In 2015, at the fund I was managing, we made an investment in a company called Reverie Language Technologies. The company was trying to solve a problem around language equality and digital exclusion, built around the reality that apps, even those being built in India for Indian users, were largely constructed in English and had little or no meaningful translation capability in Indian languages. I remember something very clearly from our first meeting with the team, when they told us that if you looked at the entirety of all the content on the internet, less than 1% of it was in any Indian language, including Hindi.

That was roughly ten years ago, and while the internet has changed beyond recognition, the underlying imbalance has not. English still accounts for almost half of all websites whose language can be identified, while Spanish, the next closest language, accounts for roughly 6%. India representing almost 20% of humanity barely registers despite the fact that India itself has now largely come online. Roughly one billion Indians use the internet, which means the access problem is no longer simply whether someone owns a phone or can afford data. The more pertinent question now is whether the digital world they have connected to can understand them in the first place.

This is not an India-only problem. Versions of it exist across Southeast Asia, Africa, Eastern Europe and much of the rest of the world, where hundreds of millions of people have joined a global internet whose linguistic architecture bears remarkably little resemblance to the languages they actually use in their daily lives.

There are plenty of reasons to worry that AI could amplify many of the inequalities already embedded in the internet. On language, however, I think there is a possibility that it could begin to reverse and eventually upend the foundational underpinnings of the entirety of what it means to be online.

You cannot meaningfully engage with India without thinking about language. The last Census found that only about 10.6% of Indians spoke English at any level, and virtually none spoke it as their mother tongue. Those numbers are now fifteen years old and English usage has undoubtedly increased, particularly among younger Indians, but the larger point remains: English became the default language of India’s digital economy without ever becoming the default language of Indians.

Ever since Reverie’s early days (the company was acquired by Reliance in 2019) Indian language interfaces have become considerably better, but localization generally meant taking an interface originally conceived in English and translating it into something else. The menus, dropdown forms, and buttons stayed the same, and so did the fundamental assumption that the person’s needs would be adequately fulfilled by a blunt-force translation.

India has 22 constitutionally recognized Scheduled Languages, but even that number is not particularly helpful if you are trying to understand how Indians actually communicate. The Census recognizes 121 languages, with many hundreds more mother-tongue varieties reported, and even those categories flatten important distinctions once you are trying to get a machine to understand an actual human being.

Take Bhojpuri, not one of the 22 Scheduled Languages, yet more than 50 million people report it as their mother tongue. That’s a striking number approaching the entire population of Italy, yet within India’s official language framework it still sits outside the list of 22. Rajasthani, with roughly 26 million speakers, is another example, and there are many more as you move down through the extraordinary linguistic complexity of the country.

Compare that with much of Europe, where languages tend to exist in comparatively standardized written and spoken forms and where translating, say, French into Italian is a reasonably well-defined machine problem. India presents something quite different, because there are hundreds of languages and dialects and, even within the major languages, enormous differences in how people actually speak them. Hindi spoken in Delhi is not Hindi spoken in Bihar, Tamil spoken in Chennai is not necessarily Tamil spoken in Dindigul, and in urban India people frequently do not speak one language at all, but move between two or three of them, often throwing in whatever current slang I am now too old to recognize.

Anyone who has spent much time in India hears something like this constantly: “Swiggy pe dosa order kiya tha, but delivery guy wrong Prestige gate pe hai, security usko andar nahi aane de raha, aur ab app bol raha hai ‘order delivered.’”

What language is that exactly? It is perfectly intelligible, distinctly Indian and completely ordinary, but it is not something that can be solved simply by taking an English sentence and translating it into Hindi. It is transliteration, code-switching, borrowed vocabulary, local context and a particular rhythm of speech all occurring at once, and that presents a very different problem from conventional translation.

Language lives in the wild, yet the solution has generally been to standardize and bludgeon and then ask the user to make a choice from a drop-down menu beside a country flag: choose English, choose Hindi, choose Polish, type your search, navigate the menu and proceed from there. Even when we successfully translated the internet into Indian languages, we were still asking Indians to communicate with online applications in a way that computers understood.

Voice AI is therefore an especially interesting frontier for India because it has the potential to upend the relationship between the user and the machine in ways that have not been done before. Instead of asking someone to navigate a banking app, perhaps they can simply tell the bank what they want to do in whatever combination of Hindi, English, Assamese or local slang happens to come naturally. Instead of finding the right government website, figuring out which department provides a particular service and then working through a series of forms, perhaps they can simply ask for what they need. The person does not necessarily have to know what the underlying application is, where the information resides, which menu contains the answer or even what language the underlying system was originally constructed in, because increasingly those are problems the machine can solve rather than problems the human being must solve in order to use the machine.

For most of the history of computing, humans have had to accommodate machines. We learned keyboards, menus, browsers, apps and all the other conventions that now feel perfectly natural only because we have spent decades being trained to use them. The smartphone made those conventions dramatically easier and brought billions more people into the digital world, but it never really changed the underlying relationship: we still learned how the machine wanted to be used.

Language models introduce the possibility that the machine can finally do the accommodating, and that means something very different in India than it does in the United States. Making it slightly easier for an American to book an airline ticket or deal with a bank is useful, but it is largely incremental. Allowing someone who has never been particularly comfortable navigating an English-language digital interface to speak naturally in Telugu, switch halfway through the conversation into English, use a local expression that never appeared in a textbook, and still accomplish whatever they set out to do is actually revolutionary.

There is, of course, a great deal standing between that idea and reality. Understanding Indian languages is not simply a matter of feeding more Hindi into a large language model. Speech data is uneven, accents and dialects are badly represented, code-switching makes training substantially more complicated, and enormous amounts of the linguistic material that exists in everyday Indian life have never been digitized at all. The conversations taking place in homes, mandis, call centers, buses, and villages across India have always existed, but for the most part they never became part of the corpus on which the digital world was built.

And that’s part of the problem, because AI learned language by consuming a digital world that was already badly skewed. One estimate puts roughly 90% of the training data behind current generative AI systems in English, while some very large languages remain almost invisible in the datasets on which these models depend. Hausa, Pashto, Amharic, Yoruba, Sundanese, Sindhi and Zulu, for example, each account for less than four-thousandths of one percent of Common Crawl.

AI learned how human beings communicate largely by consuming the internet, but the internet itself never came close to representing the way much of the world actually communicates. In effect, we trained machines on the linguistic distortion created by the previous generation of technology and are now asking those same machines to understand everyone whom that technology left out.

India may therefore be one of the places where the limitations of the current generation of AI become most obvious, but it may also be one of the places where overcoming those limitations matters the most. If AI can move beyond translating the existing digital world and instead allow people to interact with that world in their own language, accent, dialect and natural manner of speaking, then we are no longer simply localizing the internet or making another interface slightly easier to use.

The world we experience online reflects only a marginal sliver of reality; vast amounts of human language, expression and experience have simply been obscured from it. Bringing that world online will do much more than expand access; it will likely fundamentally rewire the internet as we know it. The internet flattened context, AI may actually give it its shape back.

The gulf between India as it is and India as it is viewed by the West is as old as time and stretches across every imaginable facet. As it pertains to global asset allocation, the story is no different. If a pension or endowment moves from a four to a five percent allocation based on a macro assessment, does that really make a statement or capture the essence of what is actually happening in the country? Across every corner of India bazaars are humming and consumers are spending, yet it all remains largely illegible to the West.

At an even more fundamental level, among every panel discussion which invokes the world’s fastest growing democracy and the demographic dividend, sits something more basic beyond the question of what the allocation should be, which is whether India is even investable in the first place?

By investable, I do not mean whether one should allocate capital to India. I mean whether, even if one wanted to, is there a clear and functional pathway for foreign investors to do so? The conversation around India as an investment destination assumes that such pathways exist. In practice, they exist only in a narrow and highly conditional sense. I would argue that there is actually no road. More specifically, the path exists but only for those equipped to build it themselves.

At the top end, large institutions can access India directly through the FPI route. They have the scale, infrastructure, and internal capacity to navigate registration, custody, compliance, and reporting. That system works, but it serves a very specific class of capital. Beneath that sits a far larger universe of investors, RIAs, family offices, smaller institutions, and individuals. In the United States alone, this represents well over $100 trillion of capital. For that segment, access is not meaningfully available. What exists instead is a thin layer of products that create the appearance of access without providing any real scale or substance.

The entire US-listed India ETF market sits at roughly $15 to $17 billion, anemic relative to what it is meant to represent. That is the size of an average mid-cap US company. It is roughly equivalent to the entire annual GDP of the Bahamas or Brunei. It is about half the size of Albania’s economy. These are strange benchmarks when you consider that this is the primary access point for US capital into a $4 trillion economy that is on its way to $10 trillion and represents 20% of humanity.

It’s like we decided to have a party and then get angsty and upset that no one showed up, without recognizing that we held it in an undisclosed location and didn’t actually invite anyone to begin with. A few made their way in either out of sheer luck or determination, but most remain simply completely unaware that there is plenty of samosa and bhaji on offer. This is the primary interface through which the world’s largest capital base engages with India.

Even within the limited construct of the handful of available ETFs, exposure is concentrated in a narrow set of large-cap names, banks, IT services firms, and conglomerates. The largest US-listed ETF, administered by BlackRock, has roughly $8 billion in assets. Within the fund, the top names, Reliance, HDFC, ICICI, Infosys, and TCS, represent a significant portion of the portfolio. All fine firms, but hardly representative of the broader market.

There are only a handful of ADRs with consistent liquidity and a few closed-end mutual funds that do not add much to the aggregate number. For most US-based investors operating outside the FPI framework, the opportunity set is constrained to a small subset of India, with most of it completely out of reach.

What is notably absent from that accessible layer are the parts of India that are actually shaping the economy. Mid-sized manufacturers integrated into global supply chains, exporters taking new brands to market, NBFCs extending credit into segments that formal banks do not reach, newer public companies that have come out of the last IPO cycle. These businesses are listed and they are operating at scale and they are investable in a technical sense, but they are not accessible in any practical one.

This gap is often diagnosed as a perception problem. There is a tendency to say that India is not fully articulated, and that global investors do not appreciate the scale of what has been built, which is also true. I had a discussion recently with one of the architects of India Stack along these lines. I chided him a bit and asked him why the country constantly undersells the extraordinary story of UPI and Aadhaar to the world. Why isn’t this the biggest story of the 21st century? His view was that the story is too difficult to communicate because there are no clean parallels. There is no simple reference point that allows someone to say this is equivalent to something they already understand. That may be true, but to bring it back to markets it is not the only binding constraint.

Capital will move when a story is compelling, but it is also far more likely to move when there is a structure that allows it to flow in a way that fits within existing mandates. In that sense, India does not only have a marketing problem, it has an access problem.

This is not limited to US investors as European investors face similar constraints, often compounded by regulatory limitations that restrict access to US-listed instruments. Japanese investors, despite deep economic ties to India and among the highest savings rates in the world, have very few domestic vehicles through which to participate in India. In the Gulf, access beyond the largest sovereign pools remains limited. Across geographies, capital exists but the ability to express that interest at scale in India does not.

There is a broader assumption embedded in global markets that capital naturally flows to opportunity. That is true in practice, but that infrastructure doesn’t naturally build itself. Yes, you need to solve for liquidity, for depth, for currency, but these are all surmountable issues.

Japan built a deep ETF ecosystem over decades and attracted significant capital despite nominal growth. China invested early in creating access pathways and continues to command allocation even through periods of uncertainty. India, despite its scale and growth, has not yet built that external layer with the same intent.

The consequence is that global allocation to India remains structurally lower than it could and should be. Ongoing debates about valuations, currency movements, or the return of foreign flows assume that investors are in a position to act, but in many cases they are not. An opportunity that cannot be accessed cannot be priced properly, and a market that cannot be owned at scale remains theoretical and unfulfilled. There is plenty of goli soda and pakoda to go around, and there’s valet parking and a band to boot. Most people just don’t know where to show up.

One afternoon early on in my term I was standing in line at the café with a friend chatting and as we reached the glass case containing veg puffs and other assorted snacks, I heard him mumble something that sounded like, “Samosa or chai.” It sounded like a question, as if he were asking the guy at the counter which one he should have, or which was better that day.

So I asked my friend, why did you order samosa or chai?

What do you mean?

You said samosa or chai. Why did you ask him to choose?

I didn’t. I got both.

You said “or.”

No, I didn’t.

Well what does “or” mean in Hindi?

What do you mean, ya?

I mean what does or mean. Samosa or chai.

I just told you, ya.

No, what you said. Or.

You mean aur.

Yes. Or.

No. Aur.

That’s what I said. Or.

No, he said. Aur.

What is aur?

Aur means and.

What do you mean or means and?

No. Aur means and.

So what is or?

Or is or. That’s English. You should probably know this.

So there is no or in Hindi?

There is ya. I told you already.

Ok so what is it, ya?

Correct.

What do you mean correct, yaar?

No, ya. Ya means or.

What!?

To the new firang, the difference between or and aur wasn’t audible, but beyond that the bigger mistake was trying to apply Boolean logic at the samosa counter.

You cannot say coffee or chai and then walk away with a cup of each. That defeats the entire function of “or” because or is binary. It’s an XOR gate, which by definition is exclusive. If you choose samosa, you have renounced chai. If you choose chai, you have forfeited samosa. That is the social contract embedded in the word or.

Whether it was aur or or, what I didn’t appreciate at the time was that I was being inducted into a completely new operating system, one that assumed inclusivity. Samosa did not cancel the chai and rice did not compete with dal; it assumed dal was there to begin with. In this system, or does not necessarily eliminate what follows, it leaves the door open to possibility.

This is why one-way roads don’t work in India. What one-way anyway, yaar? This way, that way, all ways, always.

Why choose when you can include? Why subtract when you can add? Why collapse possibility into a single traffic lane when there is room for both? The XOR gate belongs in the lab; the samosa counter and the chowks operate according to a more abundant philosophy.

But understanding aur and ya creates its own new challenge in that I am now completely ill-equipped for the world beyond India. Once you internalize inclusive logic, binary questions begin to feel unnecessarily cold and austere. It’s hard to unlearn that choices do not always require elimination, and that sometimes things simply accumulate. The samosa counter undid years of Western conditioning and trained me to assume nothing needed to be renounced.

As I sit here on my favorite perch in the afternoon breeze with my veg puff and adraki chai, it is now impossible for me to imagine one without the other. Naturally, I am wondering where the masala peanuts are as well.

So what actually makes it through the gauntlet of remembrance? Now that I am an ‘uncle,’ I find myself opining that much of what is popular today does not feel built for the long haul. It makes me wonder what history actually chooses to preserve, and why. Does it save what was best, what was most common, or what was most profound? Maybe it’s less about the ideas themselves and more about how they made it through history’s chopping block. To be fair, Books 2, 4, and 6 of Marcus Aurelius’ Meditations are epic reads, but when we look more broadly at the survivors, perhaps we are blindly guided by the ancient remnants of adverse selection.

If you could envision all of recorded humanity, it would resemble a landfill of beliefs, stories, and systems that arrived and vanished with equal fury. We do not know what was lost to history, so we treat the ideas we have as absolute truths when they are really just the ones that survived the long march of time. In the Indian worldview, we might distinguish between shruti, that which is revealed and eternal, and smriti, that which is remembered and shaped by context. We often mistake the surviving philosophical fragments for shruti, as if they were divine laws, when they are actually the ultimate form of smriti. We aren’t hearing the original song; we are listening to a low-bitrate file of a tribute band, stripped of the frequencies that make life rumble.

It is estimated that roughly 120 billion people have ever lived. The world then has played host to untold millions of thinkers, texts, languages, wars, fires, pandemics, and ideologies. The odds therefore are likely heavily stacked against the survival of most things. Those that endure seem to benefit from narrative simplicity and, most likely, a tremendous amount of dumb luck. “Amor Fati” thrives on Instagram sandwiched between kettlebell workouts and remote work retreats. Ryan Holiday sells it as modern gospel in The Daily Stoic, even peddling a coin emblazoned with the term so you can remind yourself to embrace your fate while waiting in line for a flat white. It fits neatly into today’s attention economy because it is clean, minimal, and “doing the work” for you, making it a comforting voice of reason during a CBD face misting.

Contrast that with the philosophies that did not make the canon. The Cyrenaics, who believed the purpose of life was immediate physical pleasure, likely stayed at the punch bowl too long to build a lasting legacy. Their ideas may have been compelling over a long weekend in Goa, but bacchanalian hedonism has an inverse relationship with staying power, or so I’ve been told. Then there were the Cynics, led by Diogenes, history’s original internet troll, who lived in a ceramic urn, defecated in public, and carried a lantern in broad daylight looking for an honest man. He practiced a form of feral nudity and anti-civility that was brilliant but fundamentally incompatible with institution building. We do not remember the philosophies that championed vagrancy although we can always hope for mean reversion.

This selection bias in philosophy mirrors the selection bias we see in the modern world. I am reminded of my colleague in India who, upon landing in New York, insisted on a detour to Patel Brothers in Jackson Heights to buy Indian spices to bring back to Bangalore. He loaded a shopping cart full of elaichi and channa, convinced they were superior because they were destined for a market with deeper pockets. He treated the ‘Export Quality’ stamp as a guarantee of authenticity, failing to realize that the best spices actually never left the neighborhood. We do the same with thoughts by importing bite-sized philosophical nuggets as productivity hacks while forgetting that a deeper and more textured experience is already being practiced across every dusty chowk in India.

Nowhere is this shadow intelligence more visible than in the workshops that the formal world has forgotten. Outside of Meerut, there is a cluster that manufactures medieval gladiator armor for film sets and museums. It is a world of blacksmiths and metal artisans producing chainmail and helmets off the grid. When you walk those lanes, you see and hear the cacophony of hammers hitting steel by men operating through tacit knowledge and mastery who do not need coins to remind themselves of Amor Fati. The reality is lived every time the power goes out mid-weld or a supply chain collapses.

I once visited a gherkin factory in rural Karnataka, down a long unmarked road where pickles were being jarred and white-labeled for Walmart and other global brands. From the outside it was a large, nameless shed with corrugated roofing. But inside, the equipment was complex and modern with stainless steel tanks and conveyor systems sorting cucumbers by size. There were barrels stacked to the ceiling and lines of brine-filled jars being capped and boxed, all while administered by a raised control center that looked like central command on the USS Enterprise. Labels were being printed for export in countless languages across as many brands with no hint of the global supply chains they were feeding into. I couldn’t believe something at that scale existed, in that place, with that level of precision. It struck me then, as it does now in Meerut, that there is an entire economy in India legible only to those who know where to look. And even then, just barely.

India is an endless study in layers of intelligence that will likely never be known because it doesn’t broadcast on a frequency that the modern world is tuned to hear. It’s not so much reading signals differently as listening to the hum of things working without expecting an explanation. Buried in the heat and dust and signal lag, you begin to understand that the volatility the Stoics feared is not something to be managed or transcended, it is simply how things are. It is the background noise of daily life.

If we’re honest, most of what we surround ourselves with is just the stuff that was light enough to float to the top. It’s the “Export Quality” version of reality; cleaned, polished, and easy to ship. But visibility isn’t the same thing as value. Real intelligence doesn’t feel the need to introduce itself at a networking event or sell you a commemorative coin. The composure people search for in packaged ideas is already being practiced, without language or attribution, in places that never needed to explain themselves in the first place.

India’s regulatory architecture is a constant story of evolution, one that grew out of the circumstances of a young nation attempting to stabilise itself after a long period of colonial extraction. In the beginning, it reflected a belief that the state needed to guide development in an environment marked by scarcity, uneven institutional capacity, and deep uncertainty. Over time this approach created a dense network of approvals, inspections, and procedural requirements that shaped how enterprise took root. Even as liberalisation in 1991 removed the most visible controls, the habits built over decades remained, and the regulatory state continued to grow through a thousand smaller processes distributed across jurisdictions. These systems were not conceived as barriers, but as they accumulated they introduced frictions that today weigh most heavily on the small and mid-sized businesses that sit at the center of India’s economic aspirations.

Digitizing Dysfunction

“The regulatory state didn’t disappear with liberalisation; it just dispersed into a thousand invisible frictions”

Liberalisation dismantled the large industrial apparatus but did not unwind the administrative instincts that had formed around it. The supervisory posture simply migrated to new domains. Environmental clearances, fire inspections, periodic renewals, municipal permits, sector-specific filings, and a growing array of digital submissions became the new expressions of an older logic. The Licence Raj no longer appeared in the form of a single ministry with sweeping authority but in a distributed set of obligations that touched almost every aspect of operating a business. Rules accumulated faster than they disappeared, since removing a requirement seemed riskier than letting it persist. Digitisation amplified these tendencies. Uploading a document made it easier for agencies to request more documents, automated reminders made it easier to create filings, and online portals made it easier for regulation to expand without appearing to do so. The interface changed but the mentality behind it remained.

Old Logic, New Portals

It is in this context that the reforms of the last several years take on greater meaning, because they represent the beginning of a reorientation in the way the state conceives of its economic role. India’s small and mid-sized enterprises, which number over 60 million, remain burdened by three structural costs . The first is capital, as borrowing can cost from 12 percent to more than 40 percent per annum, numbers that constrain risk-taking and force entrepreneurs to adopt defensive strategies. The second is logistics, as moving goods across India still costs close to 14 percent of GDP, a figure that reflects both the complexity of the country’s geography and the pace of infrastructure development. The third is compliance; in sectors such as real estate and infrastructure, the cumulative weight of inspections, certificates, sequencing requirements, and administrative ambiguity can absorb a fifth or even a quarter of total project expenditure.

The Structural Reset

“We are finally swapping the anxiety of control for the architecture of trust.”

The state has begun to address these constraints; one example is the redefinition of the MSME category. By raising the turnover thresholds from Rs. 50 crore to 100 crore for small enterprises and from Rs 250 crore to 500 crore for medium enterprises, the government allowed thousands of firms to continue operating within a lighter compliance system. Audit obligations, environmental reporting, labour filings, and procurement eligibility often change once a firm crosses specific financial ceilings. Increasing those ceilings gives firms room to grow without immediately triggering heavier regulatory demands.

Labour reforms mark another significant shift. For decades, 29 separate laws governed India’s labour market, contributing to an environment in which 42 percent of manufacturing workers are on temporary contracts. These patterns did not arise from a preference for informality but from the cost of crossing compliance thresholds. The consolidation of these laws into four comprehensive labour codes, the increase of applicability thresholds from 100 to 300 employees, the introduction of flexibility in retrenchment, and permission for women to work night shifts all signal a state that is beginning to understand that employment growth requires predictability and flexibility alongside worker protections. Implementation will be stage by stage across the states, but the framework itself represents a considerable institutional evolution.

Quality Control Orders (QCOs) provide another window into the reform process. They were created to ensure product quality but over time became a constraint for MSMEs because of bottlenecks in testing capacity. Certification delays often lasted up to 12 weeks, and with only 3,000 BIS personnel overseeing thousands of products, mandatory testing for inputs pushed firms toward a narrow bench of certified suppliers, thereby increasing costs throughout the value chain. The decision in late 2025 to modify or withdraw 76 QCOs, many relating to essential inputs such as polyester yarn, terephthalic acid, chemical intermediates, aluminum, nickel, tin, and refined zinc, is expected to reduce costs by 8 to 12 percent in downstream sectors. The Ministry of Steel’s suspension of 55 additional QCOs strengthens this shift and signals an ongoing willingness to recalibrate regulation based on data and industry realities.

The financial regulatory environment has seen a similar transformation. The Reserve Bank of India undertook a rationalisation effort that consolidated more than 9,000 circulars into 238 coherent, function-wise directions. This reduces interpretive divergence, strengthens supervisory clarity, and simplifies compliance for banks and NBFCs. It is not a reform that will appear on the front pages of newspapers, yet it represents the type of institutional housekeeping that underpins more efficient markets.

But, Lot’s More To Be Done…

Even with these reforms, institutional challenges remain. The average time to enforce a contract in India is 1,445 days, an invisible cost that shapes credit decisions, supply chain planning, and willingness to take risks. The expansion of commercial courts, the growing use of online dispute resolution, and the demonstrated effectiveness of specialised tribunals such as the NCLT all point to an emerging recognition that ease of justice and ease of business cannot be separated.

Speed has become a central theme of the reform agenda as delayed shipment and lingering permits depress investment, weaken confidence, and inflict more harm than any formal fee. India is increasingly moving toward tighter service timelines, T + 1 processing, and the idea that approvals should be automatic unless objections are explicitly raised. These shifts signal an administrative culture that is beginning to align with the tempo of a modern economy.

Digitisation is also evolving. The next stage will require re-engineering processes so that information need only be provided once and can then flow through the system without repeated input from citizens or firms. Countries such as Estonia have shown how explicit guarantees of data reuse transform interactions with the state. India is moving toward unified business identifiers, interoperable registries, and integrated digital platforms that set the foundation for similar transformation.

From Suspicion to Trust

A more forward-looking reform will involve treating regulation as a dynamic instrument rather than a permanent fixture. This will require the introduction of mandatory sunset clauses, with every new rule expiring after five years unless a clear cost–benefit assessment justifies its continuation. Such a discipline prevents the quiet accumulation of procedures that once served a purpose but no longer do, and it forces policymakers and administrators to revisit the logic underlying each regulatory demand.

Within this broader landscape, the Jan Vishwas framework marks a philosophical turning point. By shifting minor offenses from criminal penalties to civil fines, by requiring regulatory impact assessments, and by introducing risk-based inspections carried out by accredited third parties, Jan Vishwas moves India away from preemptive supervision and toward a system based on proportion, predictability, and trust.

An Architecture for Scale

India’s ambitions for the coming decades are considerable. The country aims to raise manufacturing from 15 percent to 25 percent of GDP and to grow its overall economy nearly tenfold in the next 20 years. Such goals require not only infrastructure and capital but an institutional environment that rewards initiative and enables scale.

The evolution of India’s regulatory system is still unfolding, yet its direction is increasingly clear. The country is moving from a supervisory framework shaped by the anxieties of an earlier era to an institutional architecture built for confidence, scale, and opportunity.

“AI threatens India only if you assume India is already a Western-style labor market. It isn’t.”

To accept this as India’s fate, one must first ask whether India actually operates within this framework, because for the vast majority of the country the boundaries between formal and informal work are flexible, porous, and constantly in flux. More than 80% of the workforce has no contract, no benefits, and no department to downsize, and people move between gigs and trades based on seasonality and opportunity (OECD average: 20%); rather than work being assigned from the top down, it is often assembled by workers and shaped by circumstance.

This dual system is visible almost everywhere. A FASTag scanner reads a license plate in a few seconds and deducts a toll through the country’s digital rails, yet at the same toll booth a man in a reflective vest sits in a plastic chair and writes license plate numbers by hand in a khata. The digital layer and the human layer operate side by side, and the system does not treat these overlaps as contradictions.

If one accepts this reality, it becomes clear that India’s operating environment is built on adaptability, and this is where things become interesting because AI is not confronting a neatly organized labor market. For most of the country, it is arriving in a landscape of overlapping systems, uncertain boundaries, and fluid roles, although some parts of this landscape are clearly vulnerable. The IT and BPO sectors grew out of a specific form of global cost arbitrage and directly employ over 5 million people, with millions more indirectly, and much of that work involves structured processes that AI can learn to perform. Similar pressures exist in professions that rely on routine analysis and predictable documentation, and teachers, accountants, paralegals, lawyers, and journalists will all face forms of disruption.

But the fundamental question for India is not how many jobs disappear but how work mutates. The country has lived through repeated shocks—from demonetization to GST to abrupt digital transitions and the pandemic—and each time it has reorganized itself rapidly; people found new forms of activity, new ways to earn, and new ways to navigate the altered landscape, and this capacity to reconfigure is deeply embedded in the country’s social and economic structure.

It is instructive to consider how global labor markets respond to technological change. Global labor spending, at roughly $50 trillion dollars, dwarfs global software spending at $300 billion dollars, and AI is entering the economy primarily as a mechanism to reduce labor costs. In countries where labor is expensive, the pressure to automate is immediate; Western economies face this acutely because human capital across the skill ladder is costly, and replacing people with machines becomes a rational choice.

India occupies a different position. Labor is inexpensive across factories, hospitals, services, and professional roles, and the economic incentive to automate is therefore far less forceful. Some work will shift, but the underlying calculation is different. The global cost arbitrage that powered India’s outsourcing boom will weaken in certain categories, especially lower-end tasks. About 3000 global capability centers employ nearly 2 million people, and parts of that activity will contract as AI becomes more capable.

Yet this is only a partial picture. More interesting is the emergence of new categories of “offshorable” work that did not exist previously. Radiology provides a clear example: the United States faces a severe shortage of radiologists and the gap widens each year. AI can now perform a highly accurate first pass on a scan, a technician in India can provide a confirmatory layer, and a licensed radiologist in the United States can then focus on the interpretive layer, performing more reads with greater precision. This is not a story of displacement but one of combining AI with human judgment across borders to produce faster, better, and more affordable care.

This first-pass and second-pass structure will not remain limited to radiology, as many regulated industries (think legal, compliance, accounting, risk underwriting) will adopt similar models; regulators will insist on human oversight even when AI provides the initial analysis, and India is the only country with the scale, expertise, and cost structure to operate this human layer at meaningful volume. What emerges is not the erosion of offshoring but a transformation of it; certain categories will shrink, but others will grow significantly, especially where AI augments rather than replaces human capability.

The long-term implications may be even more significant. A large share of future AI activity will be directed at tasks that were never handled by humans at all, and these tasks represent net-new categories of work that will require people who can supervise, guide, refine, and interpret AI outputs. India’s comfort with unstructured opportunity, its willingness to move across categories of work, and its capacity to recombine skills in real time create a meaningful advantage in this environment.

This intersects with India’s domestic economic story. The country is approaching an inflection point where per-capita income will cross the threshold that unleashes large-scale discretionary consumption; the top 200 million consumers already drive much of the country’s discretionary spending, while the next 500 hundred million are beginning to move from necessities toward healthcare, financial services, travel, housing, and digital goods. As this transition accelerates, it will create enormous demand for small businesses, freelancers, service providers, and new types of ventures.

India registered over 170,000 new companies last year, a dramatic shift in the country’s entrepreneurial landscape, and it now hosts one of the world’s largest communities of digital freelancers on global platforms. AI-enabled tools will become lighter, more accessible, and available in Indian languages, meaning that millions of people who were once peripheral to the formal economy will gain new forms of leverage. As consumption rises, the domestic market will become an even larger driver of employment; smaller teams will deliver high-quality services at scale, and solo entrepreneurs will take on tasks that previously required entire departments. AI will not simply replace work; it will enlarge the space in which work can take place.

Automation inside India will follow its own logic. Images of humanoid robots walking through Indian factories or public spaces make for striking illustrations, but they are unlikely to become common because the economics simply do not support them; a humanoid priced at $30,000 dollars is out of reach for nearly everyone. Where automation makes sense in India, it tends to arrive in simpler and more targeted forms, and this is already visible in manufacturing, where robotic arms operate efficiently at a fraction of the cost of a humanoid. From automobile plants to solar-module factories to large warehouses, automation is expanding in ways that match the realities of Indian labor markets.

The world is moving into a period where tasks will shift rapidly, careers will be episodic, and identities won’t be tied to a single form of work. In such a world, the traditional idea of a job matters less; what matters is the ability to adapt, learn, and shape opportunity as it appears.

India has lived in this reality for a long time.

Its informal intelligence may be its most valuable export in an AI world.

I wrote this piece along with Akhil Handa, a global C-suite financial services leader who is building the intelligence infrastructure to shape the future of SMB capital. You can follow him at akhilhanda.com

Building on my last post, I dug in further with my friend Jas, and to flip the above on its head, when Jas and I talk about technology in India we’re reminded how often the market mistakes silence for absence. In India, needs don’t always announce themselves in obvious ways, and when they do they might not be easily accessible in clean survey responses. More often they live in quiet workarounds (not jugaad) that resist neat classification. A lorry driver might ask his son to check diesel prices on an app but never download it himself, or a farmer might join a WhatsApp mandi group and lurk until the season turns. Teasing out a clean notion of consumer behavior isn’t always obvious.

The recent announcement that India will soon unveil its own sovereign AI foundational model brings that into the here and now. Trained on Indian datasets and hosted on Indian servers, it is being built as an attempt to encode the lived realities of languages, dialects, and the countless micro-patterns of social and economic life that global models fail to see. The hope is that it will allow India’s digital infrastructure to learn from the same quiet cues.

We’ll see how it does, but I am hopeful it will not make the mistake of building from the top down, with English wiring and for India 1. Something that understands fluidity and code-switching. Hinglish. Deccani. Surjapuri. Mashups of multiple languages and thought patterns in the same sentence. Something so fundamentally Indian that no one outside of India can actually imagine building in the first place.

Jas has been tracing distinctly Indian patterns for over a decade on his Digital Desh road trips, listening to how small businesses and farmers hint at what they want without saying it. The signals are there in abundance, but you have to know what to look for. “Senior, experienced people who can read these cues don’t have the time to travel!” he says when I ask him why this is not a more regular, mandatory thing to do among the investor classes. He believes India 1 is not tuned into Bharat, both both because it takes significant effort and because the results are too complex to easily interpret.

The “missed call” was a profoundly Indian phenomenon and still lives on in a post-Jio world. Beyond signaling “I’ve reached,” it might be used by a supplier to check stock or to tell a lorry to get moving. It’s still used by mega LPG companies to book your next cooking gas cylinder in Gurgaon. Although simple in theory, the underlying complexity of the “missed call” operating system cannot be grafted onto a standard product roadmap because each use case is idiosyncratic and deeply local. So how do you build for a market where the map is always in flux?

The question “how do you build trust at low cost?”—while well-intentioned—can often slip into lazy shorthand to explain why adoption lags. The missed call was an outcome of saving money while still signaling an important message or request (if you don’t pick up the call it’s zero cost, and India mass-adopted that shortcut). Sure, trust matters, but it isn’t the only switch. People adopt technology when it is useful and when it saves time, money, or effort. A UPI QR code isn’t a vote for believing in a bank as much as it’s about getting paid faster. WhatsApp didn’t win because it was trustworthy; it fit habits and had extraordinary utility. Do you “trust” WhatsApp? Probably not. In fact, their only mass media campaigns are around the fact that even WhatsApp cannot read your messages. Do you live in it like I do, regardless? Almost certainly if you’ve self-selected into reading this post.

UPI worked because it was simple, free, and solved a pain everyone felt: getting paid instantly. Aadhaar didn’t spread because rural India trusted the state; it grew because it unlocked bank accounts and SIM cards. (One can argue whether adoption was truly optional.) Each wave of technology has been about reducing friction and matching lived incentives. When it clicked, it clicked because someone made the new tool feel familiar while bringing utility.

Now that AI is everywhere, does Bharat itself need to believe in artificial intelligence, or should the goal be to make intelligence practical and desirable—to meet people where they are? There are grand solutions to be dreamt up, but there’s already so much being asked for, even if it doesn’t announce itself loudly. A shopkeeper screenshotting UPI inflows to track cash signals latent demand for better inventory and cash-flow tools. A dairy cooperative forwarding price charts on WhatsApp is, in effect, asking for predictive insights they can trust and act on. This is market research hiding in plain sight.

Listening well is hard. It means time on the road, bouncing through towns and junctions, giving up creature comforts for weeks, swimming in a sea so vast that full comprehension is likely never achieved. I’ve come to accept that understanding India’s full complexity is not unknown—it is unknowable, not only by me but by anyone. At best we form composite pictures, but we are always, without fail, the blindfolded man feeling his way around the elephant.

If you follow Digital Desh you’ll know Jas spends weeks watching who touches the smartphone, who borrows it, how questions get asked in a tea shop. Demand signals in deep India are often oblique: a question about diesel prices is really a question about working capital. Once you see that, the challenge becomes what to do with it.

AI won’t automatically solve this. Models trained on the public, English-speaking internet know very little about these small signals. Defaulting to urban patterns won’t get you far in a country that is both the second-largest English-speaking nation and one where 85 percent of people don’t actually speak it. Local context (language, transaction flows, social graphs, use cases) must be taught and curated, which means learning to listen at scale.

Incentives matter as much as trust. If an AI tool improves cash flow, a lender should reward that with cheaper working capital. If a farmer’s decisions become more accurate, it should reduce input risk and credit cost. Adoption accelerates when it ties to tangible economic upside, not just abstract efficiency.

Distribution, a perennial challenge in India, matters too. UPI’s story wasn’t just technology; it was policy and network effect, the same way WhatsApp didn’t need to be taught. For AI to diffuse, it will need similar piggybacking: chat interfaces in local scripts, open APIs for small developers to plug into payments and commerce rails, and foundational models tuned to regional realities.

The creation of a national LLM is a remarkable milestone and one that should be celebrated. But the real prize lies in what happens next: how quickly and seamlessly that intelligence travels outward, translating into tools that small businesses, farmers, and families can use with minimal friction. The brilliance won’t just be in the model itself, but in the quiet act of making it useful and of listening well enough to let intelligence feel local.

When I hear these, I wonder if the entire premise is misplaced. National pride aside, does it matter if the next great large AI model comes from Pune or Palo Alto? Sure, the former would be gratifying, but isn’t India’s most consequential and life-altering opportunity not in training massive models but in absorbing intelligence into the everyday lives of its 65 million small businesses—kiranas, workshops, traders, manufacturers, tailors, and countless others? Into the 150 million small farmers and cooperatives that would benefit from better access to weather, crop, soil, and livestock agents? Do they need the world’s most cutting-edge model, or something more practical: tools that make their purchasing decisions, cash flows, and sales opportunities sharper?

I recently debated this with my old friend Jasminder (Jas) Singh Gulati, whom I talk to regularly. He has been thinking about small business enablement far longer than most people I know. Years ago he built NowFloats as a turnkey way to give tiny shops a digital storefront, which Reliance Jio eventually acquired. Today Jas is deep into his long-running Digital Desh, an ongoing road journey into India’s heartland, listening to how people adopt technology, what enables and prohibits them, and what unlocks trust.

Our conversations wander from big themes—how UPI changed the arc of payments, how Aadhaar and KYC rails gave even the smallest merchants a digital foundation—to the small, lived, daily frictions he hears in taluks across Jharkhand. Shopkeepers running both digital and paper khatas side by side. Exporters stitching linens for overseas buyers who can accept foreign payments but can’t get a local bank to provide working capital. We have debated at length whether the transformative moves in India’s digital history were about invention or distribution and trust. Aadhaar, UPI, the India Stack weren’t necessarily deep-tech marvels by Valley standards, but they reshaped the country because of scale and reliability. And WhatsApp, not of Indian origin at all, became the operating system of commerce because it fit into existing habits.

So when the “Where is our DeepSeek moment?” question arises, we both tend to shrug. Maybe the better question is how to make AI feel like UPI or WhatsApp: invisible infrastructure that works in small, but familiar and obvious ways. Not an abstract breakthrough but a practical nudge that millions can use right now.

To borrow a term from other walks of Indian life, you can call it sacheted AI. India taught the world about the sachet, first with FMCG goods like shampoo, powders, and snacks, and then extended it to data, insurance, and anything small and affordable enough to try. Why shouldn’t intelligence come the same way? Small and contextual. A kirana store could have an agent that looks at its UPI inflows and whispers what to reorder before stockouts. A textile cluster could run a model tuned to shipping delays and overseas demand so a factory can take or decline an export order accordingly. A dairy farmer could get crop and price forecasts tied to weather and mandi data via a WhatsApp ping. None of this needs a giant, headline-making model. It needs local training data, clever packaging, and distribution along rails already built.

India is unusually well placed to do this. Because of UPI, nearly every merchant now has a digital footprint. Aadhaar simplified identity and onboarding while WhatsApp normalized conversational interfaces, while affordable smartphones and ubiquitous 4G did the rest. The next wave can build directly on rails people trust.

However, if the first wave of AI tools in India is locked inside walled gardens of a few platforms, we end up swapping one set of middlemen for another. If interfaces remain English-first or clunky, adoption will stall at the periphery. If privacy and data control are ignored, trust will evaporate. And as always, early adopters could cluster in the urban, English-speaking tier, leaving the vast majority of the country behind.

That said, it feels like time. In Jas’s latest Digital Desh work he writes about tiny merchants in Bihar who now see a smartphone as a ledger, a catalogue, a bank terminal. They already run order books on WhatsApp and collect via QR. They don’t need convincing that digital matters; they need help making better decisions with the data they’re generating.

What could help accelerate this shift? A few things: open APIs so anyone can plug lightweight models into payments and chat; local datasets that let models speak Marathi, Bhojpuri, Kodava; financing that rewards productive digital behavior (if your AI improves your cash flow, your working capital gets cheaper); above all, simple, human interfaces that make intuitive sense.

The prize isn’t a single eureka moment but something quieter: productivity gains diffused through millions of small nodes. A pharmacy shop that reorders smarter, a weaver who exports with more confidence, a food stall that manages seasonal demand. Global conversations about AI love clear symbols—“build our own GPT,” “outcompete DeepMind.” But India’s opportunity is significantly more interesting and profound. As it has done for centuries, India absorbs, adapts, and scales what matters. The country’s digital rails are already a kind of national unfair advantage. They can let intelligence seep down into the base of the pyramid faster than almost anywhere else.

When I talk to Jas about this, I’m reminded how unglamorous true change can look at first. UPI didn’t feel revolutionary in its first months. WhatsApp didn’t feel like infrastructure until suddenly every business depended on it. AI in India might arrive the same way: not with fanfare, but as a quiet new normal.

The crowd that self-selected in were those who wouldn’t eat a vegetable unless it was touching meat. Dal makhani might get a pass. Onion was acceptable, but washed down with mutton. Cucumbers slices ornamental. Naan had to be stuffed with keema. I grew up in the US, so obviously come from a meat-heavy world, but this was the deep end of the non-veg spectrum. Among this group, I often ended up being the most veg-friendly of the lot.

The green dot and red triangle divide the world pretty neatly. But that neat bifurcation can feel disconnected from daily life because so much of it plays out in the blurry middle. Even clear lines seem occasionally left open to interpretation. After all these years, I still don’t understand what a mushroom is, and I’ve seen it go both ways. It’s clearly not meat, but that doesn’t necessarily make it veg either. Some mysteries remain so.

A friend once told me about growing up in a joint family that occupied the ground and first floors of a house. On the ground floor, there was a firm belief that children need copious amounts of meat to avoid ending up stunted and left behind in maths. On the first floor, even garlic was forbidden from crossing the threshold. And yet, they lived happily amongst one another, maintaining theoretical boundaries that were occasionally crossed, although never openly acknowledged.

This led me to start noticing a broader taxonomy, the liminal spaces that exist between the green and the red. There are your standard vegetarians and non-vegetarians, of course. But then there are “Meat-Out-The-Home” types—MOTHs—who maintain a vegetarian household out of respect for tradition or kitchen sanctimony, but regularly head out for murgh malai on weekends. They’ve perfected the art of living with a foot in both worlds, and the kitchen stays cleaner on both a physical and metaphysical level.

Then there are the clandestine “Secret Meat Eaters”—SMEs—who maintain the outward appearance of strict vegetarianism but live in a gray zone that allows for the occasional exception—chicken paillard or sushi-grade sashimi. SMEs operate with plausible deniability and a carefully curated public image. I’ve shared prosciutto with people whose parents would identify their offspring as pure veg if interrogated.

I was once sitting in a dhaba and ordered dal tadka and subzi. The waiter looked at me for a moment and said, “No sir, you won’t like it.”

I asked why not.

“Because it’s vegetarian, sir.”

I told him I knew that, and that it was fine—I wanted veg. He nodded and walked off, clearly trying to process what had just happened. I know what he was thinking—that I likely eat meat, and since I can eat meat, why would I not eat meat? It glitched the system.

I will more often than not find myself in amusing discussions like the above because of my own peculiarities. For example, I will often tend, especially at lunch, to eat from the “starter’s” menu and in ways which primarily consist of “dry’s,” doing something like ordering a tandoori roti and a hariyali kebab. Upon doing something like this, the waiter will without fail look at me and say, “There’s no gravy, sir,” to which I’ll explain that I know and that I’m fashioning my own kind of kathi roll. This then will usually solicit a “What about the mains, sir?” to which I reply that this is the main. From here we usually go one of two ways—he’ll lean in and convince me that I need a “wet,” because this is all dry and therefore not tasty. Or he’ll just kind of let me be with my own idiosyncratic firang ways of bringing a bread order into the domain of the starters. Both are plausible, and neither is incorrect.

This scene also plays out in every darshini I’ve ever been to for breakfast as relates to the timing of the coffee, which I focus on first. If you order a dosa and a filter coffee, the coffee comes afterwards. However, if you order a dosa and a filter coffee and ask for the coffee to come first, the coffee still comes afterwards. Through many years of honed attempts, the secret is as follows: walk into a darshini and order a coffee. When asked if you would like anything else—khara bath, kesari bath, pongal, etc.—reply, “No thank you, only coffee.” Then wait for coffee to arrive and declare you’ve changed your mind, and proceed to order rava onion dosa, Guntur idli, uttapam, etc.

And while these are silly examples, they are allegories for life as we spend so much of our time beating our heads against the wall, trying to impose systems in Sysiphean fashion as opposed to simply reframing the problem. You don’t have to force the system to change its logic, you just have to learn to move with it, sideways, and find the workaround. The grand introduction to this for all expats in India to this notion is the FRRO office. Granted, things have become significantly easier, and much more has moved online and that’s almost a shame as it’s something I think that everyone should experience. There’s so much inherent wisdom to be gleaned from calmly navigating bureaucracy, it makes me almost miss the dread that would come over me when I was finally called to the window to produce something I hadn’t anticipated and that wasn’t even on the checklist.

There’s an entire genre of living in India that consists of gently subverting the system without anyone quite admitting it’s being subverted. Whether it’s mushrooms masquerading as meat, filter coffee arriving whenever it prefers, or the existential crisis caused by ordering dry starters as mains, the logic here doesn’t reward confrontation. It rewards style, patience, and the ability to smile while rerouting the entire operation sideways. It works in its own deeply elegant way, and even when it doesn’t you still get coffee eventually.

At the outset, he said he was grateful I had replied to his note and that he appreciated an uncle like me taking the time to meet him.

"An uncle!" I said, laughing out loud.

I asked him what exactly he thought the definition of an uncle was.

"Well, anything with a 4- in front of it," he replied.

I thought about that for a moment, and I couldn’t refute it. To a twenty-something, a four-handle must seem ancient. He wasn’t wrong. I’ve got a niece and nephew of my own, so technically I’m already an uncle. In my mind though, 'uncle-dom' was always a distant milestone—a state of mind, at the very least.

Now, it’s official. I am an Indian Uncle. So, if someone accuses me of giving gyaan, I can now steadfastly reply…well yes, that’s my job.

I knew this day would probably come, but I hadn’t thought it would be so soon. There are so many Uncle archetypes, and I wonder which one I will eventually settle into.

There’s the NRI Uncle, who’s lived in the Bay Area for 30 years. He visits India thrice a year and waxes eloquent about ancient Indian culture, but secretly likes the convenience of shopping at Sam’s Club in Sunnyvale.

There’s the Gymkhana Club Uncle, usually found on his perch with a Black Label and a bowl of boiled peanuts. He may occasionally change locations based on reciprocity.

There’s the Old-School-Cool Uncle, rocking a receding hairline that somehow comes out the other end as a ponytail. He looks weathered because he’s earned it (and that pack-a-day habit helps). He still listens to The Eagles and has spent more time in Ladakh than you have.

There’s the Fitness-Walking Uncle, always happy to meet a young entrepreneur—but only at Lalbagh at 5:30am. If you agree to that, he might just sponsor the idli-sambar afterward.

There are countless more, but which one am I? Perhaps a composite. I secretly like big box stores. I regularly engage in discussions about how Bangalore traffic wasn’t this bad 15 years ago. I experiment with Ayurvedic supplements and have read Outlive by Peter Attia. I check my step count by the hour. I forward WhatsApp messages. I will talk incessantly about how much I love my air fryer to anyone who will listen. Somehow, I now seem to be wearing the sneakers I bought only for running to more and more social occasions.

As I sat across from that young entrepreneur, I realized this wasn’t an insult, it was a recognition of great honor and the continuation of a noble tradition. As we wrapped up our meeting I thought about ordering a second filter coffee, but these days the caffeine hits differently.

Knowing I had a US passport, I think he wanted me there with him in hopes it might provide him a slight edge. Then came enthu cutlets, senti toppers, timepass gyaan, 3 Idiots, mugging up, chalta hai, arbit CP, fundas, faffing, chatur, Five Point Someone, backbenchers, chindi, bhindi, and that was just the first day.

So began my introduction to the Indian School of Business (ISB), where I showed up sight unseen in March of 2009. It’s a day I remember well for lots of reasons, and by chance I arrived on campus a few weeks early on what happened to be Holi. Although I had traveled across India for close to a year in 2006, I hadn’t passed through Hyderabad and even if I had, ISB was worlds away at the time. But here I was, going to business school in India, no turning back now. Having navigated from the airport to someplace called Gachibowli, I plopped my things down in a dorm room and set out on campus for the first time. I heard music and a lot of commotion off in the distance and I followed the trail to a block of dorms known as Student Village 3. There I stumbled into a scene of stunning chaos, and within minutes this oddball firang was covered in gulal with what was probably a doctored lassi.

I can state unequivocally that ISB was easily one of the most transformational periods of my life. Situated in what was then the fringes of Hyderabad, and what is now the heart of the city’s ‘Financial District’, it remains a special and nostalgic place. Having recently been back on campus for my 15-year reunion, I was reminded of the sentimental magic on those evenings when the warm wind whips through the Atrium rustling papers and occasionally sending the furniture skidding across the room.

Opening a bank account, getting a SIM card, navigating the intricacies of the FRRO, taking the shared auto from the main road into the city, the first morning wake-up from the campus peacocks, understanding the importance of birthdays in India, my first trips to Paradise, Shadab, and Charminar, all of it happened in the first few days. Not long after I’d come to appreciate why it helped to have a Marwari CA in my accounting study group.

There was a running joke around campus that, during my time on the trains, I’d probably seen more of India than a good portion of the batch combined. Whether that’s true or not, I don’t know, but I certainly didn’t understand India. Yet, I remember how many classmates asked to travel with me during our term breaks. They wanted to visit the places I’d gone—places they hadn’t really thought about—and it struck me then that in a country as vast and layered as India, you don’t have to stray far to be a tourist in your own land.

It’s true that we often take for granted what’s most familiar, and I’m certainly guilty of it myself. When something is part of the daily backdrop, it can feel ordinary even if it would be extraordinary to anyone seeing it for the first time. So many of my recollections and things that are etched in my memory aren’t necessarily profound, and they seem ridiculous now. I remember the first time many years earlier I saw bananas being cut off a stalk, because bananas didn’t grow on stalks, they grew in bunches of five in supermarket aisles with a sticker on them, or so I was led to believe as a child. Not knowing what should be embarrassing is freeing, and if nothing else it provided good fodder for my classmates who would often cast me in scenes for student projects: “Kutte kamine, tu mujhse bachkar nahi ja sakta. Main tera khoon pee jaoonga.”

There was a More store on the ISB campus that sold all kinds of sundries. I would stop in on occasion for a Thums Up, but I had a standing bulk order of Kinley soda and would have the store manager order cases of it for me every few weeks. I liked drinking seltzer, or sparkling water, on ice and would often go through a bottle or two a day, occasionally with a splash of kokum concentrate or mixed in with aam ras. One day I went to inquire about a couple of cases I had on order and the manager saw me and asked me to wait a moment. He then made a phone call and requested that I wait five more minutes. I poked around the store until one of the assistant deans of the school came in to greet me.

He pulled me aside: “Thomas we’d like to speak with you, we’re very worried about you.”
“Yes sir, happy to speak, worried about what?”
“The store has informed us that you are ordering cases of Kinley on a regular basis.”
“Well, yes.”
“Thomas, we are worried about your drinking problem.”
“Sir, what do you mean? I have barely been drinking at all this year.”
“Well then what are you doing with this much Kinley?”
“I drink it, sometimes with juice.”

He looked at me wanting to believe that I was in fact not putting away a few bottles of Old Monk and Royal Stag every night, and after some more back and forth and a laugh, I think that he finally did. I think about this story a lot, because while it’s funny and I found it quite endearing, it’s been a constant reminder of how we can reach two wildly different conclusions about something so seemingly innocuous depending on the starting frame. What felt so routine to me was simply inconceivable to everyone else, and that learned experience was worth its weight in gold.

Once the first term got underway, I ran uncontested for the student body position of Head of the Food & Beverage Committee, which probably should have been a fairly obvious signal as to why no one else wanted the job. I would receive dozens of emails a week about canteen grievances: the food is too spicy, not spicy enough, too oily, too repetitive, too much bitter gourd, not enough veg, not enough non-veg, too much biryani, we should have biryani more often, for what we’re paying we should get three rotis instead of two, and so on. This topic, as you know, runs deep, and every month I’d take my notebook with collected suggestions to meet with the catering department to try to make sense of the seemingly contradictory comments. The one that was pretty consistent was that people wanted an easier, quicker lunch option, and while I am hesitant to celebrate my achievements, for this one I do take credit for as an enduring contribution to ISB’s culinary legacy—the mini meal. We turned the small side canteen that largely sold puffs and samosas into a meals joint, and we bought the compartmentalized trays and rolled them out as a trial. Two rotis, rice, papad, dal and veg, pickle and curd for fifty bucks. What a legacy to have left behind.

I arrived at ISB with the vague notion of building a career in India but wasn’t sure exactly how I would get there. Somewhere along the way, what once felt foreign became familiar and then more like home to me than anywhere else. If there is a lesson that has lasted, it is that understanding comes from showing up, getting it all wrong, and letting the world nudge you in ways that don’t feel obvious at first. There are few things that feel as comforting to me as landing at Shamshabad, driving past the National Fisheries Development Board (which is itself shaped like a fish), and arriving back on campus to while away time in the Atrium as the warm breeze whips through and the placards go flying.

Ghar lautna, khud se milne jaisa hota hai.

Upon closer inspection, it turned out that Amul—or more precisely, the Gujarat Cooperative Milk Marketing Federation (GCMMF)—had partnered with a cooperative in Michigan to launch a fresh milk brand in the U.S. directed squarely at the NRI community, with a flagship high-fat 6% milk that lends itself to making desserts and ghee. (American whole milk is generally around 3.25%). The choice of partnership made sense as Verghese Kurien* had studied in the state, graduating with a master’s in mechanical engineering from Michigan State University in 1947, before returning to India to create what is arguably one of the most consequential livelihood organizations in modern history. He took India from milk deficiency to the largest dairy market in the world and transformed milk from a scarce luxury into a daily, affordable, nutrient-rich staple. But Amul setting up in the US wasn’t just a diaspora gesture, it marked a move from exporting labor and raw goods to exporting finished products, complete with their own operating systems.

Some have succeeded, some have faltered, but they’re all part of a shift in India’s outward posture. Haldiram’s has gone into more mainstream supermarkets and Sula turns up on the occasional restaurant menu. Tanishq has recently opened seven stores across the country. In 2017, Bira 91 was the official beer sponsor of the Tribeca Film Festival in New York, although the brand eventually refocused back on India. Somehow, even after all these years, FabIndia couldn’t get it right, resisting e-commerce even in India.

There are countless examples of Indian companies, methods, or craftsmanship quietly powering the world. Tata’s ownership of Jaguar and Land Rover has revived the brands.** Ashok Leyland supplies to London’s iconic red buses. In cities across the U.S., if you look down, you’ll see “Made in India” stamped across manhole and drain covers. Ninety percent of the world’s diamonds are polished in Gujarat. Over thirty percent of the U.S. generic pharmaceutical market comes from India, along with spices, textiles, and teas. Premium shoe brands that may be “finished” in London or Florence are often structurally Indian underneath it all. Often though these have remained obscured and white-labeled for larger brands.

There are now thousands of Indian SaaS companies plying the corridor, marketing and selling into the U.S., competing on better technology, better pricing, a more unified customer experience, or some combination of all three. Meanwhile, Indian IT services giants like TCS, Wipro, and Infosys quietly power the back-end office systems that keep much of the world running.

One place where America could really use a dose of Indian pragmatism is healthcare. While Narayana Health and Aravind Eye Care deliver complex world-class surgeries at scale at a fraction of the cost, the U.S. system is bloated, predatory, and universally despised. I recently called my doctor for a routine appointment and was told the scheduler would call me back—in six weeks—to book something three months out. On another occasion, I was charged $400 for a ten-minute telehealth visit, despite having insurance. I needed a shot of Amul after that one.

What’s missing across much of the American consumer experience isn’t just quality, it’s consistency and the ability to withstand stress. Maybe it’s time for Indian companies to stop white-labeling and show up with their full playbooks.

Should IndiGo come and rescue flailing JetBlue and Southwest, and just get back to basics? (The Nut Case would do really well too.)

Could Jio disrupt the cable and mobile mafia and finally break the bundle for good? Imagine $10 fiber blanketed across rural America.

Can the UPI come and clean up Venmo, PayPal, and the extractive 1970s-era credit card networks?

Can GVK or GMR come and build airports for the 21st century?

Can Indian Railways step in and show Amtrak how to run a railroad? Their tagline could be: "Literally, anything would be better." And they would be right.

Maybe lending and credit would benefit from SHG dynamics—more reliance on social collateral than rigid, bank-led underwriting criteria. Community accountability could rebuild trust in broken U.S. neighborhoods.

How about DigiYatra moving people through airports at scale and putting CLEAR out of business in the process?

And yoga, wow does yoga need a bailout given its been incarnated as $200 leggings, pithy slogans, paddleboards, and a gong ceremony.

And Saravana Bhavan drive throughs. America would run much better on filter coffee and idli.

Some of these examples are funny, but not entirely a joke. As Indian companies have sharpened themselves against the scale challenges of the home market, they have built both resilience and pragmatism that feels increasingly rare in many parts of the US. Indian brands can succeed in America, and the country needs them much more than it realizes.

-------

*As an aside, I have never forgotten an essay that Manu Joseph wrote about Verghese Kurien at the time of his passing in 2012. It stood out because it captured not just the scale of Kurien’s achievements, but the small human details (the stubbornness and the wit) without trying to smooth them over. It was the story of a man who stayed, built, and fought quietly and a rare piece of writing: clear, sharp, affectionate, and honest that struck me at the time.

I would tell Manu a few years later, upon meeting him in Delhi, that the essay contained what I thought was one of the best lines I had ever read:

“He was a man who saw the world as a conflict between the clever and the foolish, and took the side of both to push his plans through.”

Link to the essay

** Given that Tata Sons is majority-owned by the Tata Trusts, the profits from those companies ultimately help fund education, healthcare, and other initiatives back in India. What was once extracted now flows in the other direction.

Today, we live in an age of satellite imagery and real-time GPS. We can zoom in on license plates from space and estimate travel times down to the minute. And yet, in the ways that matter most, much of the world is still regarded as a dragon’s lair—not because it’s physically unmapped (it’s often rendered in extraordinary detail), but because it remains unexamined, misread, or ignored altogether.

I can take a street view of a town in rural India and see a snapshot of its physical markers: kirana shops, roads, multi-story buildings. But how do I peer into these places beyond the physical and observe transactions, livelihood-generating opportunities, and intention to repay loans? If I’m looking at alternate data sources, perhaps I could use satellite imagery to show how much of the town is lit by night or crop patterns to gauge a harvest. We could look at patterns in COD e-commerce orders, data on ATM restocking frequency, or IRCTC ticketing trends to track seasonal migration. These, and numerous other proxies, can help triangulate reality on the ground. But how would capital allocation know to assume there are dragons out there—unless someone has actually gone and looked? These places are, for all practical purposes, contextually unmapped.

Every year, my friends at Blume publish the Indus Valley Annual Report, a comprehensive review of the Indian startup ecosystem. It looks back at notable financings and themes and offers a forward-looking view of where the market might be headed. At nearly 200 pages, it is data-rich and widely appreciated across the Indian venture ecosystem.

Page 76 of this year’s report jumped out at me. It divides India into three consumer segments:
• India 1 ("Mexico"): 30 million households (140 million people) with per capita income around $15,000. The true consuming class, and the core market for most startups.

• India 2 ("Indonesia"): 70 million households (300 million people) with per capita income around $3,000. Aspirational consumers—heavy app users, smaller ticket spenders.

• India 3 ("Sub-Saharan Africa"): 205 million households (one billion people) with per capita income around $1,000. Largely considered unmonetizable today.

This is the market as it appears through conventional economic lenses, and from the perspective of global capital allocation—based on how the narrative is generally portrayed—I don’t think they are wrong. Because of how we assess GDP per capita and easily measurable discretionary income and employment, this is how the market in India presents itself. Even if the real addressable market for venture is only the size of one Mexico, it’s still a growing (and deepening, as Blume notes elsewhere in the report) underserved opportunity for both consumer discretionary and basic spending.

There’s an inherent discrepancy here though, which we’ll call the total addressable narrative vs. the actual addressable capital deployment. On the one hand, if you take the view that the market for venture is actually only one Mexico, and maybe some sliver of an Indonesia, then you have something significantly smaller than an India-sized opportunity to put money to work with the hopes of generating outsized returns. At the same time, most of the investment narrative in India builds using the following phrases, or offshoots thereof: “demographic dividend,” “rising middle class,” “one billion aspirations,” “Bharat” as a theme, and so forth. So which one is it? The market appears to be advertising the hopes and dreams of 15% of the world’s population but only channeling money toward 15% of that population. (8 billion × 15% × 15% = 180 million people, or one Mexico and a little bit of Indonesia).

While we have outwardly objective, statistically measured indicators to quantify rural vs. urban, formal vs. informal, employment data, GDP per capita, income and expenditure, they don’t stand up all that well to more complex definitions. Depending on the frame, you can arrive at wildly different conclusions. We know that GDP per capita in the country hovers around $2,800, but if I am managing a VC fund investing in India 1, this is largely meaningless as a picture of everyday reality, and that number itself is significantly dragged up by a relatively thin sliver of the population. Thinking in medians and modes gives a better picture of what we're working with more broadly across the country (also the topic of another post).

The Blume slide makes its argument cleanly: the lower echelons of India 2 and certainly 3 are unmonetizable. But is that a fixed truth? Or is it a reflection of how narrowly capital markets have defined what counts as monetizable?

I was listening to a podcast recently with Neil deGrasse Tyson—astrophysicist, Renaissance man, polymath extraordinaire. The topic was whether we’re alone in the universe—the answer, probably not. But the more fundamental question he raised was how would we even recognize intelligence if we encountered it? Do we even have the capacity?

Tyson pointed out that humans share 98.8% of our genetic material with chimpanzees. It’s in that narrow 1.2% gap that we go from poking termite holes with twigs looking for lunch to launching satellites, composing symphonies, and posting latte foam art on Instagram. Now imagine life that’s not 1.2% smarter than us but 3%, or even 10%. Perhaps the reason for Fermi’s Paradox (the contradiction between the likelihood of extraterrestrial life and the total absence of contact) isn’t that they’re not out there. Maybe they actually came, took a quick look, and found us about as fascinating as we find a pile of leaves, and moved on because they didn’t think we were worth interrupting.

His point is that we are only capable of looking for what we already recognize. When we imagine aliens, we picture something vaguely humanoid—upright on two feet, breathing oxygen, curious about us, something we’d be able to communicate with. But what if intelligence, in its more advanced forms, isn’t immediately recognizable?*

If you’ll allow me a small leap, I see that same bias playing out in how the West, and specifically global capital, looks at much of the world. It misses the march of time that has developed an intelligence that should be worth promoting but instead is viewed largely as a no-go zone. India is an old place, and it is the wisest place I know—if you want to confirm that for yourself, just go sit under an ancient banyan tree for an hour. It’s systems (social, moral, economic) weren’t necessarily engineered to meet the immediate needs of short-term gain, but they have endured, and they scale while also being optimized for externalities. By contrast, much of the modern West is still a well-capitalized startup in search of long-term resilience and meaning. As one ages, one hopes that humility sets in. In this case, a refresher in respect for your elders is probably in order.

If you know where to look you’ll find many promising business models out there that service erratic cash flows through income-linked repayment structures, rather than rigid EMIs. Some provide productive asset financing to rural and peri-urban households—not for phones, but for milling machines, irrigation equipment, or secondhand vehicles that generate income. Others are building digital credit rails around informal savings groups, not just banked individuals—taking the logic of JLGs and adapting it to highly informal communities. There are also embedded finance models within supply chains that finance inventory for kirana stores or offer working capital to small vendors based on GST data, UPI transactions, and even offline purchase behavior.

There is tremendous economic activity where the dragons are purported to live, but it’s fragmented, non-linear, and often invisible to traditional underwriting. I keep coming back to that spotlight diagram: a narrow beam concentrated on a bright corner of a much larger landscape. The beam illuminates the familiar ground of structure and predictability, but just beyond its reach the light fades. The rest of the world does not actually disappear when we do this—it just simply becomes easier to ignore.

*Letting go of ego should, in theory, free one from embarrassment. But in this case...I spent part of my weekend watching Indiana Jones and the Kingdom of the Crystal Skull—the fourth in what was supposed to be a trilogy.

I can’t imagine I need a spoiler alert here, but at the end of the film there’s a scene where the alien beings (upright, vaguely humanoid, surprisingly familiar) depart Earth in their craft, leaving behind a trove of treasure buried in rubble.

One character asks Indiana Jones, “Where did they go? Did they go back to space?”

Indy replies, “Not into space. Into the space between spaces.”

Someone else asks, “What about the treasure?”

Indy answers: “The treasure wasn’t gold...it was knowledge.”

Take from that what you will. At the very least, now you don’t have to sit through the rest of the movie.

I arrived at the old airport in Kolkata without much of a plan. I had been traveling through Southeast Asia and ended up in Bangkok, where I searched for the cheapest flight to anywhere in India. I found one on Drukair with a stop in Kolkata on its way to Bhutan, which seemed as good a starting point as any. At the time, I had no real sense of the difference between Kanpur, Kozhikode, or Kolkata—but the fare was cheap and the route was short, so I started there.

By then I had already been on the road for nearly a year. I’d taken many train journeys across Asia, including the Trans-Siberian from China through Mongolia and west to Moscow. I’d hitchhiked, motorbiked, ridden in shared autos and trucks, taken countless buses, and even stumbled into a brief internship with the US State Department in northern Laos. I’d slept in hammocks, on a sheep farm, train station benches, and rooftops.

After about a year of this, I remember thinking to myself that I had a reasonably good understanding of how to figure things out on the road, and that I could find my way into situations and get myself out of a box. Like when I entered Laos from Vietnam at one of the most out of the way crossing points, an exercise that culminated in my passport eventually locked in a desk by a Vietnamese border agent.

To reach many hub destinations, one could usually take an express bus or piece together a series of local buses and small transports, and I would often choose the latter. I had nowhere to be and I would see new things, although I was also trying to save as much money as I could.

Imposing constraints—five dollars today for all my meals (and ideally a place to sleep)—forced me to explore narrow lanes and figure out how to order food in small stalls and find a bathroom. On multiple occasions, across China and Vietnam, I recall very patient restaurant owners trying to figure out what I wanted to eat. I would often just end up back in the kitchen where I would point to a fish, some vegetables, and the wok, and then make some sort of mixed-up commotion with my hands. They would laugh and I would end up with a stir-fry or fried rice. It wasn’t that hard.

But back to India.

Landing in West Bengal was something else entirely, even before I made it out of the airport. India was more, and although I didn’t know exactly why, I did know that everything else in my life until then had felt like the dull edge of a knife. The fact that this place even existed, and had been there all along, only made me realize how profoundly naïve I’d been all my life. It was one of a countless stream of lessons in humility to come.

There was a little information kiosk near baggage claim where I inquired about finding a place to stay for a few days. Sizing me up as the shaggy-haired, and generally clueless backpacker that I was, the woman told me to find Sudder Street. I wandered out into the cacophony of Ambassadors for my first auto negotiation and soon ended up at one of those classic backpacker crossroads that I had both been in and tried to avoid across Asia.

These places often resemble the cantina scene from Star Wars. They were the old circuits where travelers of all stripes—drifters, grifters, bongo players, and long-term wanderers historically gathered, cheap and loose, creating their own strange little worlds inside the bigger local culture. You could recognize them by the endless array of signs advertising banana pancakes, chakra readings, money changing, cheap massages, or overnight buses to somewhere even more far-flung. There are some notorious ones: Khao San Road in Bangkok, Pai in northern Thailand, Vang Vieng in Laos, Hampi, Rishikesh, Anjuna Beach and so forth. Sudder Street fit the bill.

One of the things that struck me about India early on was that, as large and intense as it was, I could get by with English in a way that I couldn’t across China and Vietnam. In those places, there was a foreignness compounded by the language barrier, and it made it obvious that these places were not accessible. India, by contrast, was equally as foreign, but there was this strange familiarity that made it both seemingly accessible and distant. Granted, I was very much in the middle of English-speaking India, and I have since spent plenty of time in non-English-speaking India—the obvious majority of the country—but at the time, in my mid-20s, this contrast both perplexed and allured me.

I came across Egg Devil, “chops”, and something called a cutlet that was vaguely familiar. After a few days of kulfi and kathi rolls and jhalmuri, I found my way to Howrah Junction, and its endless crossroads of porters, platforms, and all kinds of travelers coming and going. I remembered standing in massive Chinese train stations, staring up at flickering boards of Chinese characters, laughing to myself, and then thinking “hmm...how am I going to get out of this one?” Howrah was no less sprawling, but it was both easier and harder. It had Tatkal and clear signage, but I still had no idea where to queue or how the platforms worked. It was in those early contradictions and odd juxtapositions that things slowly began to settle in.

On the rails for the first time, sleeper class trains carried me across that first stretch of the country, and many others after that. I had no itinerary, no real sense of how far I was going, or what would happen when I got there. This was before smartphones, and I really didn’t do much planning, relying on instinct and surrendering to the universe. I still have this backpacker’s hangover in that I like to show up to places without a hotel booking and just figure it out, granted technology has largely made this a non-issue.

When everything is new, even the mundane is novel. Things you would normally take for granted and not even notice during the day-to-day of life suddenly become things to ponder. The cool night air on the berths, the world outside blurring by in a cacophony of sounds I had never heard before. Voices, languages, the chaiwallah making his way up the aisle with his distinctive ‘chai garam, chai garam’ intonation. Train whistles, the caws of unfamiliar birds, curious questions about why I wasn’t in AC 1, and offerings of home-cooked meals from stacked tiffin boxes that I had never seen before.

The first time someone asked me if I wanted chaas, I said, “What’s chaas?”

“It’s buttermilk”

And having only a vague notion that buttermilk was something you used to bake a muffin, I remember thinking, what am I going to do with buttermilk right now?

There were periods of long silence, as there always are on the train. I learned the rhythm of the stops, five minutes here to hop out and buy a bottle of water and a samosa, a few there to stretch my legs and find some biscuits. Trusting the rails to open things up for me one station at a time, I had never felt so unmoored and I had never felt so alive before.

It was only much later that I understood the full scale of what I was moving through. Indian Railways moves 23 million passengers a day (more than the entire population of Australia) across 68,000 kilometers of track and to more than 7,300 stations. Only armies and supermarket chains employ more people, and none carry a country forward in quite the same way.

It was never designed as a single system. It grew outward and sideways, a federation that came together by folding its many contradictions and idiosyncrasies into itself. It continues to expand and decommission, combining bygone relics with the latest Japanese technology in permutations and combinations that would make little sense anywhere else on earth. It continues to build around what exists.

About a month ago I took a breezy 11-hour ride through the Western Ghats from Bangalore to Mangalore in the Vistadome, with its expansive panoramic windows and swivel chairs. Although high-speed trains are now compressing journeys that once took days, the essential nature of Indian Railways endures. You can ride a Shatabdi Express with a hot meal delivered to your AC car; yet just a track away, a slow second-class sleeper rattles along carrying families, students, workers, and maybe even a clueless backpacker with a Tatkal ticket and no destination. These are the foundational arteries of the country, and difficult to describe unless experienced.

I have learned, unlearned, reassessed, and tried to make sense of so much over the years. Most of what I thought I knew has shifted, expanded, or been discarded altogether. But there is one conclusion I reached early on that still holds true — maybe even more so today: the only real way to see India is by hanging out of the side of a train car, watching the world roll by and letting it shape your perspective, one chai garam at a time.

This wasn’t just a scientific curiosity; it was a reminder that just because we don’t see something doesn’t mean it’s not real. Our understanding of the world is shaped by the limits of language, perception, social norms, coping mechanisms, and inherited frameworks. And often, those limits can obscure things that are hiding in plain sight.

Take urbanization in India.

India’s urbanization is unfolding differently from the neat, linear progression you might read about in textbooks. According to conventional Western industrial history, the model is simple and straightforward: rural villagers move to cities, integrate into structured workforces, and become part of the consumption economy. Anyone who’s traveled to India’s edges knows that this story doesn’t wholly apply.

To start with, migration in India is not just a one-time shift from rural to urban areas; it is a natural, seasonal ebb and flow that mirrors agricultural cycles, festivals, and labor demands. Consider the kharif, rabi, and monsoon seasons, which dictate the planting and harvesting cycles for crops like rice and wheat. During these peak periods, there is a significant demand for agricultural laborers in specific regions, drawing individuals from their home villages in search of work. As the harvest concludes and agricultural activity slows, many of these laborers return to their native places with their earnings. This cyclical movement is also intertwined with social and cultural life: festivals, marriages, and family obligations that often serve as compelling reasons for temporary return migration. Furthermore, the informal economy in urban and peri-urban centers offers temporary employment opportunities in construction or small-scale manufacturing, providing a vital supplementary income for those whose livelihoods are primarily tied to the land. This constant ebb and flow of people is a pragmatic adaptation to the economic and environmental realities that shape livelihoods across India.

I’ve traveled across these places and seen firsthand how this works. This movement isn’t simply about moving from one place to another, it's about shifting in and out as conditions change. This fluidity gives rise to an economic phenomenon that doesn’t fit neatly into the traditional rural-urban dichotomy. It’s the rise of peri-urban India.

Peri-urban is the in-between—not quite city, but no longer rural. These are messy, vibrant zones of economic activity shaped by migration, logistics, and affordability. Think Bhiwandi, Sanand, Hosur, Ranchi’s outer ring, Coimbatore’s industrial sprawl, the outskirts of Lucknow, or the Gurgaon-Manesar-Dharuhera corridor. None of them feature in tourist brochures and many startups aren’t keen to serve them, yet all are critical nodes in the country’s supply chains and labor networks. Take a place like Ghaziabad as a prime illustration. On paper, it’s a municipal corporation. In practice, it’s a fragmented collection of peri-urban clusters—itself straddling the edge of Delhi’s NCR but never quite inside it.

While India’s official urbanization rate is pegged at around 36% (World Bank, 2022), studies from UN-Habitat and others suggest the real urban footprint could be closer to 50% when you account for electrification, construction activity, and migration flows. That missing 200 million people—or one Brazil—aren’t in cities or villages. They’re in-between.

These in-between zones are full of contradictions. Construction is booming but permits and zoning are nonexistent. Online shopping is common, yet addresses are often improvised based on landmarks. Apps deliver Limca but piped water may be absent. Work is abundant but workers are invisible to labor statistics. Industrial sheds sit next to farmland, and informal settlements abut logistics hubs. Autos navigate construction that spills out into the streets and water tankers slosh along. In local markets, vendors sell both mobile phone accessories and fresh vegetables for the evening’s subzi. The lives of residents are similarly hybrid; a young person might commute to a tech job in the city center but return home to a family deeply rooted in tradition and community. This in-between is a dynamic collision of old and new, where the conveniences of urban life are often accompanied by the challenges of inadequate infrastructure and layered with deeply ingrained rural customs.

Are these places just waypoints enroute to some formal urban destination—or are they the destination? The idea of a polished, linear rural-to-urban trajectory doesn’t hold here. Yet India’s governance, credit systems, housing models, healthcare delivery, and even data collection remain oriented toward neat categories—urban or rural, formal or informal.

A Problem of Definitions

The Census of India’s definitions compound the invisibility. Urban is classified in one of three ways:

• Statutory Towns: Areas with a local municipal body (municipality, corporation, cantonment board).

• Census Towns: Areas that meet all three criteria:

  1. A population of at least 5,000

  2. At least 75% of the male workforce in non-agricultural employment

  3. A population density of at least 400 persons per sq km

• Urban Agglomerations: A continuous spread of a core urban area and its adjoining outgrowths, totaling at least 20,000 people.

And rural? It’s everything else, meaning Rural = Not Urban

This binary classification leaves no space for the in-between. It erases places with urban character but rural classification. It ignores the flux of population, the decline of agriculture, and the shift toward services and industry. These areas are often “stuck” administratively, missing out on infrastructure, planning, and visibility, even as they function like small cities.

A Side of Peri-Peri

Peri-urban India falls through the cracks. These areas lie outside municipal boundaries, so they receive no urban services while at the same time remain excluded from rural schemes. Yet, these engines of industry and consumption are home to over 600 million people. Gaps in classification have real-world consequences. Fast-growing zones remain rural on paper and therefore miss out on investment and infrastructure. No one is quite responsible—urban governance is too weak, rural governance too rigid. Some areas deliberately stay rural to access subsidies or avoid taxes, sacrificing long-term infrastructure for short-term gain. And investors, clinging to the usual suspects, overlook these hubs and miss the steady, long-term growth bubbling under the surface.

You don’t need to go far to feel the movement and energy in these transitional zones, and nowhere is the Olo reality more visible than on the outskirts of Bangalore. From Devanahalli’s luxury real estate around the airport to Hoskote’s industrial corridors lined with paying guest hostels (PGs), these areas look urban, act urban, but remain administratively rural. Anekal, on the city’s southern fringe, encompasses parts of Electronic City yet relies on panchayat services. There’s no formal signpost that tells you that you’ve crossed from rural to urban and back.

And that’s the thing—most reports I read or investor conferences I go to still portray India through outdated binaries. India 1, 2, 3. Rural vs. Urban. Formal vs. Informal. Developed vs. Emerging. But life doesn’t work in binaries here. Honestly, it doesn’t anywhere—India’s just more upfront about it.

While the major metros are undoubtedly charging ahead at breakneck speed, with a proliferation of fine dining, coffee shops, breweries, bakeries, and creature comforts beyond imagination, the next India isn’t taking shape among the usual suspects. It is coming together at the edges — on the fringes of cities, in the spaces that were never planned but have grown anyway. In half-finished industrial parks next to fields and truck depots, and in bazaars where cash auctions and QR codes coexist. Away from the spotlight and in the big middle—that’s where the action is.

Can you spot the Olo?

*Maybe you noticed that “olo” is hiding in plain sight—the middle three letters of “color.” It’s been there all along, just waiting for someone to see it.

India is formalizing. And yet, it remains profoundly informal. We toss around terms like “formality” and “informality” to describe economic activity—but do we even know what they really mean? Over 85% of India’s workforce is considered “informally employed”, but when something accounts for the overwhelming majority of how a system functions, maybe it’s not the exception. So why does formality matter? And more importantly, is it even definable?

In the Western lexicon, “formality” implies structure, regulation, and legitimacy. It’s framed as a kind of progress—a journey from chaos to order, from the unorganized to the organized. “Informality,” by contrast, is seen as a deviation: temporary and messy, impoverished and backward—in need of an upgrade. That’s a narrow, very linear (and limited) view of the world. Who promised linearity in the first place?

Is a business formal because it has a GST number? Because it pays taxes? Uses a QR code? Ola is formal—but what about its drivers? Like gig workers everywhere, they exist in a gray zone: not quite employees, not quite independent. But in India, that ambiguity is layered over a labor market where protections are scarce and informality is the norm. A kirana shop might have a POS terminal, yet still run on local credit and trust. A factory might have all its licenses and still pay workers off the books. Formality feels less like a binary classification and more like a spectrum—a patchwork stitched together in the margins. It’s not about structure so much as visibility to the state, and visibility doesn’t always mean protection, stability, or legitimacy. That’s the paradox: what appears formal on paper may not be in practice, and what seems informal might actually function more effectively than we expect.

So what does "formalization" actually mean? Does it mean eating dosa with a knife and fork?

The Western lens often tries to impose linearity and familiarity onto systems that are fundamentally non-linear—systems that loop and adapt as a matter of design. When these systems resist being easily mapped or measured, they’re often misread as dysfunctional.

If we can accept that India is a place that is at once ancient and modern, improvisational yet precise, disparate and cohesive, then perhaps its informality isn’t a flaw to be fixed and it’s the very quality that makes it work. That doesn’t mean systems can’t or shouldn’t evolve—they can and they will—but there’s a difference between fixing something as opposed to improving it, and that’s a nuanced distinction that often gets lost. And while I wholly cringe at the cliché, it fits here: informality is a feature, not a bug.

To developmental economists, "informality" often feels like a dirty word. But informal systems—whether social, economic, or digital—aren’t primitive, nor are they necessarily temporary. They’ve been optimized over time to handle volatility and risk.

Nassim Taleb would call these systems antifragile: not merely robust in the face of shocks but actually improved by disorder. They absorb stress, reroute around disruption, learn, and recalibrate. As opposed to completely breaking they bend, flex, and can often emerge stronger and wiser in the face of a monsoon, a supply chain crisis, or a pandemic lockdown. Systems built on trust, redundancy, and adaptation snap into motion without waiting for central-level policy memos. In a world that faces such great uncertainty on every front imaginable, which societies will be best prepared not only to handle external shocks but also to respond and adapt to them before they reach a breaking point?

Visit any wholesale mandi or bazaar in Surat, Madurai, or Rajkot and you’ll see intricate networks and transactions governed not by paperwork, but by trust and relationships. Goods move in bulk without invoices. Prices adjust on the phone in real time. Credit is extended based on a nod. Logistics are coordinated via WhatsApp, not ERPs. Despite the ubiquity of digitization, it isn’t creating order in the Western sense, it’s actually supercharging informality. WhatsApp is the CRM. YouTube is where skills are learned and taught. UPI functions like a digital khata. Digitization here doesn’t conform to traditional ideas of structure—it strengthens decentralized, intuitive, and informal systems.

This isn’t dysfunction. It’s what Dee Hock—founder of Visa—describes as a chaordic system: part chaos, part order. These systems aren’t centrally controlled or totally anarchic—they’re alive with intelligence at the edge. They scale without hierarchy and operate without top-down control.

However well-functioning these systems are, they face a threat to their continued existence: the danger of over-monetization. Hock believed that by reducing everything to a monetary value, we destroy value. Not just economic value, but human and community value. When goods are extended on credit and services are bartered the relationship, not the contract, underpins trust. These are non-monetary exchanges of value—and Hock believed they were essential to a functioning society—try monetizing every relationship in your life and see how fast they break.

What happens when we measure worth only in Rupees? We lose sight of the community capital, natural capital, and social resilience that can’t be priced and we start assuming that value only exists when it's visible to capital providers. We begin designing systems that erase—rather than enhance—what already works. Systems that are not easily quantified and that don’t show up in dashboards are seen as flawed, risky, and therefore unpriceable by investors.

India’s informal systems haven’t endured for centuries because they were waiting to be fixed. They’ve endured because they solve problems and adapt to complexity instead of resisting it, and they do so without reducing people to economic units or commodifying every transaction. In Hock’s view, monetized systems risk making people invisible—seen only as means to an end, not ends in themselves. In India’s informal economy, people remain visible through relationships, reputation, and reciprocity—just look at how microfinance joint liability groups (JLGs) underwrite based on social capital.

Anyone trying to develop a thesis on India has to start with this understanding. Non-linearity isn’t the exception—it’s the rule. Humility is required to interpret it. Western frameworks prize legibility, scale, and order, so investors seek solutions that impose order—but by favoring only what looks familiar we miss the action in the margins. This distorts how capital gets allocated, and we chase formality not because it’s more valuable, but because it’s more visible. Progress therefore is subjective; in lending, for instance, it often looks like pulling people into formal frameworks that were never designed to underwrite them in the first place.

Countries are often compared to one another based on perceived stages of development, but India is not a “failed” version of the West—nor is it 20, 50, or 100 years “behind” some imagined global timeline. Perhaps its collected wisdom places it ahead. Or maybe it’s simply on a different arc altogether existing in a time and space that’s unfamiliar to those not attuned to its rhythms. Mumbai isn’t a chaotic New York with poor infrastructure, and Delhi isn’t Washington DC with more pollution. These aren’t incomplete copies, they are rich with history, layered with logic, and built for time and scale. To mistake them for knockoffs is to miss the genius of adaptation—and to misunderstand the future. India isn’t catching up, it’s just carving its own path. Until we recognize this, we’ll keep misreading the signals and imposing linear maps onto circular terrain.

*On a Lighter Note: A Cautionary Tale
In 2014, then New York City Mayor Bill de Blasio was photographed eating pizza with a knife and fork at Goodfellas on Staten Island. #Forkgate erupted, upsetting the sensibilities of New Yorkers everywhere. As a native son of Brooklyn—and a regular at Udupis and Darshinis far and wide—this one hits a little too close to home.

This is how trust in public institutions erodes

There’s the emerging middle class. The much-hyped demographic dividend. A surge in formal jobs. Digitization fixing the so-called scourge of informality. The whole “China Plus One” play. The list goes on.

And while I’m heartened by the attention that my adopted motherland is getting, there’s one thing that always gnaws at me when I read these pieces: the stories we hear aren’t always the stories that matter.

We might begin with Maya, a central concept in Indian philosophy. Traditionally, it refers to the illusion or appearance of the world—the idea that what we perceive with our senses is not the ultimate reality, but a veil that obscures a deeper, truer essence. It reminds us that clarity isn't always found in what's visible or measurable, and that the truths that matter most often lie beneath the surface. In the Indian worldview, Maya isn’t just deception—it’s the seductive pull of the obvious, the trap of mistaking form for substance. And in the context of understanding India, it’s a crucial lens: because without it, we risk being enamored by the image and missing the essence.

This truth is central to begin to understand India.

There’s an old parable about a drunk man searching for his lost keys under a streetlamp. A passerby stops to help. “Are you sure you lost them here?” he asks. “No,” says the man, “I dropped them in the park.” “Then why are you looking here?” “Because the light’s better.”

That, in a nutshell, is how the India story often gets told.

We look under the streetlamp—because the light’s better—while the keys lie somewhere in the dark.

Over the years, I’ve sat in air-conditioned rooms with investors, fund managers, and policymakers watching the same polished narratives play on loop. I’ve also spent time wandering through India’s wholesale mandis and dusty chowks, talking to family businesses, small manufacturers, and local traders. I’ve logged thousands of kilometers in the back of India Rail—and not once have I thought I had it all figured out.

But here’s what I do know: the gap between how India is presented and how it actually works is vast.

This series is about that gap—and the strange, beautiful, often contradictory truths it contains. It’s about peeling back the rhetoric and asking: what’s really going on here?

We’ll dig into themes like:

– Why formality and informality aren’t enemies

– How digitization is strengthening informality, not replacing it

– What gets missed when India is mapped through a Western lens

– Where we might be headed—and how we might actually get there

This isn’t a tribute to hustle culture, nor is it a romantic ode to jugaad. It’s about understanding India on her own terms.

The India I’ve spent the last 15 years trying to understand—and that has quietly shaped me in return—isn’t trying to mimic the West. It loops, it adapts. It builds systems that are resilient not because they’re perfect, but because they’re designed to survive in the wild. It delivers public goods and commerce through mechanisms few in the West would recognize or believe possible—just look at UPI.

And yet, even as UPI rewires financial life at breakneck speed, it doesn’t make India less Indian. In fact, it might be making it more so. If that already makes intuitive sense to you, great. If not, I’ll try to show you what I mean—through lived examples in the essays to come.

At heart, this series is a love letter to Bharat Mata and her enterprise. It’s a celebration of the entrepreneurs who power this economy: the street vendors, the small traders, the family firms, the informal networks that keep things humming while flying under the radar. It’s about celebrating what’s real in India, even if it doesn’t fit neatly into the Western playbook.

So if you’re also walking through India with more questions than answers—welcome. You’re in good company.

There’s magic in the in-between.