ‘wickedness’ is precisely what produces founders capable of navigating other wicked environments? What if two decades of solving for chaos, building payments infrastructure around cash-dependent economies, creating logistics networks across fragmented geographies, selling software to enterprises with procurement processes that make no sense, what if all of that produced a generation of founders with an unusual skill: the ability to reason across broken contexts?

This brings us to AI.

AI is the most ‘wicked’ environment we have seen in technology.

The capabilities shift quarterly. In fact, today’s state-of-the-art is next quarter’s baseline and moats that seemed defensible evaporate. Business models that worked six months ago need reinvention. In short, there is no playbook.

The founders who will win are not the ones with the deepest narrow expertise in a single vertical; they are the ones who can see patterns across domains, adapt when assumptions break, and build for a world that doesn’t exist yet.

The Science study found that world-class performers engaged in more multidisciplinary practice and showed more gradual early progress. They sampled before they specialized, they developed what Epstein calls “range.”

For twenty years, the services economy sent Indian engineers into the complexity of global enterprises: banking systems, healthcare compliance, retail logistics, and manufacturing operations. It wasn’t ‘glamorous’, it was often dismissed as ‘not real product work.’

But something was happening beneath the surface: an entire generation was developing involuntary range. They learned to solve problems across contexts that had nothing in common except their messiness.

Now those people, and the people they trained, are starting companies. And they are starting them in an era when AI is amplifying the generalist advantage. The bottleneck is no longer “can you build sophisticated ML infrastructure?” The bottleneck is “can you see which problems are worth solving, draw the right analogies, and adapt when the technology shifts under your feet?”

We truly believe that ‘Range’ founders can.

Solve for India’s complexity, and you develop muscles that transfer. Solve for AI’s uncertainty, and you need those same muscles.

The ‘wicked’ edge isn’t a handicap; it’s actually the point we want to drive home.

And, there are signs that the ‘wicked’ system is leading to good outcomes or, at the bare minimum, showing signs of potential.

The Numbers Story

Something significant happened this year. Indian startup ecosystem did not just survive the funding recalibration, they arrived on the public markets with conviction.

13 companies went public in 2024, raising over ₹29,000 crore collectively. That momentum carried into 2025, with 100+ IPOs hitting the markets. Notable ones include Meesho, Ather Energy, Urban Company, Groww and BlueStone Jewellery. While the stocks will have their ups and downs in the public market, there is no denying the momentum. The deeper story is what this signals about ecosystem maturity.

Bengaluru alone produced 32 unicorns between 2020 and 2024. Perhaps most tellingly, the average time taken for startups to attain unicorn status had fallen from 8.4 years in 2022 to only 5.5 years in 2023. With AI, this will only decrease further. We are already seeing this in the global ecosystem.

The flywheel—founders becoming angels, executives becoming operators, capital recycling into new ventures—is turning faster than ever.

The broader economy provides a tailwind. India is now the world’s fourth-largest economy, growing at 6.5% while peers stumble through uncertain terrain. The path to the third-largest by 2030 looks increasingly inevitable, with projections pointing to a $7.3 trillion GDP by 2030. When the IMF calls you the fastest-growing major economy amid global headwinds, something structural is at work.

Two Themes We Are Leaning Into

Vertical AI: Where the Real Value Gets Built

The AI conversation has matured considerably since the ChatGPT moment of late 2022.

Two years in, the froth is clearing. Horizontal tools that promised to transform everything have started running into a stubborn reality: enterprises want outcomes, not experiments.

They’re shifting budgets away from one-size-fits-all models and toward vertical AI systems built to the specific demands of their industries. Global spending on domain-specific AI topped $5 billion last year, and at a sustained compound annual growth rate exceeding 40%, it’s on track to reach nearly $47 billion by 2030.

This is where India has an edge worth betting on. Two decades of building for global enterprises created a deep bench of engineers who understand domain complexity intimately, not just the technology layer, but the workflows, the edge cases, and the reasons why generic solutions fail. The winners will be companies that combine deep domain expertise with AI capability, not those that bolt one onto the other as an afterthought.

Consumer Enablers: The Infrastructure Layer

India’s consumption story doesn’t need retelling. But behind every consumer company scaling to millions of users, there’s a layer of enabling infrastructure: logistics networks, commerce platforms, supply chain tech, fintech rails, that’s equally investable and often more defensible.

These aren’t ‘glamorous’ categories that generate breathless headlines, but they compound quickly, building moats that are difficult to replicate once established.

Our Portfolio in 2025

We made three new investments this year and doubled down on one. When we look at these teams, we see the pattern we described earlier, not because we set out to find it, but because it emerged from the work.

Game State Labs: We co-led a $2 million seed round alongside Neon Fund. Aashbir Bhatia, Ashwin Ramakrishnan, and Jagveer Gandhi are building the data infrastructure layer for gaming studios, enabling them to understand player behaviour with a depth that existing analytics tools simply can’t match.

What struck us about this team wasn’t ‘gaming pedigree’, it was the diversity of contexts they had navigated before converging on this problem. Ashwin spent time at gaming giant EA, but his career spans Jio, Telstra, and Airtel. Aashbir came through strategy consulting at Kearney and Arthur D. Little. Jagveer’s path is the most winding: autonomous vehicles at Drive.ai, imaging software at ImageKit, teen banking at Akudo, and fintech at Uni Cards.

They’re not gaming lifers who decided to add AI. They’re pattern-matchers who recognized that the data infrastructure problem in gaming looks a lot like problems they had seen, and solved, in other industries.

Redacto: We led a $1.4 million seed round alongside Antler India. As India prepares to implement the Digital Personal Data Protection Act, Redacto is building the AI-powered privacy and governance platform that enterprises will need. Amit Kumar spent 20 years building at Microsoft, Google, GRAB, and Ola before founding a BNPL platform. Shashank Karincheti brings deep cybersecurity experience from his time as Deputy CISO at Razorpay. They’re people who’ve seen data and privacy problems from enough angles to know what enterprises actually need.

TyrePlex: We led a $2.4 million round for this B2B tyre marketplace. Puneet Bhaskar, Rupendra Pratap Singh, Jiveshwar Sharma, and Nikhil Kalra have built a platform that’s transforming an unstructured market: 90% of India’s tyre dealers are multi-brand outlets operating without modern tools. Puneet’s path to tyres ran through textiles, fintech, edtech, wellness, fashion e-commerce, and used cars. TyrePlex gives dealers procurement efficiency, inventory management, and data-driven decision-making. They have already achieved impressive revenue growth since 2022, and the $13 billion Indian tyre market (growing to ~$30 billion by 2030) represents a massive opportunity.

Vaaree: We led a $4.6 million pre-Series A round, doubling down on our initial seed investment. Garima Luthra and Varun Vohra are building the rails for how India will shop for home furnishings in the future. The founding team includes a serial entrepreneur who built and sold companies in tech services and healthcare, and an operator who worked across quick commerce, insurtech, and fintech.

They are not home decor people, they are people who have seen how consumer businesses work across enough contexts to spot what’s broken in this one. Their AI-powered platform VibeCheck is changing how consumers discover and style their homes, while their factory-direct model maintains industry-low return rates below 5%.

Other Portfolio Highlights

Beyond this year’s investments, several portfolio companies hit meaningful milestones:

• Stupa, an AI-powered sports technology platform for racquet sports, signed multi-year SaaS contracts with Table Tennis Australia, Swiss Tennis & Padel, and Czech Table Tennis. What started as tournament management software is becoming the infrastructure layer for federations globally. Stupa has also launched FrameX, a replay-based DRS built on high-speed cameras, which will be deployed for the Prime Volleyball League. It has enhanced its Badminton DRS with automated line calling for serves.

• Webless, an AI-powered discovery engine that converts B2B website visitors into buyers, achieved SOC II compliance, a threshold that matters for enterprise sales, and converted all design partnerships into paying customers. The product now supports multi-modal search and multiple languages.

• Bidso, a design-led original design manufacturer (ODM) for outdoor toys and play equipment, more than doubled its locked-in manufacturing capacity. New export agreements validate the thesis: IP-led manufacturing with a software layer can compete globally, not just on cost.

• Fabriclore, a platform on a mission to become the operating system for fabric procurement, grew 4X since we invested. Repeat rates nearly tripled while achieving positive contribution margins. Custom fabric at low minimum order quantities is becoming a habit, not a one-time experiment.

• Flash, an AI-powered personal shopping assistant, expanded globally to target premium shoppers across the US and India. They have launched Flash AI to offer end-to-end agentic e-commerce capabilities to premium power shoppers. They have created a novel feature that allows any product URL to be appended to flash.co to generate a detailed product review, leveraging insights from YouTube, Reddit, Blogs, etc.

What We Can Add

A question we ask ourselves: what can we possibly offer founders who’ve navigated five industries before landing on gaming, or built and sold two companies before touching home decor, or are already making a ton of progress?

A line from Christopher Clarey’s biography of Roger Federer, The Master, has stayed with us. Severin Lüthi, Federer’s long-time coach, was often asked what he could possibly tell the greatest (arguably) tennis player in history. His answer: “He wants to hear it again, even if it’s a small thing like ‘stay lower to the ball’. Because he says sometimes, I also forget things, or I just don’t realize.”

Range founders don’t need investors who pretend to know their industries better than they do. They need partners who notice when they’re standing too tall, who can say the obvious thing at the right moment, precisely because they’re watching from outside.

That’s the role we try to play. We are not the experts, we are just keen observers. We are sometimes just the voice that reminds you to stay lower to the ball.

Advanced Manufacturing: A Thesis Takes Shape

We spent much of 2025 deepening our thinking on advanced manufacturing. The opportunity is structural: global customers actively seeking non-China sources, government incentives at historic levels, and a generation of founders finally choosing atoms over bits.

Our view remains consistent. The companies that will matter aren’t contract assemblers competing on labour cost arbitrage. They are design-led, IP-rich businesses that happen to manufacture in India, companies that own what they create rather than rent out their workforce. This includes manufacturing infrastructure platforms that create tech layers on top of fragmented production capacity.

Our portfolio company Bidso, IP-led tech-enabled ODM aggregating capacity with a software layer, represents this thesis in action. More will follow as we find the right teams attacking the right problems.

The Headwinds Are Real

We’d be doing ourselves a disservice if we pretended the path ahead is smooth.

The geopolitical environment has grown more complex. Trade corridors are being redrawn as the global order reshuffles. For India, this cuts both ways: opportunity in diversification, risk in unpredictability. Companies with strong domestic demand and exposure to multiple markets are better positioned to weather these shifts.

The AI landscape still has froth in the system. Not every AI startup has a defensible moat, and not every enterprise buyer is ready to deploy every other tool just because it seems exciting. The gap between technical capability and real business outcomes remains wide. Founders who can bridge it thoughtfully will win. Those building features disguised as companies will not.

These are real considerations, but they exist alongside, not instead of, the structural tailwinds we’ve described.

Looking Ahead

India at 100 isn’t some random slogan. This country is building something that would have seemed improbable a decade ago. The exits are happening, the flywheel is turning, and capital is being recycled by operators who understand what it takes to build.

We’re genuinely excited about 2026.

Here are a few predictions based on our worldview and our bets:

• AI-powered consumer enablers will move as quickly as enterprise AI, if not faster: Most predictions for 2026 focus on agentic AI transforming enterprise workflows. We think the enabling infrastructure behind how people shop, discover, and make decisions will surprise everyone with how quickly it embeds into daily life. The interfaces are intuitive, and consumer behaviour shifts faster than procurement cycles.

• AI regulation will create categories, not just compliance burdens: As AI becomes central to business operations, regulation is catching up: EU AI Act, a patchwork of US state laws; we will see more laws across countries. Every wave of regulation creates infrastructure opportunities. GDPR spawned an entire category of privacy tech companies. AI governance will be larger, and the companies building the picks and shovels of compliance, privacy, and security will have a long runway.

• The most durable companies of this cycle will solve hard, unsexy problems where complexity is the moat: Fabric procurement, tyre dealer networks, sports federation software, contract manufacturing with design IP, these categories don’t make front-page headlines, but they compound and build defensibility that’s difficult to replicate.

Fragmented, offline-first markets aren’t uniquely Indian: auto parts, building materials, or industrial supplies are messy everywhere. But India forces you to build for low margins, long-tail customers, and chaos from day one. Solutions that work here can travel.

We are particularly drawn to these structural problems, not because they are ‘AI-first,’ but because the problem itself is hard enough to be a moat. AI will be used to solve pieces of it, but these aren’t enterprise data plays where the model is the product. The complexity is the product, and that’s what gives them shelf life.

India’s manufacturing story fits the same logic. The winners won’t be contract assemblers competing on labour arbitrage, they will be design-led, IP-rich businesses that own what they create, companies where the hard problem is the point.

• India’s deep tech moment will move from promise to production: For a long time, India’s deep tech ambition has felt like something always just over the horizon. That is beginning to change with government incentives, targeted funds, and mission-driven programs. They are starting to translate into real capacity being built on the ground.

What matters is not the size of any single fund or announcement, but the signal they collectively send. Capital is being aligned with relatively long-cycle, technically demanding sectors: semiconductors, advanced manufacturing, space, energy systems, in a way that reduces early risk for founders willing to work through uncertainty. This won’t create overnight winners, and it shouldn’t. Deep tech compounds slowly, then all at once.

India’s advantage here is familiar: founders who have learned to operate under constraints, iterate without perfect information, and make progress without relying on abundant capital. As incentives lower the cost of the first few steps, we expect more teams to take the plunge. The result will be a steady emergence of design-led, IP-owning companies that use India as a capability base rather than a cost base.

• Consumerization will deepen, and Bharat will start showing real output, not just intent: Consumerization in India is often discussed as a story of aspiration. What we’re beginning to see is the next phase: aspiration converting into output at scale. The expansion into Bharat is no longer just about access or awareness; it’s about systems: logistics, payments, distribution, manufacturing, and services, finally reaching the threshold where demand can be met consistently.

The last decade digitized attention, the next one will monetize it. As consumer behaviors diffuse across geographies, the real winners will be companies that build for fragmented realities from day one: low margins, inconsistent infrastructure, diverse preferences, and long-tail demand. This applies equally to bits and atoms, services and goods.

What’s changing now is velocity. Consumer platforms are getting better at coordinating complexity, and that coordination is unlocking tangible output: more products shipped, more services delivered, more repeat usage across markets that were previously hard to serve. Solutions built for India’s messiness are starting to travel because they work on-ground.

This is not convergence toward a Western consumer model. It’s something more interesting: a distinctly Indian consumer system taking shape, one that rewards founders who can operate across broken contexts and still deliver reliable outcomes. And, like we keep harping often, this can and will be scaled to other similar geos.

None of these trends will operate in isolation. They will interact, compound, and create second-order effects that will matter more than any single shift on its own.

Will we be right about all of it? Probably not. Markets humble everyone eventually.

But that’s precisely why we back founders with range, people who aren’t optimizing for a single outcome, but building the capacity to adapt when the world shifts.

The ‘wicked’ edge compounds, and we are here for the long game. Here’s to 2026!

• Ankur, Karthik, Rohit

The year was 2014, and the mission cost $74 million.

In 2013, Hollywood spent $100 million making Gravity, a movie about space.

The same month that Mangalyaan reached Mars, so did NASA’s MAVEN orbiter (in fact, MAVEN reached 2 days before), at nine times the cost. It is no secret that the ISRO engineers didn’t have NASA’s resources.

ISRO couldn’t afford expensive redundancy or multiple test cycles. They had to be right the first time, with less. The constraint forced a different kind of engineering: frugal but not cheap, constrained yet innovative, and pragmatic yet audacious. Indeed, a delicate, tightrope-kind balance.

The same DNA is now emerging in manufacturing. Indian companies are beginning to compete on ingenuity, designing and engineering products the world wants to buy, with cost structures that create margin rather than just volume.

For years, India’s relationship with global manufacturing followed a predictable script (with obviously a few well-known exceptions): multinational corporations designed products in California, Munich, or Tokyo. They then manufactured those at scale in Shenzhen. And they offshored the back-office work: call centers, software maintenance, data processing, to Bangalore, Hyderabad, Chennai, Mumbai or NCR. Remember, these jobs were ‘Bangalored’? (Offensive to say the least.)

However, the truth is that India captured value at the edges. The intellectual property, the manufacturing know-how, the brand equity, et al, were accrued elsewhere.

This arrangement created a $200 billion IT services industry and several globally competitive companies. It also created a mental model that India was somehow unsuited for manufacturing. There was just too much bureaucracy, too little infrastructure, not to mention the many languages, states, and other complications.

That mental model is, however, changing now. Yes, a lot more (a whole lot more) needs to be done, but there is a feel of change in the air. There is the intoxication of possibility.

We are witnessing a shift in where value gets created. The companies that will define Indian manufacturing over the next two decades won’t be contract assemblers competing on labor cost. They will be design-led, IP-rich businesses that happen to manufacture in India. Companies that own what they create rather than rent out their labor.

At Peer Capital, this is the bet we’re making.

Three Shifts That Changed the Math

Every few years, someone declares that India’s manufacturing moment has arrived. The declarations have been premature often enough that skepticism is warranted. So why do we believe this time is structurally different?

Three shifts have converged simultaneously. Unlike previous false starts, they reinforce each other rather than depending on government will alone.

Global supply chains are actively diversifying away from China.

This is no longer a talking point; it really is a procurement directive.

It snowballed when COVID lockdowns in China shut down factories for weeks, because geopolitical tensions made single-country dependency a board-level risk, and because operations teams concluded that concentration in China was no longer acceptable.

Every multinational with significant manufacturing exposure is having the same conversation.

India has emerged as one of the top beneficiaries. In April 2025, India surpassed China in monthly iPhone exports to the United States for the first time. Tim Cook told analysts in May 2025 that “the majority of iPhones sold in the US will have India as their country of origin”.

Samsung now manufactures its flagship Galaxy S24 and S25 series at its Noida plant and exports to over 70 countries. The company is now exploring shifting additional production from Vietnam as tariff pressures mount. Google began assembling Pixel phones in India in 2024 and plans to move the majority of its smartphone manufacturing here within three years.

In laptops, Dell, HP, Lenovo, Asus, and Acer have all signed on to produce locally under the PLI scheme, 27 companies in total, responding to import restrictions that made local assembly the path of least resistance.

The pattern is consistent: companies that came to India for its consumer market are now discovering it works as an export base too.

China+1 created optionality, but it doesn’t explain why India is winning a disproportionate share.

Three capability shifts matter here:

First, quality has reached parity. Indian manufacturers are now meeting the same specifications, tolerances, and audit requirements that global OEMs demand from world-class suppliers.

Second, India offers engineering integration, not just assembly. Most alternative manufacturing destinations provide labour and logistics. India provides engineers who understand the product, who can iterate on designs, identify production efficiencies, and solve problems without escalating to headquarters. For OEMs, this collapses cycle times and reduces coordination costs.

Third, the supplier ecosystem has reached critical mass. When one anchor OEM succeeds in India, it pulls component suppliers, tooling vendors, and testing infrastructure along with it. Each subsequent entrant finds the ecosystem more capable than the last.

This is what changes the math: India is no longer a fallback option for de-risking. It’s becoming a capability-driven choice.

Indian policy has reached critical mass.

The Production-Linked Incentive schemes now cover 14 sectors with ₹1.97 lakh crore in committed incentives.

The India Semiconductor Mission has ₹76,000 crore deployed across approved projects. Results are materializing: ₹1.76 lakh crore in realized investments by March 2025, over 12 lakh direct and indirect jobs created, and ₹16.5 lakh crore in production output.

Indian engineering talent is turning toward hardware.

For two decades, the best IIT graduates went into software. First to services companies like Infosys and TCS. Then to product companies like Google and Microsoft. Then, to startups building apps and SaaS platforms.

This is now changing. Ex-ISRO scientists are founding space startups rather than retiring into consulting. Semiconductor design veterans are leaving comfortable positions at Qualcomm and Intel to build Indian chip companies. The talent pipeline that created India’s IT services industry is now pointing at manufacturing.

The Numbers Behind the Narrative

India’s manufacturing sector is valued at $1.62 trillion today and projected to reach $2.3 trillion by 2030. But the real story is what happens after.

NITI Aayog’s October 2025 roadmap, ‘Reimagining Manufacturing,’ frames the stakes clearly: manufacturing contributes just 15-17% of India’s GDP, compared to 25% in China and 27% in South Korea.

The gap represents both a deficit and a runway. The government’s target, manufacturing at 25% of GDP by 2035, would require the sector to grow at roughly 15% annually for a decade. That trajectory implies $270 billion in additional manufacturing GDP by 2035 and over $1 trillion by 2047.

Hitting 15% annual manufacturing growth while India’s overall economy grows at 6-7% requires productivity gains that manual processes cannot deliver. The math only works if Indian factories adopt automation and digital manufacturing technologies at unprecedented scale.

The global market for manufacturing efficiency technologies tells the same story from the demand side:

• Industrial IoT: $483 billion (2024), growing to $1.7 trillion by 2030 (23% CAGR)

• Industrial Robotics: $34 billion (2024), expanding to $61 billion by 2030 (10% CAGR)

• Smart Manufacturing (Industry 4.0): $350 billion (2024), projected at $791 billion by 2030 (14% CAGR)

India captures a fraction of these markets today. The opportunity for venture investors lies in backing the companies that will close that gap.

IP Over Assembly

Most commentary about Indian manufacturing focuses on assembly, the sheer volume of products that could be made here as companies diversify their supply chains. This framing treats India as a location arbitrage play: lower costs, lower tariff risk with acceptable quality.

The value capture in this model is thin. Competing on cost means margins stay compressed and defensibility remains elusive.

We are interested in companies that own intellectual property, capture margins through design and engineering, and use Indian manufacturing as a capability rather than just a cost center.

The distinction matters enormously for returns. An assembly operation might generate 2-3% operating margins even when run exceptionally well. A design-led company with proprietary technology can generate 25-35%. The former is a business while the latter is a venture-scale opportunity.

This is the lesson from IT services. The companies that built lasting value, the Infosyses and TCSes, didn’t just arbitrage labor costs. They developed proprietary methodologies, built deep client relationships, and moved up the value chain into consulting and product development. The pure labor arbitrage players got squeezed as wages rose and automation encroached.

Manufacturing will follow the same pattern, albeit faster: the engineering talent base already exists, global supply chain pressures are compressing timelines, and AI-enabled tools are accelerating product development cycles in ways that weren’t possible a decade ago.

The companies that matter won’t be competing to assemble someone else’s design at lower cost; they will be designing products, developing processes, and building technical moats that justify premium pricing.

Semiconductors: Watch the Edges

No sector generates more headlines than semiconductors.

The coverage focuses on fabs: Tata’s $10 billion Dholera facility, Micron’s Gujarat ATMP plant, and the government’s push for wafer fabrication, among others. These are significant projects, but they require tens of billions in capital, decades of operational learning, and relationships with equipment suppliers that take years to cultivate. They will be built by Tata, Reliance, and foreign partners/entities that can absorb the capital intensity and timeline.

The opportunity we find compelling sits around the fabs.

Design is where India already has global relevance. Roughly 20% of the world’s chip design workforce is based in India. They work for Qualcomm, Intel, AMD, and NVIDIA, designing chips that ship in billions of devices worldwide. The talent certainly exists.

However, at scale, Indian companies don’t capture the value of that design work. Most Indian engineers design chips owned by American companies. And obviously, the IP and the strategic control flow elsewhere.

The opportunity is founders who can build design-led semiconductor businesses that own what they create. Attacking specific verticals where domain expertise matters as much as design capability, think automotive, industrial or edge AI. These are sectors where understanding the application is as important as understanding the silicon.

Compound semiconductors offer a different entry point. Silicon fabs are a catching-up game, and India will be manufacturing established technology for years before reaching the frontier.

But gallium nitride and silicon carbide, materials essential for EVs, renewable energy systems, and high-frequency communications, are different. The global industry is still maturing. India approved its first commercial compound semiconductor fab in August 2025. This is a sector where India could build at the frontier rather than behind it.

Packaging is also underrated. Advanced packaging: chiplets, 2.5D and 3D integration, heterogeneous packaging, is becoming the bottleneck in semiconductor performance.

Intel, TSMC, and Samsung are investing billions because packaging is increasingly where performance gains come from, especially in the age of AI. India’s Outsourced Semiconductor Assembly and Test (OSAT) facilities are coming online. Micron’s Gujarat plant is an ATMP facility. The packaging layer requires less capital than fabs, has faster learning curves, and creates genuine technical moats.

Kaynes Semicon shipped its first paid chip modules in October 2025. CG Semi delivered commercial ‘Made-in-India’ chips in August. The ecosystem is all set to take off.

Electric Mobility: Beyond the Vehicle

There is no doubt that the EV opportunity in India is substantial. The market is projected to exceed $100 billion by 2030, with 1.97 million units sold in FY25 alone. But the value distribution within this market is uneven, and understanding where margins accrue matters more than sizing the overall opportunity.

Vehicle manufacturing is certainly a demanding business. It is capital-intensive, margin-thin, requiring excellence across design, manufacturing, distribution, and service. The companies that win will deserve their success. But from our vantage point, for venture investors, the risk-reward calculus is challenging.

The opportunity we find more compelling is in everything the vehicles need, as in the component layer that India currently imports.

Battery packs constitute roughly a third of an EV’s cost, and India imports nearly all its cells. The opportunity is in battery management systems, thermal management, and alternative chemistries that address supply chain concentration.

Rare earth dependency represents both a vulnerability and an entrepreneurial opening. The motors in most EVs require rare earth magnets. China controls 60-70% of rare earth supply and an even higher share of processing. Chara Technologies recently raised $6 million to scale rare-earth-free motor production to 100,000 units annually. This is the kind of company we find interesting: solving a genuine technical problem with global applicability.

Power electronics: inverters, converters, onboard chargers, are all high-value and technically complex. They are mostly imported today. The companies building these domestically will supply Indian EV makers and global OEMs seeking supply chain diversification.

Two-wheelers are where the volume is. India sold 1.15 million electric two-wheelers in FY25. The economics already work: lower purchase prices, lower running costs, and sufficient range for urban commutes. The component ecosystem serving electric two-wheelers is growing and becoming more accessible.

Space: A Genuine Moat

If we had to identify one sector where India has an authentic, defensible competitive advantage, it would be space.

ISRO spent five decades building space capability on constrained budgets. That frugality forced genuine innovation in design, materials, and manufacturing. Indian space engineers learned to do more with less because they had no alternative.

That institutional knowledge is now flowing into the private sector. When the government opened space to private participation in 2020, it unlocked one that had been building for decades. The 400-plus space companies operating today are staffed by engineers who learned at ISRO, use facilities that ISRO shares, and build on technology that ISRO transfers.

The cost advantage is structural: Skyroot targets launch costs that would be a fraction of Western equivalents. Pixxel’s hyperspectral satellites are dramatically cheaper than competitors. The design philosophy differs: optimized for cost-effectiveness from first principles rather than adapted from military programs or legacy architectures.

Launch capacity is a global bottleneck: Demand for satellite launch services exceeds available capacity worldwide. India’s PSLV has launched 433 foreign satellites with an exceptional track record. Skyroot’s new Hyderabad facility can produce one orbital rocket per month. Private Indian launch companies are positioned to capture meaningful share of the small satellite launch market.

The downstream applications extend far beyond ‘space companies’: Precision agriculture, insurance risk assessment, supply chain monitoring, climate tracking, urban planning, these are massive markets where space-based data is becoming essential. The companies capturing that value won’t all be space companies in the traditional sense, they could be agricultural analytics companies, logistics platforms, and climate intelligence businesses that use space infrastructure.

In short, the regulatory clarity and talent exists. The cost structures are also favorable. India can, without doubt, build globally competitive companies here.

Specialty Chemicals: The ‘Quiet’ Compounder

Specialty chemicals rarely feature in discussions of India’s manufacturing future. The products are invisible. These are intermediates, additives, and precursors that go into things people actually buy.

But the dynamics are exceptional. Growth in high-value segments is significantly faster at 11-12% CAGR. The real opportunity lies in high-value specialty segments where customer stickiness and margins are significantly higher.

China+1 supply chain diversification has made this growth structural rather than cyclical. Chemical supply chains are inputs to everything else, and buyers are actively seeking geographic diversification. India has the technical capability, the feedstock access, and increasingly the scale to capture this shift.

Specialization is the moat: Commodity chemicals compete on cost. Specialty chemicals compete on technical capability, reliability, and regulatory compliance. Indian companies that meet global standards—REACH compliance in Europe, EPA requirements in the US—earn margins that commodity players never will.

Electronic chemicals are the critical gap: Fabs require ultra-pure chemicals for etching, cleaning, and deposition. These electronic-grade chemicals are currently almost entirely imported. As India builds semiconductor capacity, the demand for domestic electronic chemical supply will grow. The specifications are demanding. The quality requirements are extreme. The opportunity is substantial.

This is a sector where strong operators can build durable, profitable businesses. The moat comes from process know-how, customer relationships, and regulatory compliance rather than patents. If patents can be added on top, it certainly will be the cherry on the cake.

Defense: Real Opportunity, Long Horizon

• India’s defense market is $30 billion and growing.

• Defense exports hit a record $2.8 billion in FY24-25.

• The government has mandated 75% domestic procurement.

The opportunity is real, but defense is different from commercial markets in ways that matter for venture investors. Procurement cycles are measured in years. The feedback loops that allow startups to iterate quickly don’t exist. Relationships with the government and large defense contractors matter enormously.

The iDEX program creates a wedge: The Innovations for Defence Excellence initiative has engaged 619 startups with 430 contracts. This provides a path for startups to prove technology, build relationships, and develop track records that can lead to larger opportunities.

Drones are the accessible entry point, and the UAV market is growing at 20-plus per cent annually. Regulations have liberalized, and barriers to entry are lower than in most defense segments. ideaForge holds 50% market share and ranks third globally in dual-use drones. This demonstrates what’s possible.

We are more interested in dual-use technologies that can generate commercial revenue while building defense relationships than in pure-play defense companies entirely dependent on government procurement.

Robotics: Inevitable Adoption, Indian Solutions

The trajectory of India’s robotics adoption becomes clearer when examining automotive, the sector where India performs best. India’s automotive robot density reached 148 per 10,000 workers in 2021, roughly where China’s automotive sector stood in 2010 at 131 units. Over the following decade, China’s automotive density skyrocketed to 772. India now sits at the base of that same adoption curve.

What happened in automotive will replicate across manufacturing. PLI-backed facilities need to hit quality standards that require automation. Labor costs are rising in manufacturing hubs near major cities. Global customers demand consistency that manual processes cannot reliably deliver. The adoption curve will steepen because it must.

This rising demand creates two distinct opportunities. First, India can build globally competitive robotics companies. Addverb Technologies, which raised $132 million from Reliance and produces 50,000-60,000 warehouse robots annually, already exports to global customers. So does GreyOrange. These companies prove that Indian robotics companies can compete at global scale.

Second, and this is where we see white space, is automation designed specifically for Indian manufacturing conditions. Global robotics leaders optimize for high-wage, high-volume, greenfield environments. Indian factories present different constraints: smaller batch sizes, brownfield facilities with legacy layouts, acute cost sensitivity, and the need for human-robot collaboration in less controlled settings.

The founders building for these parameters, rather than importing solutions designed for German automotive plants, will find less competition and deeper moats. They’ll eventually export.

The Platform Layer

Here’s where our perspective differs most from conventional manufacturing investment theses.

Most investors looking at manufacturing focus on product companies, the chip designers or the motor manufacturers or the rocket builders. These are, no doubt, legitimate opportunities.

But we believe a larger, more defensible opportunity often sits one layer beneath: the infrastructure that enables manufacturing.

Consider what’s worked in our adjacent investments. We’ve backed platforms that aggregate fragmented supply, create trust in unreliable markets, or bring software intelligence to analog industries. These are businesses that don’t essentially own the factories, they rather own the customer relationships that make factories viable.

The same dynamic applies to manufacturing.

India has over 2 crore manufacturing MSMEs (out of 6.82 crore total enterprises, ~32%), collectively employing crores of workers. The vast majority operate without digital production management tools. They are capable of producing quality goods but struggle to access customers, meet compliance requirements, and operate at consistent quality.

As highlighted above, the global Industrial IoT market, the infrastructure layer enabling MSME digitization, is valued at $483 billion and growing at 23% annually to reach $1.7 trillion by 2030.

Manufacturing software platforms, from ERP to MES to quality management systems, constitute a significant portion of this market. Add in AI capabilities and you have sure shot winners. The companies that solve coordination problems across India’s fragmented manufacturing base will capture value at scale.

Our portfolio company Bidso exemplifies this model. It operates as an IP-led, asset-light ODM that aggregates capacity across multiple facilities, with proprietary software managing quality control and production coordination. The moat is being the trusted interface between brands and factories, owning the design, controlling the quality, without owning the machines.

The question we are asking is, what’s the Bidso for auto components? For electronics assembly? For precision parts? For packaging?

Zetwerk has reached unicorn status as a B2B manufacturing marketplace. Platforms like these create value by solving coordination problems—matching demand with supply, ensuring quality consistency, managing compliance—rather than by operating production lines directly.

We find this model compelling. It’s capital-efficient, you don’t need to own factories to capture value from manufacturing, and it’s defensible; customer relationships and supplier networks create switching costs that pure manufacturing doesn’t. And it uses India’s complexity as an asset rather than fighting it as a liability.

What We’re Looking For

Across sectors, our investment criteria reflect consistent beliefs about what kinds of companies will win.

• Global ambition from day one: India-for-India is a fine business, while India-for-the-world is a venture business. This doesn’t mean exports on day one. Early revenue from domestic customers is often the right path to product-market fit. But the ambition should be global from the start: products designed to meet international standards, unit economics that work at export pricing, and a founder mindset oriented toward competing worldwide. The plan matters as much as the current execution.

• Defensibility that creates pricing power: We are not interested in companies competing purely on cost. Cost advantages erode as wages rise and competitors emerge. We look for proprietary technology—patents, trade secrets, process know-how, or network effects that create switching costs. Defensibility that justifies premium pricing and creates durable margins.

• Capital efficiency despite hardware economics: Hardware can be capital-intensive, but the best founders find paths to validation that don’t require $50 million before product-market fit. We look for founders who understand these paths.

• Founders who understand customers: The best hardware/advanced manufacturing founders didn’t start with a technology looking for a problem. They started with a problem they understood deeply, often from operating in adjacent industries, and developed technology to solve it. Customer obsession matters more than technical pedigree.

• Businesses that can reach profitability: Capital efficiency matters, and we back founders building businesses with plausible paths to profitability.

What We’re Not Doing

We are explicit about where we won’t invest.

• We’re cautious about pure-play R&D: R&D is essential, and many of the companies we back will be deeply technical. But venture fund cycles create real constraints. A company requiring 5-7 years of laboratory work before commercial viability faces a difficult fit with typical fund timelines, regardless of how promising the technology. We need to see a realistic path to revenue within a timeframe that works for our investors. Outliers exist, and we’ll evaluate those case by case.

• We’re not competing in capital-intensive commodity manufacturing: If the primary competitive advantage is scale and the path to winning requires billions in capex, this is a game for large conglomerates.

• We prefer diversified revenue paths: Government contracts can be valuable anchor customers, and various government programs have created/will create real opportunities for startups to build credibility and relationships. But companies that depend entirely on government procurement are difficult to manage at the startup stage. We look for founders building businesses with commercial revenue streams alongside any government work, not because government contracts are unattractive, but because diversification creates resilience.

Where we play

The manufacturing renaissance we’ve described requires two very different kinds of capital. Building a fab or an OSAT facility is a multi-billion-dollar undertaking, that’s Tata and Micron territory, backed by sovereign-scale incentives. As you can guess, we’re not in that game.

But those fabs need chips worth fabricating. The OSAT facilities need IP worth packaging. And that’s where the opportunity sits for a fund of our size.

A fabless chip design company doesn’t need $500 million to get started. It needs enough to pay 10-12 engineers for 18 months and cover one tape-out run. That’s a $3-4 million round. Likewise, a robotics startup needs a working prototype and initial customers willing to pilot it.

The distance between ‘exceptional founding team’ and ‘company with proof points that larger funds will back’, that’s the gap we’re built to fill, and we’ll do it early enough to matter.

The Honest Version

We’re entering advanced manufacturing with conviction and humility in equal measure.

Our conviction comes from structural analysis. The shifts we’ve described—supply chain diversification, policy maturity, talent reorientation—are real and reinforcing. The early evidence suggests this cycle is different from previous false starts.

Our conviction also comes from pattern recognition in adjacent sectors. These patterns will apply to manufacturing.

Our humility comes from acknowledging what we don’t yet know. Manufacturing has operational complexity that software doesn’t. Hardware development cycles are longer and less forgiving, and in the Indian ecosystem, the execution challenges are real. We’re building networks and developing pattern recognition in real time.

This thesis will evolve.

We’re not presenting a finished view from a position of omniscience. We’re sharing our thinking because we believe the opportunity is significant, the timing is favorable, and the kinds of founders we want to back are already building. Over the next few months, we will go deeper into each of these ‘sectors’ with our micro thesis.

The 25-year parallel feels apt. IT services grew from negligible exports in 1990 to over $200 billion today, creating millions of jobs and multiple globally competitive companies. Manufacturing has the potential for a larger impact: a greater employment multiplier, deeper backward linkages and a more diversified economic base.

But that outcome isn’t inevitable; it depends on founders who see the opportunity clearly and build companies that can compete globally.

The Invitation

If you’re building in advanced manufacturing—whether in semiconductors, EVs, space, chemicals, robotics, or the platform layer that enables all of them—we want to hear from you.

We’re particularly interested in founders attacking the problems we’ve described: design-led semiconductor companies, component businesses with technical moats, manufacturing infrastructure platforms, India-to-world products with defensible IP.

We’re looking for founders with deep conviction about problems worth solving and a credible claim to being the ones who can solve them.

The new era of Indian manufacturing is being built now, by founders who recognize the structural shifts and are positioning to capture them. We intend to be their partners and peers in the trenches.

Sources include NITI Aayog’s ‘Reimagining Manufacturing’ roadmap (October 2025), Press Information Bureau releases, IN-SPACe mission data, International Federation of Robotics reports, IBEF industry briefs, market research from Grand View Research, Mordor Intelligence, and MarketsandMarkets, and coverage from Business Standard, LiveMint, Economic Times, BBC, and CNBC.

PeerCapital hosted Alt+Play to understand this shift firsthand. We brought together founders and operators who are building games in India today, not to talk about the industry in abstract terms, but to discuss what is actually changing in how games get made.

Beyond Single-Format Thinking

The room was filled with builders who have moved past the old categories. They are not focused on whether something is casual or mid-core, mobile or console. They are thinking about what players actually want and how to deliver it in ways that were not possible three years ago.

This matters because the Indian gaming ecosystem has spent too long defining itself by formats that other markets established. The founders at Alt+Play are creating their own playbooks instead.

Generative AI Is Already Here

If there was one theme that defined every conversation, it was this: generative AI is not a future technology. It is a present-day tool that is already reshaping game development.

Teams are using AI across the full production stack. Asset creation, 3D environment design, level generation, gameplay mechanics, UI design, sound production, and dialogue systems are all being influenced by AI-driven workflows. This is not experimental. This is operational.

The impact is direct. Production cycles are faster. Creative exploration is cheaper. Small teams can now build experiences that would have required significantly larger studios just a few years ago. The gap between imagination and execution is narrowing.

What PeerCapital Is Backing

Our portfolio reflects this moment of change. We are partnering with companies that understand where the industry is headed and are building the tools and platforms to get there first.

Here is what we are focused on:

Platforms enabling game creation and publishing: As gaming gets more stream as a form of entertainment, it is the creator who needs the requisite tools and support to convert ideas to reality - platforms that enable the creator are an interesting area to focus on

AI-native infrastructure powering games: LiveOps for games is complex and many workflows can benefit from the injection of AI from bottom-up essentially making them AI-native.

India-specific studios: Studios that are building uniquely for the growing gamer population in India that can also unlock monetization (beyond RMG) is a key future growth area

Why Alt+Play Mattered

Alt+Play was valuable because it cut through the noise. There were no generic panel discussions about the future of gaming. There were no pitches or demos designed for investors.

Instead, there were honest conversations about the constraints teams face, the tools they are using, and the opportunities they see. The builders in the room are not waiting for permission or validation. They are solving real problems with real solutions.

This is where the momentum in Indian gaming is coming from. Not from the headlines, but from the people quietly building better systems, faster workflows, and more engaging experiences.

What Comes Next

Over the next few weeks, we will be sharing more from the founders who were part of Alt+Play. They will explain what they are building, why it matters, and what they are learning as they scale.

If you are building something in gaming, we want to hear from you. The shifts we saw at Alt+Play are just the beginning. The next decade of Indian gaming will be defined by the teams who understand where the technology is going and how to use it to create experiences that players actually want.

We are backing those teams. If that is you, let’s talk.

But that instinct exposes a deeper bias: the belief that AI must follow the same playbook as previous technology cycles. The game may be shifting, and the next wave of AI companies might not emerge from where the last wave did.

India has a credible claim as the place to build, or at least to attempt building, that next wave.

Why do I say this?

The centre of gravity in AI is moving from model training, which is capital-heavy and concentrated, to deployment and inference, which reward efficiency, distribution, and real-world constraints.

India’s structural advantages align directly with this shift, and products designed for India’s complexity turn into durable moats in any market.

From capital to efficiency

In January 2025, China’s DeepSeek built a model rivaling GPT for $5.6 million. To put that in context, it spent less than 6% of OpenAI’s estimated $100+ million cost. Using older Nvidia H800 chips, they achieved comparable performance through architectural innovation.

Nvidia lost $600 billion in market cap in a single day. Investors and the world at large, for that matter, realized the moat was architectural, not capital-based.

This signals a phase transition, from AI as a compute arms race to AI as an engineering efficiency game. AI’s economic value isn’t just in training foundation models (though it is most certainly a crucial piece).

A lot of real on-ground value can be derived in inference, deployment, and orchestration at scale. Training is expensive but happens once. Inference happens billions of times daily, and its unit economics determine who captures value.

To be clear: this isn’t about consumer AI apps. It’s about B2B infrastructure companies selling to enterprises, but this is infrastructure that must operate at population scale because of the volumes Indian deployments demand.

Indian enterprises already operate with this mindset. A recent survey of 200 Indian companies reveals that 91% prioritize speed of deployment as their biggest buying factor, not model sophistication, not parameter count, not benchmark scores. Indian enterprises are rather asking, “Can we deploy this next week?”

The data validates the shift: 47% of Indian organizations already have multiple AI use cases live in production, compared to the previous year, when most were still experimenting. This is a clear production-first mentality that aligns perfectly with the inference economy.

This is India’s natural terrain.

Building in India forces you to solve for efficiency from day one. You have to deal with limited compute budgets, constrained bandwidth, and users on feature/basic smart phones, among other constraints. These constraints help produce robust, globally deployable infrastructure.

Before someone says, “But shouldn’t we aspire to build foundational models?”. Of course, we should, and there’s no ambiguity about that ambition.

The point is simply this: we must also double down on India’s existing strengths and compete at speed, even as we build the foundations for long-term technological sovereignty. If we don’t scale now, we risk waking up to a world where we’re merely playing catch-up.

When constraints become competitive advantage

Here’s the conventional thinking: build for the US market, then ‘expand’ to India by localising.

What if you built for India first, and that became your unfair advantage globally, especially in the ‘AI age’?

Building AI for India means:

• Multilingual by default: 22 official languages, with only a relatively small probability that two random Indians share a common language

• Low-bandwidth optimized: Rural populations on 2G/3G networks

• Voice-first interfaces: Literacy barriers require audio-native experiences

• Offline-first architecture: Intermittent connectivity is the norm

How many of these constraints apply to Indonesia? Nigeria? Brazil? All of them.

You’re not just building for 1.4 billion Indians; you’re in fact building for 5+ billion people in emerging markets. India is the largest, most linguistically diverse laboratory for this market. Silicon Valley builds for San Francisco and hopes it scales to Jakarta. India builds for constraints that already represent the global majority.

The company that cracks voice-first, multilingual, low-bandwidth AI for India doesn’t need to ‘adapt’ for Southeast Asia or Africa; they already solved the harder problem.

India has also ‘seen’ scale.

India’s digital infrastructure has been deployed at a population scale with full-stack integration. UPI processed over 100 billion transactions in 2023, while Aadhaar provides a digital identity for India’s large population. Then there’s Bhashini, India’s AI-powered national language technology mission, delivering 300 million monthly translations.

The interesting insight is what these systems produce: edge cases at scale. When you run payment infrastructure for 100 billion transactions, you encounter every possible failure mode, be it network interruptions mid-transaction, simultaneous requests from remote villages, or fraud patterns across linguistic boundaries.

This creates training data and operational knowledge that doesn’t exist anywhere else. For AI infrastructure companies, this matters enormously. A credit scoring API trained on UPI patterns understands income volatility across informal economies. Healthcare AI trained on India’s disease prevalence works in markets where Western datasets don’t apply.

India should become ‘talent and data exporters’, architects of infrastructure built on a unique deployment density that serves global markets.

The inference economy: Where value actually gets captured

The B2B infrastructure layer has clear business models. Enterprises adopting AI face a predictable constraint: AI is probabilistic, enterprises need reliability. This creates demand for guard rails, integrations, and compliance layers. To add to this, traditional IT services are experiencing margin compression.

India’s advantage lies in solving for inference economics from day one. Limited budgets mean you can’t waste tokens. These constraints produce better infrastructure. This infrastructure can win globally as AI shifts from ‘can we train a model?’ to ‘can we run inference profitably at scale?’

The opportunity is certainly there. BCG projects $100-120 billion in AI software and services revenues by 2028 from India alone.

Why growth capital follows proof?

U.S. AI companies raised $110 billion in 2024, 62% more than 2023. India’s AI companies raised $780 million in 2024, 39.9% more than the previous year. While there is certainly a very significant gap, what matters more is the trajectory.

India’s late-stage AI funding reached $554 million in 2024, and early-stage funding (37% reduction) shifted toward proven companies with market viability. Investors aren’t avoiding AI, they’re rather seeking demonstrated traction.

In 2024, India’s total VC funding hit $13.7 billion, up 43%. There is no doubt that the capital is there; it is waiting for proof.

Growth investors need to see revenue scale, global customers, and exit paths. For years, these didn’t exist at scale. Now they do.

• The exit flywheel is starting: Twelve tech IPOs in 2024 raised $3.5+ billion versus five in 2023. Exit activity held firm in 2024, inching up to $6.8 billion. What stands out is the shift in where those exits came from: public markets accounted for about 76% of total exit value, up from roughly 55% the year before.

This swing was driven by a 7X jump in IPO exit value, supported by stronger liquidity, recovering tech valuations, regulatory tailwinds, and a backlog of companies finally heading to market. Each exit proves the model works, returns capital to LPs, and creates experienced operators who become the next generation’s founders and advisors.

• Revenue scale is arriving: Kore.ai ($150M raise), Atlan ($105M at $750M valuation), and others are at the leading edge. Indian Centaurs (companies with $100 million in ARR), and unicorns are on track to generate $20–25 billion in additional revenue by 2030, according to industry estimates. With AI advancing at an extraordinary pace, India’s software founders building AI-native products could exceed this projection, setting the stage for a new phase of accelerated growth and innovation in the region’s cloud economy.

• Enterprise adoption creates proof: Partnerships with OEMs and startups are fast becoming central to GenAI execution. Nearly 60% of organizations say they are co-innovating with these partners, reflecting a growing recognition that startups are essential for speed and fresh thinking.

The increasing dependence on agile, highly specialized partners marks a shift away from purely in-house strategies. Startups offer cutting-edge capabilities and rapid experimentation that large enterprises often can’t match, making them indispensable for staying competitive in the GenAI race. When enterprises deploy at scale with measurable ROI, it validates product-market fit for global markets.

AI and SaaS funding in India reached $1.7 billion in 2024, up 1.2x from 2023. The capital will flow to companies proving they can hit $100M+ ARR with strong unit economics.

The companies building now (2024 to 2026) are creating those proof points. First-wave exits unlock second-wave growth capital. The time for India’s big AI leap is here.

The talent question

There is the talent to make this happen, too. India has 600,000-650,000 AI professionals, projected to reach over 1.25 million by 2027.

Yes, I am aware of the flip side: India files far fewer AI patents than some other countries. That is a valid concern if you are measuring frontier research output.

But patents capture only one part of the value chain. The ‘valuable work’ in AI is not limited to inventing and training foundational models; it also includes implementing, deploying, and scaling them in the real world.

Taking a foundation model and making it work reliably in production; building guardrails; cutting inference costs by 10–100x; designing human-in-the-loop systems that work at scale, none of this is ‘easier’ or ‘less important’ work. It is exactly what enterprises and societies need right now.

In that sense, the talent pool that can repeatedly do this applied, production-grade work, is just as critical as the talent pool publishing papers and filing patents. We need both, not one at the expense of the other. And again, we certainly must find ways and bring in programs that encourage more foundational research, but that mustn’t stop us from capitalizing on what we are good at.

The ecosystem producing this talent is maturing rapidly. The U.S. has a structural advantage India can’t replicate overnight: FAANG alumni/unicorn alumni starting companies with personal wealth, networks that span Sand Hill Road, and operational DNA from seeing $100M+ ARR up close. Indian big tech—TCS, Infosys, Wipro—doesn’t produce the same founder profile. Services company discipline doesn’t automatically translate to product company DNA.

But the gap is closing faster than most realize. Product companies now produce founders. Flipkart alumni have founded 598 companies, raising $23 billion across 754 rounds, including seven active unicorns, with most ventures based in India (527) and the US (53). The ecosystem spans consumer (254) and enterprise applications (185) sectors and has seen 53 acquisitions so far.

Freshworks, Razorpay, Swiggy alumni are starting the next wave. These founders bring product thinking, not services thinking, they’ve built for global customers, raised venture capital, scaled teams.

Reverse brain drain will hopefully accelerate the maturation. Historically, most stayed. Now they’re returning. H-1B restrictions, U.S. tech layoffs, and India’s maturing ecosystem are shifting the equation. Returning talent brings not just skills but playbooks.

Tamil Nadu’s ‘Tamil Talents Plan’ offers competitive pay, research grants, and relocation support specifically for AI researchers. The IndiaAI Mission’s ~₹10,300 crore allocation includes support for researchers. More initiatives will hopefully follow.

India’s ever-growing product pedigree combined with execution discipline will enable the next phase of growth. For B2B infrastructure companies selling to enterprises, the ability to navigate complex sales cycles, manage client relationships, and deliver reliably, skills honed in services companies, becomes an advantage, not a deficit.

A decade and a half ago, India was not a very prominent name in the global SaaS market. As of 2023, it has 27 unicorns and 14 centaurs collectively generating $5.9 billion in revenue. Each success creates the alumni network for the next wave.

Additionally, there’s a significant salary differential: for example, some data show that reasonably experienced AI engineers in India earning around USD 60,000–80,000 per annum, while their counterparts in the US are earning USD 150,000–200,000+. That translates to roughly a 60–70% cost-savings.

Critics might argue this is just another version of the ‘SaaS talent arbitrage’ game. But I’d respond that given how fast everything is moving and how much is changing globally, we should play to our strengths. We should train this talent, give them the chance to build and think bigger. This moment is a solid foundation and a stepping stone.

The monetization question

The haunting question for every Indian tech investor remains ‘How will the startup crack the monetization problem?’

The SaaS lesson looms large. Consumer willingness to pay for SaaS within India is still relatively low, though it is showing signs of improvement. Enterprise software has traditionally struggled with long sales cycles and price sensitivity. The successful companies—Zoho, Freshworks, Icertis, and others—built ‘India to World’ models where a large chunk of their revenue came from overseas.

For venture returns, that model, in my opinion, still holds. Indian AI companies achieve scale by selling globally. India-to-India monetization might improve given the current AI reality, but I still think that while the India pie of revenue might improve, overseas is where these companies will make a significant dent.

Double-clicking on the India-to-India monetization bit: enterprise adoption is actually happening, and enterprises are paying.

76% of Indian business leaders believe GenAI will have significant business impact. The buying center shifts from IT (cost center, price-sensitive) to Operations/CX (revenue/efficiency center, ROI-focused).

Indian enterprises pay for AI when it solves core operational problems with quantifiable returns. The opportunity is in B2B infrastructure deployed at population scale. Over 95% of organizations allocate less than 20% of their IT budgets to AI. I’d see this as a positive sign, if the tool is right, the runway for growth is massive.

For venture returns, the model largely remains India-to-World. Indian AI companies solve problems for Indian constraints, then sell that infrastructure globally, where the same/similar constraints apply. Domestic traction validates product-market fit. International revenue drives valuation.

Architectural patterns

Instead of listing company names, it’s more useful to look at the architectural patterns that will shape India’s most successful AI products.

Pattern 1: Multilingual-first infrastructure

Teams building voice and translation APIs for India’s 22 official languages quickly discover that the same technology works for Indonesia, Nigeria, and most of Latin America. A company solving for Bhojpuri and Malayalam has already solved a harder problem than one building only for English.

India’s sovereign AI strategy will lean on Small Language Models (SLMs) because they match the realities of emerging markets: multilingual usage by default, edge-ready deployment for low-cost devices, and tighter control over compliance and costs.

The design assumptions here are clear. Users code-switch between regional languages and English mid-sentence. Also, dialects vary across districts within a state. Several languages have minimal datasets. This forces architectural choices, efficient embedding spaces, transfer learning approaches, and voice-first interfaces that make these systems stronger and more adaptable everywhere.

Silicon Valley is still fixated on bigger foundation models. India must optimize for models that run offline, work on feature phones, and handle multiple languages natively. What looks like a constraint is actually the blueprint for serving billions.

Pattern 2: B2B infrastructure built for population-scale deployment

The winning approach is to build enterprise infrastructure that can handle population-scale volumes. Conversation-intelligence systems process billions of interactions. Voice AI manages millions of concurrent calls. Translation layers handle hundreds of millions of monthly requests. These are enterprise contracts with mass-market load requirements.

The data validates this pattern. When Indian enterprises deploy GenAI, the top three functions are Operations (63%), Customer Experience (54%), and Marketing (33%). These are enterprise operations running at scale.

When high-bandwidth connections and powerful devices cannot be assumed, model distillation becomes standard. Edge strategies are built from day one. Caching and pre-computation are essential. These constraints produce infrastructure that is far more efficient than typical Western equivalents.

This is the layer where business models are clear, customers have budgets, and the product solves real pain, not a nice-to-have.

Pattern 3: Regulatory and compliance wrappers

Indian AI companies working with banks, hospitals, and government agencies build sophisticated compliance layers from the start. These same wrappers become exportable products because enterprises worldwide face similar risk, security, and data-residency requirements.

The domestic market forces you to meet the expected standards early. That compliance layer—think access control, audit trails, encryption, data localization—quickly becomes a competitive advantage in global enterprise sales.

Pattern 4: Emerging-market specialization

Healthcare models trained on tropical disease patterns, agricultural AI optimized for smallholder farmers, and credit scoring built on alternative data from informal economies are not ‘India-specific’. They are emerging-market infrastructure products with worldwide relevance.

These solutions make sense only when building for constraints: diagnostics for low-resource clinics, credit models for borrowers without credit histories, and education platforms that run smoothly on feature phones. Silicon Valley mostly ignores these markets; Indian builders understand them deeply.

The result is a consistent pattern: B2B infrastructure, designed for constraints, deployed at massive scale, and exported globally. These products aren’t built only for India, and they’re not consumer apps. They are infrastructure systems that work in India, and therefore work everywhere else.

GTM and distribution: Building the go-to-market muscle

All said and done, I do acknowledge that there is a question still lingering: ‘Can Indian companies crack enterprise GTM in outside markets better than how they did with the SaaS movement?’

I believe that AI infrastructure distribution works differently from traditional enterprise software.

The API-first advantage changes the playbook: Infrastructure and API businesses follow developer-led adoption, not top-down enterprise sales. A developer discovers your API, integrates it into their workflow, usage scales, and eventually converts to an enterprise contract.

Indian AI companies building API-first infrastructure—translation layers, conversation intelligence, voice AI, computer vision—benefit from this model. Product-led growth reduces dependency on expensive field sales. Developer adoption in one geography creates pull in others.

The co-innovation model accelerates GTM: As highlighted above, 60% of Indian enterprises co-innovate with AI startups. These are joint development agreements. When an Indian AI company co-develops a solution with an Axis Bank or a HUL, they’re creating a proven vertical solution they can replicate globally. ‘Built for one of India’s largest banks’ becomes the wedge for banking customers worldwide.

This is already working. Indian SaaS companies serve global customers with 70-90% of revenue from international markets.

Yes, scaling to $100M+ ARR requires GTM sophistication. But the distribution muscle for AI infrastructure is different from SaaS.

Product-led growth, API-first adoption, co-innovation models, and developer communities, these are channels where Indian technical talent excels.

Asymmetric returns, structural advantages

India offers venture-scale returns with structural cost advantages and earlier paths to profitability.

Indian companies often enjoy lower burn-rates, meaning longer runway and less dilution while achieving early milestones (first revenue, break-even, product-market fit).

Looking ahead, we should expect exit pathways for India-based AI firms to grow, including from international strategic buyers. The ecosystem is stepping up: funding is increasing, new startups are emerging, and leading Silicon Valley funds are either setting up dedicated India vehicles or, at the bare minimum, exploring the possibility.

Early movers benefit from lower valuations, access to top founders before competition intensifies, and the ability to define category leadership. Wait two years, and you’re competing with multi-billion dollar funds for the same deals.

From India, for the world

Here’s the thesis in one line: India won’t beat Silicon Valley at training foundation models (at least not immediately). It will win by building the infrastructure layer that makes AI actually work, for India first, and therefore for the world.

The phase transition from training to inference, from research to deployment, from consumer apps to B2B infrastructure—this is where India’s structural advantages compound.

The picks and shovels opportunity, where business models are clear, customers are enterprises with budgets, and India’s advantages in data, deployment, and efficiency all compound. This is exactly where Indian AI companies will capture value.

Silicon Valley will continue to lead in foundational research. China will industrialize deployment domestically. India will architect the infrastructure layer that makes AI work everywhere else: the orchestration platforms, compliance wrappers, inference optimization, enterprise deployment tooling, and specialized APIs that serve the global majority.

If you’re building AI infrastructure—the deployment tools, orchestration layers, enterprise enablers, or specialized APIs that make the ecosystem actually work—India offers structural advantages that no other geography replicates.

The infrastructure is being built. The only question is whether you’re laying track or waiting for the train.

This is precisely the thesis we’re backing at Peer Capital. We partner with founders who have global ambitions from day one. We believe that if you can solve for India’s complexity, you’ve already built the muscle to win everywhere else.

I’d love to hear from you: ankur@peercapital.in

Sources

https://www.bain.com/insights/india-venture-capital-report-2025/

https://www.bvp.com/atlas/the-rise-of-cloud-ai-in-india-2024

https://www.ey.com/content/dam/ey-unified-site/ey-com/en-in/insights/ai/documents/is-india-ready-for-agentic-ai-the-aidea-of-india-outlook-2026.pdf

Player data is gold, but studios are still mining with pickaxes.
Every studio today needs to have a granular view of their players. The world’s biggest games—BGMI, Clash of Clans, Fortnite have evolved into super hit titles. A big reason for this is the studios’ ability to run LiveOps - the art and science of running, optimising, and personalising games after launch.

Yet, even as mobile gaming is slated to become a $400 billion market by 2032, most studios still lack granular visibility into their own player data. Product managers depend on overworked analysts - event tracking is rigid and insights arrive late. The opportunity to understand players’ behaviour and subsequently personalise player experiences is lost.

The Problem

Studios shipped features and designed events off historical data. Players then behaved in unpredicted ways, and by the time the team understood why, those players had already disengaged or churned. The core gap: understanding behavior as it happens and responding in real time, not in hindsight.

Therefore, it has become imperative for a studio to have robust tooling and infrastructure to understand player data. But tools that are available in the market weren’t built for gaming. They flatten complex player behaviour into generic dashboards, forcing studios to rely on in-house analysts just to answer simple questions:

Why did a player drop off? What bundle should they see next? What went wrong in the recent release?

While studios have definitely shown an interest to improve their tooling, they have continued to rely on analysts to run simple data querying and analysis.

Why Now

The industry has shifted structurally. The Free-to-Play model dominates most of mobile gaming. User acquisition has become harder post-IDFA. Studios are under pressure to extend player LTV, not just download counts.

Meanwhile, AI-driven personalisation has become table stakes with players demanding hyper-personalised gaming experiences. But without strong analytics, personalisation remains a pipe dream for studios.

The rise of Generative AI in analytics has made it possible for startups to reimagine the data and the analytics stack for studios.

The global spend on game analytics and operations tools is pegged at $8B, projected to double by 2027.

Enter Game State Labs

Game State Labs (GSL) is building the predictive analytics engine powering next-generation LiveOps. Combine this with the hyper-personalization offering that the team is working on and it becomes a potent mix to provide a winning pitch to studios.

GSL helps product managers and game designers understand player behaviour at the most granular level—the “game state” before every action, purchase, or drop-off—and convert those insights into real-time personalisation.

The team behind GSL comes with highly relevant experience, having been in the trenches for many years solving the same problem in multi-national companies.

Aashbir Bhatia,an IIT-Delhi alum and also an ISB grad, brings rich experience in strategy and sales from his time at AT Kearney.

Ashwin Ramakrishnan, an RVCE alum and ISB grad worked at EA Games earlier, has led LiveOps for Real Racing, NFS Mobile, and Plants vs Zombies at EA games during his prior role.

Jagveer Gandhi, as a repeat entrepreneur, is the engineering brain in the mix, having been in the startup ecosystem through his prior roles at Akudo and Uni Cards

Together, they’re building what legacy tools couldn’t:

• Plug-and-play dashboards for player insights.

• Natural-language querying powered by LLMs.

• Personalisation engine that adapts each game in real-time.

Competitive Landscape

Horizontal tools like Amplitude and Mixpanel are being force-fitted into gaming workflows. These tools were never built for gaming market

GSL’s differentiation lies in starting where value is created—analytics as the wedge, personalisation as the moat.

By capturing game data at its source and becoming the system of record, GSL earns the right to power the next wave of in-game personalisation—a blue-ocean opportunity.

Our Belief

At PeerCapital, we back founders who build for inevitability.
Tomorrow’s studios won’t just monitor downloads—they’ll monitor player sentiment, game state, and lifetime value in real time.

Game State Labs is building that operating system for LiveOps—where analytics, personalisation, and monetisation converge.

We’re thrilled to partner with Aashbir, Ashwin and Jagveer as they redefine how studios play, analyse, and win.

Every enterprise today is a data company. Tech companies are unbelievably proficient at collecting, storing and analysing our data, and that’s quite literally by design. Data now lives across CRMs, SaaS tools, cloud buckets, and vendor systems- often without visibility or governance. This has led to increasingly familiar instances of data leakages and regulatory breaches, the blast radius of which is largely unknown to stakeholders within the organisations.

Data Privacy in the AI era

Despite the existence of Governance Risk and Compliance (GRC) tools, Data Privacy remains as the unsolved layer**-** the layer that asks, “Where is my sensitive data? Who can access it? Why do we even have it?”

With enterprises embracing AI across business functions, these questions have only grown harder to answer. AI systems thrive on data, and in doing so, they multiply its movement, transformation, and exposure.

For an emerging economy like India, AI adoption is a force multiplier for digitisation at scale- but nowhere are the stakes higher than in regulated industries like BFSI - where data is both the lifeblood and the liability. Banks and Insurers are increasingly using AI for customer engagement, fraud detection, and risk monitoring, often through a mix of in-house systems and third-party vendors. Every such integration expands the threat surface- from prompt injection attacks via voice agents to inadvertent data exposure in LLM training pipelines.

With the Digital Personal Data Protection Act (DPDPA) coming into effect alongside RBI’s Framework for Responsible and Ethical Enablement of AI (FREEAI), compliance isn’t optional- its foundational. Enterprises now need systems to honour the consent rights of an individual’s data, operationalise what they could and could not use, what their vendors could use, and really build trust with all stakeholders?

Our exploration revealed a fragmented market: Privacy consulting firms offering manual, static audits; point solutions offering disjointed modules; global new-age privacy automation platforms, expensive stacks that lack local context; and GRC providers that fall short of remediation. The result? A clear and widening white space.

Our search ended at Redacto

Redacto is a data privacy automation platform that helps enterprises discover, manage and protect sensitive data with the ability to track data lineage. The company is built by operators and practitioners who have lived and breathed data and privacy through their entrepreneurial and corporate journey across Grab, Google, Ola and Razorpay, in the past. Amit, Shashank and the founding team bring the perfect recipe to make this tall task come true.

Redacto is built on the fundamental principle that every single data that has come into the system should be tagged for its purpose, and data should be consumed only under the ambit of this purpose.

While most point solutions handle audits, consent or static vendor checklists, Redacto builds a privacy command centre, operating at the intersection of:

• Data Governance: Enabling visibility into where their data lives.

• Privacy Automation - enabling consent, classification, deletion, and compliance.

• Vendor Risk Management - ensuring enterprise data remains secure across third parties.

In short, Redacto brings together what’s long been siloed - compliance, privacy, and data security - into one adaptive, automated, auditable system.

Market Opportunity

We believe we’re at the beginning of a structural wave - the convergence of regulatory enforcement (DPDPA), AI-led data risks, and compliance automation has created a once-in-a-decade window.

Data privacy is where cybersecurity was 15 years ago- moving from a cost centre to a boardroom priority. The global privacy management software market is growing at 39.5% CAGR, with an estimated revenue forecast of $30B by 2030. We believe this is a $3B addressable market in the BFSI sector alone in India.

The Competitive Landscape

Globally, companies like OneTrust, Vanta, and Securiti.ai have validated the need for privacy automation (and most of them are present in India). But these complex solutions are found to be expensive, and rarely contextualised for India or emerging markets. Locally, most enterprises rely on stopgap measures such as Big4 audits, spreadsheets, and disjointed tools.

Redacto’s India-first architecture, built by practitioners who’ve lived these problems, offers a clear edge.

We heard this repeatedly in our diligence calls:

“We don’t just need compliance reports - we need solutions with auditable trails and remediation measures fitting to the India context.

That’s the whitespace Redacto is filling.

Our Belief

At PeerCapital, we back founders who see around corners - those who don’t just build for the present but engineer for what’s coming. Tomorrow’s enterprises won’t just monitor service availability; they’ll monitor privacy availability - how well they can prove, in real time, that customer data is safe, compliant, and retrievable.

Redacto is building that operating system - where data visibility, vendor trust, and compliance automation come together.

As one CISO put it in our diligence calls: “With cybersecurity, we’ve secured the walls. Now we need to know what’s inside the vault.”

We’re thrilled to partner with Amit, Shashank, and the founding team as they redefine how enterprises protect what matters most -their users’ trust.

Why? Because they've mastered something most people don't fully appreciate: the Lindy Effect.

The Lindy Effect says that for non-perishable things like technologies or ideas, the longer they've been around, the longer they're likely to stick around. It's counterintuitive in a world obsessed with disruption, but it explains why your company is probably still running on software that feels like it was designed in a different century.

But here's the twist: understanding why incumbents are so sticky also reveals exactly how to beat them. And right now, AI is creating the biggest opportunity in decades to do just that.

Why Enterprise Software Gets Stickier Over Time

Legacy enterprise software doesn't persist because it's the best technology. It persists because it accumulates three types of gravitational pull that get stronger with age.

Control Points: Where Business Actually Starts

Think of a control point as the place where someone begins their workday. Not just any software they use, but the primary dashboard where their function's most important work happens.

The history of enterprise software is really a story of control point transitions. In the 1980s, if you worked in finance, your day started with logging into SAP or Oracle to run reports and update ledgers. These systems owned the "finance control point" because they stored all the company's financial truth.

Then came the 2000s and suddenly your sales team's day started in Salesforce instead of whatever clunky CRM system came before. Salesforce didn't just store customer data better, they made the daily experience of being a salesperson dramatically easier. They captured the "sales control point."

The most valuable control points share a few characteristics. First, they sit close to revenue. If your software directly impacts how money gets made, you're in a good spot. Second, they own decision moments. When someone needs to make a choice that kicks off new work, being their go-to tool gives you incredible expansion opportunities. Third, they become integration hubs where other systems need to connect.

Take something like Stripe in payments. They didn't start as a control point, but now developers begin their payment integration thinking with Stripe's API. That's control point capture in real time.

Workflow Gravity: When Software Becomes Habit

Workflow gravity happens when people don't just use your software, they live in it. Their muscle memory, their mental models, their daily rhythms all get built around your interface and your way of thinking about problems.

Consider how Figma captured the design control point. Designers didn't switch from Sketch just because Figma had better features. They switched because Figma made collaboration feel natural in a way that matched how design teams actually work. Once teams built their workflows around real-time collaboration, going back to file-based design tools felt like going back to fax machines.

The stickiness comes from three things. First, time investment. When someone spends hours every day in your software, switching means relearning how to do their job. Second, institutional knowledge. Teams develop shortcuts, workarounds, and best practices that only work in your system. Third, identity formation. People start thinking of themselves as "Figma designers" or "HubSpot marketers." Their professional identity gets tied to your platform.

This is why consumer social networks are so sticky, and why enterprise software companies study them obsessively. Instagram doesn't just store your photos, it shapes how you see the world. Enterprise software that achieves this level of workflow integration becomes just as hard to replace.

Data Gravity: The Architectural Lock-in

Traditional data gravity worked like this: your system became the source of truth for some critical business identifier. Customer IDs, employee records, product catalogs. Once other systems started referencing your data, switching became an architectural nightmare.

But AI is changing this game completely. Large language models can now automate most data migration tasks that used to require armies of consultants. What once took months of professional services work can increasingly happen over a weekend.

This creates a weird paradox. Easier migration helps startups steal customers from incumbents, but it also makes those startups more vulnerable once they've won. The old moats are crumbling, but new ones are forming.

The new data gravity isn't about storing canonical records. It's about generating behavioral data that gets more valuable over time. Gong doesn't just store call recordings, it captures the patterns of how deals actually get done. Notion doesn't just store documents, it learns how teams think and work together. This behavioral data is much harder to replicate because it emerges from actual usage over time.

Here's the kicker: this behavioral data is exactly what AI models need to be useful. Static customer records are boring to an LLM. Rich interaction data that shows how work actually happens? That's gold.

Account Gravity: When Users Control Budgets

Account gravity is simple but powerful: your biggest advocate is also the person writing checks. When the CEO uses your product daily and loves it, renewal conversations become formalities.

This shows up in two ways. First, executive engagement. When C-level people are hands-on users, they become internal champions for expanding usage. Look at how Slack grew within organizations. It often started with executives who got frustrated with email and started using Slack for their own team coordination. Once the CEO was in Slack all day, rolling it out company-wide became inevitable.

Second, integration centrality. When you become so important that other vendors have to play nice with you, you've achieved something special. Every new tool the company considers has to work with your API. Every vendor demo includes a slide about integrating with your platform. That's when you know you've made it.

Where Traditional Gravity Breaks Down

Despite these powerful forces, cracks appear in incumbent armor. Understanding where traditional systems fail reveals exactly where VSaaS companies can break through.

The Professional Services Trap

Legacy systems stay sticky partly because they've built entire ecosystems around painful implementations. System integrators make millions on multi-year ERP deployments. The worse the initial experience, the more locked in customers become because they've invested so much in making it work.

But this creates opportunity. If you can deliver immediate value without the implementation nightmare, you can flip the script. Instead of customers dreading your deployment, they can see value on day one.

Look at how Airtable attacked traditional database solutions. Instead of requiring months of schema design and professional services, they made it possible to build useful applications in minutes. The ease of getting started became a competitive advantage, not just a nice-to-have feature.

The Short Half-Life Problem

Traditional systems optimize for data that stays valuable for a long time. Customer records, employee information, financial data. But many business problems actually depend on data that loses value quickly.

Real-time inventory levels, dynamic pricing signals, customer sentiment, competitive intelligence. These all require fresh data and immediate action. Legacy systems built for different paradigms struggle with high-velocity use cases.

This is where modern VSaaS companies can excel. Cloud-native architectures and AI-powered automation can process and act on short half-life data much more effectively than systems designed in previous eras.

How Winners Actually Break Through

Let's look at some companies that have successfully challenged incumbent systems, focusing on strategies that actually work.

The Surround and Squeeze Strategy

Instead of frontal assault, smart companies start by integrating with legacy systems while providing dramatically better workflow experiences around them.

Take Ramp in corporate spend management. They didn't try to replace accounting systems like NetSuite or QuickBooks directly. Instead, they built the best possible experience for making purchases and managing expenses, then integrated seamlessly with existing accounting workflows. Finance teams started living in Ramp for day-to-day operations while still closing books in their traditional systems.

Over time, Ramp accumulated so much transaction data and workflow control that they could start offering financial planning, bill pay, and other services that traditionally lived in accounting systems. The legacy ERP gradually got relegated to a background database.

Lattice did something similar in HR. Instead of trying to replace Workday directly, they focused on performance management and employee engagement. Workflows that traditional HRIS systems handled poorly. Once they owned those high-engagement touchpoints, expanding into core HR functions became much easier.

The Vertical Deep Dive

Some companies succeed by going incredibly deep into specific industries that horizontal solutions serve poorly.

Toast revolutionized restaurant point-of-sale by understanding that restaurants aren't just retail stores with different inventory. They have unique workflows around table management, kitchen operations, delivery coordination, and staff scheduling. By building specifically for restaurants, they could offer capabilities that generic POS systems couldn't match.

Procore did the same thing in construction project management. Construction projects have unique requirements around document control, change order management, and field-to-office communication. General project management tools like Microsoft Project couldn't handle the complexity, creating space for a vertical solution.

The key insight: industries with unique workflows and high switching costs are often underserved by horizontal solutions. Going deep can create sustainable advantages that general-purpose tools can't replicate.

The AI-First Transformation

Some companies are using AI to leapfrog traditional interfaces entirely, creating fundamentally different user experiences.

Consider what AppFolio built with Realm-X for property management. Traditional property management software required dozens of clicks to create a tenant prospect record. Navigate to the people tab, enter contact info, find source fields, add property relationships, submit forms. Classic enterprise software drudgery.

With Realm-X, the workflow becomes: "Frank Johnson called about the downtown apartment." The system automatically creates the complete record with all the right relationships and data. This isn't just better UX; it's a completely different paradigm that makes traditional interfaces feel ancient.

Or look at what Harvey is doing in legal workflows. Instead of making lawyers navigate complex document management systems, they can simply describe what they need: "Draft a merger agreement similar to the Tesla deal from 2018 but with Delaware jurisdiction." The AI handles all the traditional software complexity behind the scenes.

The Evolution from CRUD to Intelligence

We're living through the biggest interface paradigm shift in 30 years. For decades, enterprise software was basically fancy forms for creating, reading, updating, and deleting database records. Users did all the cognitive work while software handled data storage and simple calculations.

The Historical Arc

In the 1980s and 1990s, software like Siebel and PeopleSoft were essentially database front-ends sold to managers who wanted consistent processes and reporting. The software wasn't designed to help users; it was designed to capture data and enforce workflows.

Salesforce changed the game in the 2000s by making software genuinely easier to use. Cloud deployment, intuitive interfaces, real-time collaboration. Suddenly enterprise software could be as approachable as consumer web applications.

The 2010s brought workflow automation and systems of action. Tools like Slack, Notion, and Airtable focused on making daily work more productive rather than just storing information.

Now we're entering the AI-native era where software can actually assist with thinking and decision-making rather than just data management.

The Interface Revolution

This isn't just about adding chatbots to existing software. The entire interaction model is changing. Instead of navigating through menus and forms, users describe what they want to accomplish in natural language.

Instead of learning how to use software, users teach software how they work. Instead of adapting human processes to fit software limitations, software adapts to human preferences and mental models.

This creates massive opportunity for VSaaS companies that build AI-native experiences from the ground up rather than bolting AI onto legacy architectures.

The Current Market Reality: Opportunity and Risk

Despite all the hype about AI transformation, the reality on the ground is more complicated. Recent research shows that only 5% of custom enterprise AI tools actually make it to production. There's a massive gap between pilot projects and real business impact.

The Deployment Gap

Most enterprise AI initiatives fail not because the technology doesn't work, but because they don't integrate well with existing workflows. Companies build impressive demos that fall apart when real users try to incorporate them into their daily routines.

Meanwhile, over 90% of knowledge workers are using consumer AI tools like ChatGPT for work tasks. The gap between shadow AI usage and official enterprise deployment reveals something important: people want AI assistance, but they want it integrated into workflows they already understand.

This creates opportunity for VSaaS companies that can bridge this gap. Build AI capabilities that enhance existing workflows rather than replacing them entirely. Focus on immediate utility rather than impressive technical demonstrations.

Investment Pattern Insights

Current enterprise AI spending reveals some interesting biases. About 50% of budgets go to sales and marketing use cases, even though back-office automation often delivers better ROI. This happens because sales results are easier to measure, not because they're necessarily more valuable.

Mid-market companies are moving much faster than enterprises. Average pilot-to-production timeline is 90 days versus nine months for large enterprises. This suggests that VSaaS companies should consider mid-market entry strategies that allow for faster iteration and learning.

General-purpose AI tools are seeing higher success rates than highly specialized solutions. This might seem to contradict the vertical software thesis, but it actually suggests that flexibility and adaptability are more important than narrow optimization in early AI deployments.

Building VSaaS Companies That Last

So how do you build a vertical software company that can actually challenge incumbent systems and create lasting value?

Pick Your Battles Carefully

The best VSaaS opportunities tend to share several characteristics. Highly fragmented markets with lots of potential customers. Large, growing total addressable markets. Industries that are technology-reluctant and underserved by existing solutions. High revenue, low margin operations where efficiency gains create substantial value. Heavy reliance on human knowledge work. Relationship-based businesses with sticky customer contracts.

But beyond market characteristics, look for places where incumbent solutions are genuinely inadequate. Not just "could be better" but "actively frustrating to use" or "missing critical functionality." The switching costs need to be worth overcoming.

Build Real Differentiation

AI capabilities provide temporary advantages, but sustainable differentiation requires more. Domain expertise integration where you teach AI models industry-specific knowledge that generic solutions can't match. Workflow optimization that matches how work actually gets done in your vertical, not how software vendors think it should be done. Outcome focus where you measure success by business results rather than software utilization.

The goal isn't to build better enterprise software. It's to solve business problems in ways that make the software interface feel invisible.

Create Your Own Gravity

While traditional data gravity is weakening, new forms of defensibility are emerging. Behavioral data accumulation where rich interaction patterns improve your AI models over time. Workflow integration depth where switching would require retraining entire teams. Network effects where customer value increases as more participants join your platform.

The companies that master this transition will capture enormous value by solving real problems for underserved markets. But success requires understanding the forces that create software durability and building platforms that can accumulate similar gravitational advantages.

The Window Is Open

We're in a unique moment where the technological foundations of enterprise software are shifting faster than incumbent systems can adapt. AI capabilities, changing user expectations, and new data architectures are creating opportunities that haven't existed since the move to cloud computing.

But windows don't stay open forever. The companies that move decisively now, with clear strategies for building sustainable advantages, will define the next era of enterprise software. The question isn't whether disruption will happen, but which companies will execute the transition most effectively while the opportunity is still available.

The winners won't necessarily have the best AI models or the biggest datasets. They'll have the clearest understanding of why software becomes sticky and how to build platforms that accumulate the same gravitational advantages that have made incumbent systems so persistent.

That's the real opportunity in VSaaS: not just building better software, but building software that gets better and more valuable with time.