Milky Mist’s Q1 gives a pretty clean answer, and it’s not really a demand story. It’s a distribution story layered on top of a regulatory story — and those two things compound differently than organic demand growth does.

The distribution problem got solved for free

Start with the boring but structurally important part: quick commerce.

No FMCG brand is scaling meaningfully right now without Blinkit and Instamart doing a chunk of the heavy lifting. That’s not a hot take, it’s just how shelf economics have changed. Getting into 50,000 kirana stores takes years of relationship-building, credit terms, and margin give-backs to distributors. Getting onto Blinkit’s dark-store network gets you in front of a metro-dense, high-repeat, impulse-buying audience almost immediately and the app itself does a lot of the merchandising work that used to require a sales force.

I’ll use myself as a data point, because it’s a decent proxy for the mechanism: Milky Mist’s high-protein Skyr yogurt has more than 19,000 4-star reviews off Blinkit. That’s not brand loyalty in the traditional sense, it’s proximity plus habit. The product showed up in my cart rotation because it was there, repeatedly, at the moment I wanted it. That’s what quick commerce distribution actually buys a brand: not just reach, but frequency.

This matters for how you read the Q1 numbers, because “yogurt category +153% YoY” isn’t purely a product story. It’s partly a distribution unlock. A category that was previously constrained by cold-chain retail logistics suddenly has a channel that solves cold-chain at the last mile by design.

The part that actually matters: unorganized to organized

Here’s the thesis, and it’s the one worth sitting with.

India’s organized paneer market is only 5–7% of the total. Read that again, over 90% of the paneer sold in this country is loose, unbranded, made by local vendors with no consistent quality control, no food safety certification, and often, no actual paneer.

That last point isn’t rhetorical. FSSAI has been actively cracking down on “analogue” paneer products cut with vegetable fat and sold as the real thing. Combine that with GST changes that are compressing the cost advantage unorganized players relied on to undercut branded competitors, and you get a formalization wave that’s being driven by regulation, not by consumer preference shifting on its own.

This distinction matters more than it sounds like it should. Demand-driven category growth is fragile, it slows in a weak quarter, it’s vulnerable to a cheaper competitor, it depends on continued marketing spend. Regulation-driven formalization is sticky. Once analogue paneer gets pushed out and GST closes the informal-sector cost gap, that unorganized capacity doesn’t quietly come back next quarter. The market structure itself has shifted.

And when a category re-shuffles from unorganized to organized, it’s rarely an even split among branded players. The company with the widest existing distribution and the strongest shelf recall doesn’t just grow in line with the category, it disproportionately captures the share that unorganized players are shedding. That’s the setup here. Management explicitly flagged that cheese is next in line for the same treatment, again “thanks to FSSAI.”

If you’re trying to build a real thesis instead of a growth-momentum trade, this is the slide to spend time on. The Q1 print is just the first visible evidence of it.

Capacity: they’re not even close to constrained

The other thing that stood out on the call: current capacity supports roughly 3.5x of FY26 revenue. Paneer specifically is running at only around 50% utilization. Meaningful headroom also remains in ice cream, yogurt, and cheese.

This is worth flagging because it changes how you should think about the growth runway. A company posting 44% revenue growth while sitting on 50% utilization in its core category isn’t growing into capacity constraints, it’s growing well ahead of needing new capex for the core business. That’s a very different risk profile than a company whose growth requires successfully executing three new plant commissionings on schedule.

They did commission a new cheddar cheese plant this quarter regardless taking installed cheese capacity to 120 MT/day, which reads less like “we’re capacity-constrained” and more like “we’re pre-positioning for the cheese formalization wave before it fully arrives.” That’s a deliberate sequencing choice, not a reactive one.

The whey protein angle — the part most people will skip past

This is the detail buried in the concall that I think deserves more attention than it’ll get.

Milky Mist’s cheese production generates roughly 1 million litres of cheese whey per day. Right now, that whey is sold off as a relatively low-value commodity, whey powder, sold B2B. It’s a byproduct stream, not a product line.

They’re now building a Whey Protein Concentrate (WPC) unit, expected operational in 15–18 months. The stated intent is threefold: internal consumption, B2B sales, and the interesting one — a B2C play. Management is explicitly framing this as margin-accretive to the bottom line, not just a capacity-utilization exercise.

Think about what this actually is: every branded protein company in India right now is fighting for shelf space and paying for supply chain relationships to source whey protein concentrate. Milky Mist already has the raw input flowing out of its own cheese plants as a byproduct. It doesn’t need to build sourcing relationships or compete for supply, it needs to build downstream processing and a brand. That’s a materially easier problem than the one every pure-play protein D2C brand is solving right now from scratch.

It won’t move the needle in the next two quarters. But 15–18 months out, if executed, this is a second growth vector layered on top of the core dairy formalization trade and one that’s structurally cheap for them because the raw material is already a sunk byproduct of the business they’re already running.

The Q1 print itself

With all of the above as context, the actual quarter:

• Revenue: ₹973 Cr, up 44% YoY

• EBITDA: ₹144 Cr, up 77% YoY

• EBITDA margin: 14.8%, up from 12%

• PAT: ₹64.7 Cr, up 9x YoY

Category-level internals, largely driven by a strong summer season:

• Ice cream: +60% YoY

• Paneer: +34% YoY

• Cheese: +38% YoY

• Curd: +27% YoY

• Yogurt: +153% YoY, the standout, and the category most directly tied to the quick-commerce distribution unlock and protein positioning discussed above

PAT growing 9x against 44% revenue growth is the number that should catch your eye that’s operating leverage working almost too cleanly, and it’s worth watching whether that margin expansion (12% → 14.8%) holds as they scale into the new cheese and yogurt capacity, or whether some of it was a favorable one-off from mix (summer-driven ice cream and yogurt demand tends to carry better margins than the paneer base business).

Where this leaves the thesis

Strip away the quarter-specific noise and there are three separate, stackable tailwinds here, not one:

1. Distribution — quick commerce solved the last-mile problem for free, and it shows up disproportionately in categories like yogurt that benefit from cold-chain-light delivery.

2. Regulatory formalization — a 93-95% unorganized paneer market being compressed by FSSAI enforcement and GST changes, with cheese next in line. This is the structural, multi-year piece.

3. Capacity optionality — 3.5x headroom on existing infrastructure, plus a whey protein unit that converts an existing byproduct stream into a new branded category.

None of these are guaranteed to play out cleanly regulatory enforcement can be inconsistent, quick-commerce economics can shift, and a WPC brand launch 18 months out is still just a plan on a concall transcript. But the combination of “regulation is doing the demand creation for you” plus “you’re nowhere near capacity constrained” is a more durable setup than most consumer growth stories get.

If you made it till here, thanks for your time!

Subscribe now

Share

Leave a comment

The story everyone is telling is about a CEO’s contract.

I think that story is mostly wrong — or at least, badly incomplete. Here’s why.

The Timeline Doesn’t Match the Narrative

Sashidhar Jagdishan’s term as CEO ends October 26. Macquarie has put out research suggesting a 50/50 chance he gets a six-month extension versus a full three-year renewal.

That’s not analysis. That’s a coin flip dressed up with a research header, about an event that resolves in eight weeks.

Here’s the problem with hanging the entire selloff on it: the steepest fall in the stock happened between January and April. The succession debate is a summer story. It explains the last leg of the decline, not the first $40 billion of it.

If you’re holding this stock for a decade, an eight-week resolution to a CEO contract is noise. Loud noise, but noise.

What’s Not Noise

To be clear, I’m not saying ignore everything around the CEO situation. There’s a real pattern worth watching:

• The chairman’s abrupt exit

• Governance allegations: reviewed and dismissed, but raised

• Scrutiny over deposit handling

• A board that needed an RBI meeting before it could decide on its own CEO

No single item on that list is disqualifying. But together, they tell you something: the machine has friction.

Friction at the top is a watch item. It is not, by itself, a sell signal. And it’s not the question that actually matters here.

The Question That Actually Decides This

Here’s what determines whether this is a buying opportunity, and it has nothing to do with October:

Is HDFC Bank at a two-year low because the franchise is damaged or because the price finally caught up with post-merger reality?

The 2023 merger with HDFC Ltd changed the bank’s underlying economics in ways the market spent two years arguing about:

• Higher cost of funds

• A credit-deposit ratio that now caps loan growth

• Margins that no longer resemble the old, pre-merger HDFC Bank

Through 2024 and 2025, the market debated whether this was a temporary drag, something the bank would grow out of, or a permanent reset. The 2026 chart suggests the market has stopped debating and reached a conclusion.

The Three-Question Checklist

This is the part that actually matters, and it’s short:

1. Is deposit growth holding?

2. Is asset quality clean?

3. Is ROA stabilizing, or still sliding?

If the answer to all three is yes, then a 27% drawdown in India’s best-run large private bank is exactly the kind of pessimistic phase you wait years to buy into.

If ROA is structurally lower: if the 2x price-to-book era of old HDFC Bank is simply over, then this isn’t a discount. It’s a re-rating. The stock is cheaper because the business is worth less. And cheaper is not the same as cheap.

Where the Real Answer Lives

I don’t know which of these two scenarios is true yet. Neither does anyone claiming certainty based on a CEO headline, that’s a guess wearing a research note.

The answer isn’t going to come from an RBI approval letter in October. It’s going to come from the next two quarterly results, from what deposit growth, asset quality, and ROA actually do in the numbers.

The CEO decision will move this stock for a week. The balance sheet will move it for a decade.

If you made it till here, thank you for your time.

Subscribe now

Share

Leave a comment

Everyone’s talking about Welspun. Nobody’s talking about the smaller name standing next to it, absorbing the same tailwind: Man Industries.

This isn’t a “buy the small-cap peer” post. It’s a mechanism post. I want to walk through why an order booked in Arkansas, for a business unit with no direct overlap with Man Industries, still tells you something real about a ₹5,000-crore Indian pipe maker with a plant in Saudi Arabia. Then we’ll get into what Man Industries actually is, what management is projecting through FY29, and why I think of it as a proxy rather than a standalone bet.

First, what does Man Industries actually do?

Strip away the ticker and the model is simple: buy steel plate, weld it into large-diameter pipe (LSAW, HSAW), coat it against corrosion, and sell it against multi-year tenders to national oil companies, gas utilities, and water boards.

That last part matters more than it sounds. Man Industries isn’t selling into a spot market. It’s not a commodity producer hoping steel prices cooperate. It’s a build-to-order manufacturer, revenue visibility comes from an order book and a bid pipeline, not from inventory turning over. When you see the words “order book” attached to this company, that number is the forward revenue schedule, filled in years in advance by buyers who’ve already committed capex.

That’s also why 2026 has been a structurally different year for the company than any prior one. In May, Man Industries completed the acquisition of National Pipe Company (NPC) in Saudi Arabia for roughly ₹1,000 crore not just capacity, but pre-qualified vendor status with Saudi Aramco, ADNOC, and QatarEnergy. That approval alone is worth more than the steel; it’s the thing that takes competitors years to earn and money can’t fast-track.

Layer on a new stainless-steel seamless pipe plant coming up in Jammu (targeting chemical, nuclear, and defense-grade applications) and a coating facility in Dammam, and you get a company that went from “India pipe maker with an export book” to “dual-geography platform sitting inside a Gulf capex supercycle” in about twelve months.

The numbers management is putting on the table

Here’s where it stops being a story and starts being a set of commitments you can hold management to.

Metric FY26A FY27E FY28E FY29E
Revenue ₹3,592 cr ₹5,000–5,500 cr +25–30% (min.) +22–25%
EBITDA ₹450 cr. ₹600–650 cr ₹750–850 cr. ₹1,000–1,200 cr (target)
Margin 9–12.5%. ~11–12% ~12–13% ~12–14%

Read that FY29 target again: more than doubling consolidated EBITDA off a FY26 base, in three years.

Q1 FY27 gave an early proof point: consolidated EBITDA up ~92% year-on-year to a record ₹155 crore, even though the Saudi acquisition had only been consolidated for about 40 days. Standalone PAT was the highest in the company’s history. The India business is still running at 50–60% utilization, meaning there’s real operating leverage sitting unused before a single new tender gets signed.

None of this is guaranteed. Gross margin actually dipped sequentially in Q1 (management attributes it to order mix, not pricing, worth watching next quarter). Jammu has already slipped once, from an October 2025 target to March 2027. NPC still has to prove it can integrate at the margins management is promising. Projections are projections.

Why I think about this as a proxy, not a pipe stock

Here’s the actual thesis, and it’s bigger than one company’s order book.

Man Industries sits at the intersection of two things I think are underpriced in how retail investors think about India:

1. India moving up the steel value curve. For decades, Indian steel exports have skewed toward flat and long products - commodity-grade, thin-margin. The interesting shift happening now is fabricators pushing into coated, welded, specialty products (stainless seamless pipe for nuclear and defense applications is about as far from commodity steel as it gets). Man Industries’ Jammu plant is a direct bet on that migration.

2. The India-plus-Gulf infrastructure capex cycle. Gas grid expansion, water transmission, city-gas distribution in India. Aramco network expansion, the Master Gas System, desalination buildout in Saudi. These are both real, multi-year, government-and-NOC-funded capex programs and Man Industries’ order book is a leading indicator of that spend, because pipe orders get placed years before a pipeline is actually laid.

You don’t need to correctly time individual tenders in Riyadh or predict India’s next gas-grid phase. You need to own the company sitting in the middle of the value chain that gets paid regardless of which specific project moves first.

That’s the proxy argument. It’s also why the stock is a higher-beta, lumpier way to express that view than owning a diversified steel major, earnings depend on the timing of large project awards, not steady-state volume.

Back to Welspun — why their order matters for Man Industries

This is the part that made me want to write this piece.

Man Industries’ own management, on the Q1 FY27 call, described the current environment in explicitly structural language: “What we are seeing isn’t a one-off up cycle, it’s a structural multi-year shift.” They cited energy security diversification, water and desalination buildout, and industrialization via “giga projects” as the drivers and flagged the Strait of Hormuz disruption as accelerating a new structural tailwind in Asia that “didn’t really exist 18 months ago.”

That’s a management team talking their own book. Fine to be skeptical of that on its own.

But then Welspun: a larger, more closely watched peer, in a completely different geography (the U.S., not the Gulf), for a different client, in a different product line, books the largest order in its history. Management there called U.S. demand “very buoyant,” citing LNG export infrastructure and rising power-sector needs. And the execution window for that order? FY28 through FY29.

That’s the exact window Man Industries is guiding its own biggest growth numbers to.

Two manufacturers, two continents, two completely independent order flows, converging on the same multi-year window. That’s not proof of anything, one data point plus one data point is still just two data points. But it’s the kind of external corroboration that turns a management talking point into something worth actually underwriting. If the cycle were Man-specific, or Gulf-specific, you wouldn’t expect a North American LNG order to land on the same timeline for unrelated reasons.

The way I’d frame it: Welspun and Man Industries aren’t competing for the same orders. They’re picking up different, complementary strands of what looks like the same global re-investment wave in energy security, water infrastructure, and pipeline capacity. Welspun’s order is the louder, easier-to-see version of the same thing Man Industries’ bid pipeline (~₹24,000 crore, 70% MENA-weighted) is quietly building toward.

The takeaway

Man Industries isn’t a story about one hydrogen-ready pipe manufacturer having a good year. It’s a story about a company sitting at the exact intersection of India’s steel value migration and a genuinely structural, multi-geography infrastructure capex cycle — with a management team that’s now put specific, checkable numbers on the table through FY29, and a peer’s record order that just gave those numbers some real-world backing.

Watch the Q2 FY27 print for two things: whether that gross margin dip was really just mix, and how fast NPC’s contribution scales now that a full quarter is in the numbers. That’s where the thesis either firms up or needs revisiting.

Not investment advice. Do your own research. I hold no position in either company at the time of writing — check back for updates if that changes.

If you made it till here, thanks for your time!

Subscribe now

Share

Leave a comment

But the headline numbers aren’t the story. The story is buried in one line most people will scroll past: 70 lakh eye tests conducted this quarter, up 40% year-on-year.

That number is the entire business model. Everything else: the stores, the factory, the Gold memberships, the China JV, is downstream of it.

The business Lenskart is actually in:

Every eyewear company thinks it’s competing for the same customer: the person who already knows their power, already owns a pair of specs, and is deciding whether to buy their next one from Lenskart, Titan Eyeplus, or the local optician.

Lenskart isn’t fighting that fight. It’s not trying to win share of an existing, finite pool of glasses-buyers. It’s expanding the pool.

The mechanism is simple and almost boringly effective: once a person gets an eye test done, they buy glasses within months, not eventually, not “someday,” but reliably fast. The eye test isn’t a service Lenskart offers alongside retail. It’s the retail funnel’s first and most important step.

So the company’s real KPI isn’t footfall or conversion rate in the traditional sense. It’s tests conducted. Get someone into the chair, get the exam done, and the purchase follows almost automatically. That’s why 63 lakh of those 70 lakh tests happened in India alone, and this is the number that matters half of them were first-time exams.

Half. Not repeat customers coming in for an annual check. First-timers. People who had never had a professional eye exam in their life, walking into a Lenskart store and getting one.

The size of the untapped pool:

India has roughly 78 crore people who need some form of vision correction. The overwhelming majority of them have never had a proper exam. Not “never bought glasses from a modern retailer” never been tested at all.

This is the number that should reframe how you think about Lenskart’s addressable market. It’s not competing for a slice of India’s existing eyewear-buying population. It’s sitting on top of a latent-need pool measured in hundreds of millions of people who don’t yet know they’re customers — because nobody has told them they need correction.

Every eye test is a conversion event in the truest sense: it turns a person with undiagnosed, unaddressed need into an active, paying customer. That’s a fundamentally different growth mechanic than same-store-sales optimization or SKU expansion. It’s demand creation at the population level, one exam at a time.

Why this industry stayed small and fragmented for decades:

Here’s the part that actually explains the last 20 years of Indian eyewear retail, and why it took a company like Lenskart to break the pattern.

The bottleneck was never lenses. It was never frames. Both are commodities, manufacturable at scale, sourceable globally, easy to differentiate on style but not on access.

The bottleneck was finding a qualified person to write a prescription.

India trains very few optometrists relative to its population. That single supply constraint kept the entire downstream industry small and fragmented because you can’t sell glasses to someone who’s never been told they need them, and you can’t tell them that without a trained professional in the room.

Every eyewear player before Lenskart operated within that ceiling. More stores didn’t solve it. Better marketing didn’t solve it. The constraint was upstream, in a healthcare workforce shortage that retail alone couldn’t fix.

Lenskart’s actual innovation, the one that doesn’t show up in the topline numbers was recognizing that the company wasn’t in the eyewear distribution business first. It was in the diagnostic access business first, and eyewear distribution second.

How they’re solving it:

The fix has been remote optometry: a centralized pool of qualified optometrists conducting exams remotely across the store network, decoupling “an eye test happening” from “a trained optometrist being physically present in that store.”

The scale-up here is the real signal in this quarter’s results: 786 stores now on remote optometry, up from just 168 a year ago. That’s not incremental rollout — that’s a near-5x expansion in twelve months. It’s the clearest evidence that Lenskart has cracked a repeatable way to route around India’s optometrist shortage rather than waiting for the country to train its way out of it.

Layered on top of that is an AI self-testing pilot — an early bet on removing the human bottleneck almost entirely for a subset of cases. It’s still in pilot, so the jury’s out on accuracy and regulatory comfort, but directionally it’s the same logic taken one step further: the constraint was never physical retail capacity, it was diagnostic throughput, and every lever gets pulled to expand that throughput.

Manufacturing and distribution followed not the other way around:

Once you see the eye-test-first logic, the rest of the quarter’s data points read as consequences, not separate initiatives.

• Hyderabad factory, 50 million pairs/year capacity — you don’t build manufacturing at that scale unless you’re confident demand generation upstream (the tests) is reliable enough to fill it.

• China JV stake raised to 70% for frames — vertical control over frame sourcing, again downstream of confidence in demand, not a bet made in isolation.

• 93.5 lakh Gold members, subscription revenue up 57.4% — a subscription model only works if customers have a reason to return regularly, and regular eye checks are exactly that reason. Gold membership is really a retention wrapper around the diagnostic relationship.

Lenskart solved the prescription bottleneck first. Manufacturing scale and distribution reach were built around that solved bottleneck, not alongside it as parallel bets.

The store density test and why it’s passing:

This is the part of the results that should worry competitors more than the revenue growth itself.

Lenskart now runs 2,725 stores across India, covering 95%+ of pin codes, with an internally mapped total opportunity of roughly 11,800 stores. That’s still early — under a quarter of the way to their own estimate of saturation.

The standard retail fear at this stage of expansion is cannibalization: new stores opening near existing ones start eating each other’s sales, same-store-sales growth slows, and the whole densification strategy starts looking like diminishing returns.

That’s not what’s happening.

Same-pin-code sales grew 24.3%. Same-store sales grew 18.3%.

Read those two numbers together carefully. If new stores were simply splitting existing demand within a pin code, same-pin-code growth would track below or in line with same-store growth — the pie would be getting redistributed, not enlarged. Instead, the pin-code-level number is higher than the store-level number. That only happens if the new stores are pulling in customers who weren’t shopping at Lenskart before at all — which loops straight back to the eye-test thesis. New stores mean more testing capacity, more first-time exams, more newly-created customers, and the pin code’s total demand pool actually grows.

And this pattern holds across metro, tier 1, and tier 2+ markets, meaning it’s not a big-city halo effect. It’s structural.

The marketing line that confirms all of this:

One more data point, easy to skim past: marketing spend dropped to 4.8% of revenue, from 5.7%.

In a business that was genuinely fighting for share of an existing customer base, falling marketing intensity alongside 43% revenue growth would be a red flag, either underinvestment or a data anomaly. Here it reads as confirmation of the thesis: when your growth engine is word-of-mouth around a healthcare service (get tested, discover you need glasses, tell your family), you don’t need to spend as much to acquire the next customer. The eye test itself, and the social proof it generates, is doing the acquisition work that a paid marketing budget would otherwise have to do.

The takeaway:

Lenskart isn’t a glasses company that happens to offer eye tests. It’s a diagnostic access company that happens to sell glasses as the natural output of that diagnosis.

The industry’s ceiling for decades wasn’t capital, retail real estate, or even brand, it was a shortage of qualified optometrists gating the entire funnel at its narrowest point. Lenskart identified that constraint, built remote optometry infrastructure to route around it, and only then invested aggressively in manufacturing capacity, frame sourcing, and store density because none of that capacity matters if you can’t get people diagnosed in the first place.

The store density numbers aren’t proof of market saturation. They’re proof of the opposite: every new store is still finding uncaptured demand, which means India’s 78-crore-person vision-correction gap is still, by Lenskart’s own numbers, barely scratched.

If you made it till here, thanks for your time!

Subscribe

Share

Comment

On Thursday, the Lok Sabha voted to delete that sentence. No debate. A voice vote, taken while the House was consumed by a protest over something else entirely.

If you scrolled past this in your feed, you’re forgiven — the headline writing itself (”UPI to become chargeable”) is wrong, and the accurate version is boring enough that most outlets didn’t bother. But the accurate version is the important one. So let’s do it properly.

The line that got deleted

The provision is Section 10A of the Payment and Settlement Systems Act, 2007. It barred banks and payment system providers from charging — directly or indirectly — on the electronic payment modes listed under the tax law’s digital payments clause. UPI was on that list. That’s the entire mechanism behind “Zero MDR.”

Here’s the part worth sitting with: Zero MDR was never a policy stance. It wasn’t a “for now” decision the government could reverse with a press release. It was written into a statute. Changing it required an Act of Parliament — a vote in both Houses, on the record, with a debate that any journalist could quote from.

That’s what got removed this week.

What the amendment actually does

To be precise, because precision is the whole point here: the amendment does not impose a fee on UPI. Nobody is charging you anything today. Read that twice, because it matters for what comes next.

What it does is smaller and, in a way, more significant. It deletes Section 10A’s reference to the tax law’s list of payment modes, and replaces it with a line empowering the central government to notify which payment modes stay free.

Look at what changed in the architecture, not the outcome:

• Before: Removing the free guarantee required a Bill, passed by both Houses, with a debate on record.

• After: Narrowing the guarantee requires a notification — signed by the executive, no Parliamentary vote, no debate required.

Same destination, reachable now with a tenth of the friction, and none of the paper trail. That’s not a fee. That’s a new pen — and whoever holds it can decide, later, without asking anyone.

Where things stand today — not tomorrow, not “soon”

• Lok Sabha passed the amendment on 6 August.

• Rajya Sabha has not taken it up.

• Nothing has been notified.

• No merchant, anywhere, is paying anything new.

If you see a headline claiming UPI is now chargeable — it’s wrong, full stop. The mechanism for a charge now exists. The charge does not.

The political fight, and why you should discount both sides

The argument started the same day the Bill passed.

The Opposition’s claim: this is Washington leaning on India. The Finance Minister’s claim: any MDR would fall on merchants, not customers.

Here’s the uncomfortable bit — both of these statements are true to the letter, and both come from people with an obvious incentive to say them. That’s not a reason to ignore either claim. It’s a reason to discount both and go looking for what’s actually documented.

What’s documented

• In March 2026, the US Trade Representative listed India’s digital payments policies as a foreign trade barrier.

• Last month, the US imposed 25% tariffs on Brazil under Section 301 — and Brazil’s Pix payment system was explicitly named in that case.

So the external pressure is real, and it’s on paper. What is not on paper is a direct causal line from that pressure to this specific clause in this specific Bill. That link is currently just one politician’s assertion, met by another politician’s denial. Neither side has produced evidence. Until someone does, treat it as an open question, not a settled fact — regardless of how satisfying either narrative feels.

The domestic explanation nobody’s running with

There’s a simpler, homegrown story sitting underneath the geopolitical one, and it’s genuinely under-discussed: India has been quietly unwinding zero MDR since 2023. RuPay credit transactions on UPI already carry a merchant fee above ₹2,000. This amendment isn’t the start of that trend — it’s the legal scaffolding catching up to a direction the system was already moving in.

“Merchant only” is correct — and incomplete

The Finance Minister’s defense — that any MDR lands on merchants, not customers — is technically accurate. It’s also missing the second half of the sentence.

When merchants get charged a transaction fee, they have two options: pass the cost on, or stop accepting the payment method. Kirana shops, small hospitals, and neighborhood restaurants have historically picked the second one. “The customer doesn’t pay” is only true until the merchant decides UPI isn’t worth accepting anymore — at which point the customer pays in a different currency: inconvenience, cash-only counters, or a card surcharge that mysteriously appears instead.

Letter of the law: correct. Behavior on the ground: a different story.

Where the trade angle actually has teeth

If you’re looking for where US pressure genuinely shows up, it’s not in the UPI clause — it’s in what got bundled alongside it.

The same Bill:

• Exempts Foreign Institutional Investors from tax on government securities interest.

• Widens tax relief for foreign companies using Indian data centres.

That second item — data centre tax relief — is a stated American ask. This is the part of the package where the trade-pressure story has actual documentary support, not just competing press statements. The UPI clause got the headlines; the data centre clause did the quieter work.

The uncomfortable economics nobody wants to say out loud

Free was never free. Somebody funded the rail — through subsidy, through cross-subsidy, through banks absorbing cost as a customer-acquisition play. India built a real-time payments system that the rest of the world now studies and copies. What it never did was settle, permanently and transparently, who pays to keep the lights on.

That bill was always going to come due eventually. This week, the legal mechanism to present it arrived — even if the invoice hasn’t been issued yet.

Revisiting my own take

I’ve written before that India exports payment sovereignty while everyone else plays defence — and I still believe that, as a description of UPI’s global standing. But intellectual honesty means updating the sentence, not just repeating it: the product being exported is a free rail, and India has just removed the statutory guarantee that it stays that way. That tension is real. I’m not going to write around it.

What this does — and doesn’t — do to the card networks

Here’s a prediction I’ll make plainly: this changes less for Visa and Mastercard than the optimists in that camp think.

Making UPI marginally less free does not make a card competitive with it. The moat was never the fee — UPI’s advantage is interoperability, ubiquity, and infrastructure lock-in. A shrinking moat is actually more dangerous to bet against than no moat, because it still looks safe right up until the point it isn’t.

The date that actually matters

Everyone’s watching August. The date worth watching is December 2026 — when the 30% market-share cap on third-party UPI apps comes due.

As of December 2025, two American-owned apps handled over 80% of UPI transactions. If that cap is enforced as written, it forces a restructuring of UPI’s competitive landscape that dwarfs anything a merchant discount rate notification could do. If it’s not enforced — or gets quietly diluted the way this week’s amendment quietly rewired Section 10A — that tells you something too.

The summary, one more time, because it’s worth repeating:

A law changed this week. A fee did not.

Watch what gets notified — and watch whether December’s cap survives contact with the same instinct that just moved this decision from Parliament’s desk to the executive’s.

If you made it till here, thanks for your time!

Subscribe now

Share

Leave a comment

There’s a moment in HBO’s Billions where Bobby Axelrod — hedge fund king, perpetual SEC target, a man who has never in his life needed anyone’s permission — decides he wants to own a bank.

Not invest in one. Not partner with one. Own one.

Everyone around him treats it like a midlife crisis. Why would the guy who beats the system want to join it?

Axe’s own answer is the whole thesis in one line:

“I want to own the bank so that I don’t have to rob it.”

It sounds like a throwaway flex. It isn’t. It’s a compressed summary of one of the more interesting undercurrents in modern finance — why capital, once it gets big enough, stops wanting to borrow and starts wanting to own the source. And this isn’t Hollywood invention. It’s a real, well-worn playbook.

The Yonkers Play

In Season 5, Axe’s bank ambitions get tangled up with his hometown redevelopment plans in Yonkers. He’s not content rebuilding the neighborhood store by store — he applies for a bank charter from the New York Department of Financial Services, so that he’s not just the developer of his own opportunity-zone projects, he’s the lender too.

Nobody gets to say no to him. There’s no bank committee to convince, no rate to negotiate, no outside approval standing between Axe and Axe’s own capital.

Chuck Rhoades — the state AG who’s spent years hunting him — tries to block the charter. And that’s really the second half of the motive: a bank charter doesn’t just get Axe money. It moves the entire fight to a different battlefield, with different regulators, different rules, and a fresh shot at legitimacy in a world that’s never fully accepted his money as “clean” enough to matter.

Underneath the strategy, that’s the real engine: status. Axe has spent the whole show clawing his way from a working-class Yonkers upbringing into billionaire rooms that still look at him sideways. Owning a bank isn’t just a funding source. It’s the establishment being forced to call him one of their own.

Why Real Institutions Do the Exact Same Thing

Strip away the drama and the mechanism underneath is completely real. Here’s why actual funds, firms, and corporates chase bank charters.

1. Stable, cheap funding Hedge fund capital is scared capital. LPs can redeem. Prime broker leverage can vanish the moment the market gets nervous. Bank deposits are the opposite — sticky, insured up to $250k, and depositors don’t panic-pull the way fund investors do. A bank charter converts a firm’s capital base from “hot money that flees at the first bad headline” into something closer to permanent.

2. Access to the Fed’s safety net This is the single biggest real-world driver, and it’s not subtle. When Goldman Sachs and Morgan Stanley converted to bank holding companies in September 2008, it wasn’t a strategic pivot — it was survival. The conversion gave them access to the Fed’s discount window and made them eligible for TARP. Lehman and Bear Stearns didn’t have that option. One of those differences is a big part of why one set of firms is still standing.

3. Vertical control over your own capital Own the bank, and you don’t just borrow money — you originate loans, hold deposits, and fund your own deals without needing anyone else’s yes. This is Axe’s Yonkers logic exactly: stop financing your own projects through someone else’s discretion.

4. A different regulator, a different fight Hedge funds live under the SEC’s microscope — insider trading, disclosure, manipulation. Banks answer to a completely different stack: the OCC, FDIC, state banking departments, the Fed. It’s not necessarily lighter — banking is one of the most regulated industries on earth — but it’s a different lane, sometimes a more predictable one than an open-ended SEC enforcement posture.

5. Legitimacy “Shadow banking” — hedge funds, PE, non-bank lenders — has always carried a stigma of being the unregulated, opportunistic side of finance. A bank charter is a signal: we’re not outsiders anymore, we’re inside the system now.

The Real List

Pure hedge funds chasing bank charters is actually rare in real life — short holding periods and leveraged, illiquid bets don’t pair well with deposit-insurance rules. What’s far more common is fintechs, auto lenders, and wealth managers wanting the same structural control Axe wanted. Here’s who’s actually done it:

Crisis-driven conversions (2008)

• Goldman Sachs — converted to a bank holding company for Fed discount-window access

• Morgan Stanley — same move, same week, same survival logic

The ILC (Industrial Loan Company) route — the real-world Yonkers charter An ILC lets a commercial company own a bank without becoming a full bank holding company under the stricter Bank Holding Company Act. This is the actual regulatory loophole Axe’s dormant-charter scheme is a dramatized version of.

• Walmart — applied 2005/06, withdrew in 2007 after political and banking-industry backlash — the closest real echo of “should commerce be allowed to own banking”

• Square (now Block) — approved 2020

• Nelnet — approved 2020

• Rakuten — applied 2020

• SoFi — dropped its ILC bid, later got a full national bank charter in 2022 via acquisition instead

• UBS — already runs UBS Bank USA, historically the largest ILC in the country

• Thrivent Financial — approved 2024

• GM Financial and Ford Credit — both want to finance their own car loans in-house, conditional approvals in early 2026

• Stellantis, Nissan, OneMain Financial — applications pending

• PayPal and Affirm — applications filed within the last year

• Edward Jones — first filed in 2020, withdrew in 2022, refiled in 2025, finally approved — a six-year saga almost as drawn-out as Axe’s Yonkers fight

The pattern is telling: it’s mostly fintechs wanting to hold customer money directly, and auto/wealth firms wanting to be their own lender — not hedge funds looking to escape the SEC. Axe’s version is more dramatic than reality, but the underlying want — stop needing anyone’s permission for your own capital — is identical.

The Takeaway

Every layer of this — fiction and fact — reduces to the same idea. Once a firm gets big enough, the constraint stops being “can we find money” and starts being “can we control the terms on which we get it.” A bank charter is the ultimate answer to that question: you’re no longer asking for capital, you’re the one who decides who gets it.

Axe’s line holds up better than most one-liners TV finance drama tends to produce:

“I want to own the bank so that I don’t have to rob it.”

Control the capital. Control the rules. Stop needing to bend either.

If you made it till here, thanks for your time.

Subscribe now

Share

Leave a comment

Nine days. That’s roughly the gap between the peak and the forced sale becoming public.

The story has been framed two ways online. One version is a conspiracy: Citadel talked up a Fed rate hike, waited for the panic to break Leopold, then bought his book cheap. The other version, the one I want to walk through, is less dramatic and more useful — this is what leverage does, mechanically, every single time someone runs it at scale without enough liquidity behind it.

The trade that made him famous

Aschenbrenner, a 24-year-old former OpenAI researcher, built his fund around one directional thesis: AI hardware wins, AI software loses. Long chipmakers like SK Hynix. Short software names like Adobe. For a year and a half, that trade printed. It’s a clean thesis — compute is the scarce input, software margins get compressed by AI-native competition, and being early and concentrated in it is exactly what took the fund from a standing start to $45 billion in NAV.

But NAV isn’t the number that matters when you’re trying to understand what happened next. NAV is assets minus liabilities, the investors’ equity in the fund. It is not the size of the fund’s actual market positions. And Situational Awareness was reportedly running leverage of roughly 4x.

At 4x leverage, a $45 billion NAV doesn’t support $45 billion of market exposure. It supports somewhere close to $180 billion of gross exposure: longs and shorts combined. That distinction is the entire story. Exiting “the public book” didn’t mean $45 billion changed hands. Multiple reports now suggest the actual unwind was well north of $100 billion in securities.

Where leverage turns on you

Leverage is a magnifier, not a strategy. It doesn’t care which direction you’re right or wrong in, it just scales the outcome. And in mid-July, Aschenbrenner got hit on both sides of a book that was supposed to be hedged against exactly this kind of scenario.

His long book: the chip and AI-infrastructure names fell over 30% in about two weeks. His short book: Adobe, the trade meant to offset that downside went up instead of down. A long/short portfolio is supposed to reduce your directional risk. Here, both legs moved against him simultaneously. At 4x leverage, that isn’t a bad month. That’s a liquidity event.

The letter that didn’t land

On July 24th, with losses mounting, Aschenbrenner wrote to investors. He called the selloff “some of the most attractive opportunities” the fund had seen, and asked for fresh capital by August 1st.

It’s worth sitting with that letter for a second, because it’s the moment the story stopped being about markets and started being about liquidity. Being right about long-term value doesn’t help you if your prime brokers are marking your collateral daily and your capital call doesn’t land before the calendar catches up with you. The fund needed cash on a timeline the market wasn’t going to give it.

The money never came.

The rate note, and why it matters less than it sounds

On July 28th, Frank Flight, who runs macro strategy at Citadel Securities, published a note predicting a surprise Fed rate hike, a move the market hadn’t priced in. That note is real; it’s on record. What followed it is the part that gets contested: AI stocks, already down sharply, sold off further into the panic. The Fed, when it met, left rates unchanged.

This is the single data point the “Citadel rigged it” theory hangs on, and it’s worth being precise about what it does and doesn’t establish. A market-making desk publishing a directional rate call that turns out to be wrong is not evidence of manipulation on its own desks publish wrong calls constantly, and being early or incorrect on a Fed prediction isn’t illegal or even unusual. What made this note consequential wasn’t malice, it was timing relative to a fund that had no room left to absorb any further downside. A note that would have been a rounding error for a fund at 1x leverage was existential for a fund at 4x, ten days from a margin call.

Same morning, three brokers

Thursday, July 30th. Bank of America, Goldman Sachs, and JPMorgan, the fund’s three prime brokers issued margin calls on the same morning. When your collateral is public equity, marked daily, and it’s dropped 35–47% on core positions, the broker doesn’t wait for you to find a plan. It calls the loan or forces the sale.

There’s a specific reason the fund’s private holding a stake in Anthropic, reported around $5 billion survived this while the public book didn’t. Private company stakes don’t have a daily observable market price. They can’t be continuously re-marked against a margin threshold, so they can’t be called as collateral the way a liquid, exchange-traded position can. That’s the entire reason Aschenbrenner walked out of this with something left standing. It wasn’t strategy in the moment, it was that the private stake was never eligible to be called in the first place.

With no room left to negotiate, the entire public book, longs and shorts, together went out in one block trade before Thursday’s open. Millennium and Jane Street were both reported to have looked at the book. Citadel bought it.

So — did Citadel do this on purpose?

Every outlet reporting this story, from the Journal to CNBC to Bloomberg, has been careful to draw the same line: the timing is suggestive, but nobody has established intent. That’s not me being diplomatic, it’s genuinely where the reporting stands as of today.

What I’d push back on is the framing that this needs to be a conspiracy to be a damning story. Taking the other side of an overleveraged, forced seller during a drawdown isn’t some hidden dark art on Wall Street — it’s one of the most well-worn plays that exists. You don’t need to coordinate a rate call to benefit from someone else’s leverage breaking. You just need liquidity, patience, and a willingness to move fast when the call comes. Griffin has told a version of this story before, from 2007: a rival fund needed to dump a $30 billion book before a Monday open, the competing bidder’s team went home for the night assuming they’d pick it up in the morning, and by 6 AM Citadel owned the entire portfolio.

That’s not a new playbook. It’s the same one, run again on a bigger number, on a younger manager, nineteen years later.

The actual lesson

If there’s a single mechanism-level takeaway here, it’s this: leverage doesn’t create returns, it borrows them from a future version of you that has to survive volatility you haven’t experienced yet. A fund can be right about the thesis — AI hardware over AI software may well still be the correct multi-year call — and still get liquidated, because being right on direction and surviving on timeline are two completely different problems. Aschenbrenner’s fund didn’t fail because the chips-over-software trade was wrong. It failed because it was sized for a world where nothing goes against you for more than a few days at a time.

That’s the risk nobody prices in on the way up. It only shows up once, and it shows up all at once.

When you’re over-levered, facing a drawdown, the guy sitting across you at the table, is cutthroat and he hardly cares.

If you made it till here, thank you for your time.

Subscribe now

Share

Leave a comment

Let me back up and explain the loop first, because you need the mechanics before the India angle makes sense.

The loop, plainly

Nvidia just announced over $750 billion in combined AI deals. The headline number is the Korea deal: more than $500 billion with SK Hynix’s parent company, to build out over 2 gigawatts of AI data center capacity. For context, that’s enough power draw for roughly 1.5 million homes. Separately, Nvidia is reportedly in talks to guarantee up to $250 billion of OpenAI’s lease payments on a data center project in Ohio, and to directly finance another $350 billion of OpenAI’s own chip purchases from Nvidia.

Read that last part again slowly. Nvidia sells the chip. Then Nvidia helps pay for the chip it just sold. It gets paid on the sale either way, whether the AI demand behind that data center ever actually materializes or not.

Here’s why this isn’t just a philosophical quibble, it hits the income statement directly. Nvidia books that chip sale as revenue the moment it ships, full stop. Doesn’t matter that Nvidia itself supplied part of the financing that made the purchase possible. So the top line keeps printing strong, quarter after quarter, and nobody outside the deal room actually knows how much of that “demand” was manufactured by Nvidia’s own balance sheet versus real customer pull.

And Nvidia’s not doing this alone, for what it’s worth. Google is backstopping Anthropic’s lease payments across five data centers: functionally a $35 billion loan dressed up as an infrastructure deal. SoftBank has put close to $65 billion into OpenAI and had to take out a $40 billion bridge loan just to cover that single position. This is becoming the house style for how AI infrastructure gets funded right now: the seller becomes the lender becomes the guarantor, all in the same transaction.

Jensen Huang’s been asked about this directly, and he’s called the “circular” label ridiculous, his argument being that these financing arrangements are a small fraction of what these companies ultimately raise from other sources. That’s a fair point to put on the table, and I don’t think he’s lying when he says it. But it’s also not really the point. The concern was never that circular financing is the majority of Nvidia’s revenue. It’s that it’s enough to distort the growth number everyone’s using to justify the valuation and distortion at the margin is exactly how bubbles get their air.

Michael Burry’s been saying a version of this too, about a separate Nvidia deal tied to Elon Musk’s xAI his allegation being that billions in chips are sitting off balance sheet, and that the risk eventually lands on American retirees through insurance products that hold this exposure without anyone quite realizing it. I’d treat that one as an allegation, not a confirmed fact, but it rhymes with everything else in this piece closely enough that it’s worth knowing it’s out there.

If AI demand ever comes in below what’s been underwritten here, the chipmaker, the buyer, and whoever financed the deal all take the hit together. Same foundation, same collapse, no separation between them.

Now here’s where India comes in

While all of that’s been unfolding in the US, a completely different but structurally related thing has been happening at home. Look at who’s building on Nvidia in India right now, and it’s not a short list.

Reliance is building large-scale AI supercomputing infrastructure and local-language models. Tata Communications is standing up massive data center capacity running on Nvidia chips end to end. L&T — a construction and engineering company — is now building what Nvidia calls “gigawatt-scale AI factories.” Yotta Data Services is running sovereign cloud infrastructure on Nvidia GPU clusters. Netweb is manufacturing the servers domestically. On the software side, Infosys is building small language models on the Nvidia stack, TCS has an entire business unit dedicated to Nvidia AI Enterprise and Omniverse, and Wipro and Tech Mahindra are training developers and deploying autonomous agents through the same partnership. Even further out, Addverb is training humanoid and warehouse robots on Nvidia’s simulation tools, and Hero MotoCorp — a two-wheeler manufacturer — is running product engineering on Nvidia-accelerated infrastructure.

That’s infrastructure, software, and physical application, three layers deep, all resting on one vendor.

Here’s my actual worry, and it’s not the same worry people have about the US side. In the US deals, Nvidia is financially entangled with the buyer — it’s on the hook too, which at least means its incentives are somewhat aligned with the deal succeeding. In India, from what’s public right now, these look like straightforward commercial partnerships. Reliance and L&T aren’t being financed by Nvidia the way OpenAI is. They’re customers and infrastructure partners, full stop.

Which means if global AI demand disappoints and Nvidia needs to protect its own numbers, it doesn’t have the same reason to keep propping up the Indian side of this. It has no loan outstanding to Reliance that it needs to protect. It has no lease guarantee with L&T sitting on its books. If capacity needs to get rationed, or capex plans need to get quietly delayed, or GPU allocation needs to shift toward whichever region is contractually locked in through financing — India isn’t the region holding Nvidia’s own money hostage. India is the region Nvidia can walk back from with the least self-inflicted damage.

And that’s the part that doesn’t show up in any of these press releases. Companies here have already made the sunk-cost commitment — the concrete’s getting poured, the workforce is getting trained on this specific stack, the CAE systems are getting rebuilt around it. That commitment is a one-way door for Reliance or Hero MotoCorp. It is not a one-way door for Nvidia. If a chip vendor needs to protect its own margins in a downturn, it cuts the relationships that cost it the least to cut — and the deals with the least financial entanglement are always the easiest ones to trim first.

I want to be careful here, because I haven’t seen anything suggesting Nvidia is currently planning to pull back from India, and none of these companies have signaled distress. This is a risk case, not a prediction. But when I look at who’s exposed if the broader AI financing loop in the US ever comes under real pressure, the honest answer is: everyone in this loop has some kind of leverage over Nvidia except the Indian side of the relationship. Reliance and TCS and L&T don’t hold Nvidia’s paper. Nvidia just holds their order books.

What I’d actually be watching

Whether any of these Indian companies start dual-sourcing chip architecture — even partially — as a hedge, rather than going all-in on one vendor relationship.

Whether Nvidia’s India deals ever pick up the same financing language the US ones have — guarantees, backstops, direct lending. That would actually be a good sign, strange as that sounds, because it would mean Nvidia has skin in the game here too.

And whether Nvidia’s own commentary on future capex, on any earnings call, starts drawing distinctions between “committed” infrastructure spend and “planned” spend. That’s usually the tell for where a company plans to cut first if it needs to.

None of this means don’t participate in the theme — India’s AI infrastructure build-out is real, and these are serious companies making serious bets. It means understand exactly what you’re exposed to when you own it. Right now, a meaningful chunk of India’s AI story is downstream of decisions being made in Santa Clara about capital allocation in a financing structure most of these Indian partners aren’t even party to.

If you made it till here, thanks for your time!

Subscribe now

Share

Leave a comment

That’s the headline. That’s the number that will get quoted in every brokerage note this week. And honestly, it’s not what caught my attention.

A far quieter disclosure did.

Tucked into the same announcement was a single line: the company has secured an order to develop a perception stack and provide engineering R&D services for robotics and Physical AI platforms.

Most investors will read that, shrug, and scroll past it. It sounds like a small contract win — the kind of line that shows up in a hundred corporate announcements a year and means nothing.

I think this one might mean something. Here’s why.

Every robot has to solve one problem before it does anything else

Forget the humanoid hype reels for a second. Forget Optimus doing yoga on stage. Strip a robot down to first principles and it needs exactly one thing before it can do anything useful: it needs to understand the world around it.

Motors give a robot the ability to move. But movement without understanding is just motion — potentially dangerous motion. Perception is what tells a robot where to move, what to avoid, and how to interact with whatever’s in front of it.

If Nietzsche had been a robot, he wouldn’t have written “I think, therefore I am.” He’d have written “I sense, therefore I am.”

Without perception, a robot isn’t autonomous. It’s blind. And a blind robot in a warehouse, a factory, or your living room is not a product — it’s a liability.

How robots currently “see”

Today’s autonomous systems build their picture of the world using a combination of sensors: cameras, LiDAR, radar, and ultrasonic sensors, each compensating for the others’ blind spots.

LiDAR gets most of the attention because it’s remarkably precise — it’s the sensor most associated with self-driving cars and high-end robotics. But precision comes at a cost. LiDAR units are expensive, and that cost becomes a serious obstacle the moment you try to deploy sensors at scale — across an entire fleet of warehouse robots, or eventually, across millions of humanoid units.

If robotics is going to go from demo-stage novelty to mass deployment, someone has to solve the cost problem in perception. That’s the opening.

Enters Novelic

A few years ago, Sona BLW acquired Novelic, a Serbian company specializing in millimetre-wave radar technology. At the time, this probably registered as a minor bolt-on acquisition — the kind of deal that gets a paragraph in the annual report and is quickly forgotten.

Novelic’s core business is radar for in-cabin sensing: accident detection, occupant detection, the kind of tech that makes your car seatbelt chime go off because it’s mistaken your gym bag for a passenger who forgot to buckle up.

Trivial-sounding use case. Directly relevant technology.

Radar has real structural advantages over optical systems. It performs reliably in darkness, fog, and rain — conditions where cameras and even LiDAR can struggle. And it offers a meaningfully lower-cost alternative for certain sensing use cases.

To be clear: radar isn’t a LiDAR killer. It’s not going to replace cameras or LiDAR wholesale. The realistic scenario is radar becoming an increasingly important layer in a multi-sensor perception stack — filling in where the other sensors are weak, and doing it more cheaply.

Which means the company that already owns radar expertise has a head start on one piece of the puzzle that every robotics company will eventually need to solve.

The pieces start lining up

Here’s where it stops looking like a coincidence and starts looking like a plan.

1. Motors, driveline systems, precision components — Sona Comstar’s existing core automotive business, built over decades.

2. Radar sensing — acquired through Novelic, originally for automotive in-cabin applications.

3. Perception software and engineering R&D services — the new disclosure, explicitly aimed at robotics and Physical AI platforms.

Individually, each of these reads as a normal auto-components company doing normal auto-components things. Together, they start to resemble the building blocks of a robotics supply chain: the actuation (motors), the sensing (radar), and now the software layer that turns raw sensor data into something a robot can actually act on.

That’s not three unrelated initiatives. That’s a stack.

Why this matters beyond one company

Zoom out and the direction of travel becomes hard to ignore. Tesla has repeatedly signaled that Optimus could become one of its most consequential long-term products — not a side project, a pillar. BYD and a growing list of global manufacturers are pouring capital into robotics and embodied AI.

Nobody knows exactly how fast this market will scale. Adoption timelines for humanoid and industrial robotics remain genuinely uncertain, and I’d treat anyone who claims precise visibility into 2030 robot unit volumes with some skepticism.

But the direction is clear enough. Physical AI is where a meaningful chunk of global capex attention is heading, and the companies that own critical sub-components of the perception layer — the unglamorous, unsexy, “just a sensor supplier” companies — stand to benefit regardless of which humanoid brand ultimately wins.

The market’s blind spot

Right now, the market still prices Sona BLW as what it has always been: an automotive component supplier, valued on automotive multiples, tied to automotive cycles.

Management’s actions suggest they may be underwriting a different future — one where the same engineering expertise that goes into building electric drivetrains also goes into building the nervous system of autonomous machines. Motors that move EVs today could, in a different form, move robots tomorrow. Radar built for keeping a car’s occupants safe could, in a different context, keep a warehouse robot from colliding with a human coworker.

That’s the bet, at least as I read it. Whether it actually pays off commercially is a separate, much harder question — R&D service contracts don’t automatically convert into meaningful revenue lines, and “engineering R&D services” can mean anything from a genuine strategic pivot to a small pilot project that goes nowhere.

The takeaway

Most people will read Sona BLW’s 10X revenue target and form an opinion about the company based on that number alone.

I’d argue the more interesting story is sitting one paragraph below it — in a disclosure most investors will skim past without a second thought.

Ignored today, that perception stack announcement may eventually be remembered as far more than just another engineering contract. Or it may be forgotten entirely, one of a thousand ambitious-sounding corporate announcements that never turned into real revenue.

Either way, it’s worth watching what management does next — not what they say in the investor deck.

If you made it till here, thanks for your time!

Subscribe now

Share

Leave a comment

Net profit up 5%. Provisions down 79%. Advances growing north of 15%.

By lunchtime, the stock is down nearly 5%, and thousands of crores in market cap have evaporated.

If you only read the press release, this makes no sense. If you read the balance sheet the way you're supposed to, it makes complete sense. And the gap between those two readings is basically the entire skill of analyzing a bank quarter.

So let's do it properly.

Rule one: profit is not information until you know where it came from

Every bank quarter deserves the same opening question. Not "did profit grow" — but did profit grow because the bank lent more money well, or because it stopped setting aside money for losses?

Those are very different businesses wearing the same headline number.

HDFC Bank's net profit came in at ₹19,060 crore, up 5% year-on-year. Look one line down and the story shifts. Provisions — the money a bank sets aside for loans that might go bad — fell from ₹14,442 crore to ₹3,060 crore. That's a 79% drop, and it does almost all the heavy lifting behind the 5% profit growth.

Provisions falling isn't inherently bad news. It can mean the loan book is genuinely healthier. But it also means the "growth" in profit isn't coming from the bank doing more of its actual job — lending — better. It's coming from an accounting release.

So the real question becomes: how's the actual job going?

The core engine is the slowest of the big four

Net Interest Income — the difference between what a bank earns on loans and pays on deposits, the purest measure of the lending business — grew just 6.7% this quarter. That's the slowest among India's big four private banks.

Put it next to the same weekend's results and the picture sharpens fast:

• ICICI Bank: NII growth 15.9%

• Axis Bank: NII growth 23%

• Kotak Mahindra: NII growth ~11%

• HDFC Bank: NII growth 6.7%

Same economy. Same quarter. Same interest rate environment. Four different outcomes. The market didn't punish HDFC Bank for existing in a tough environment — every bank was in the same environment. It punished HDFC Bank specifically for having the weakest core engine among its peer set.

That's the mechanism. Markets reward relative strength, not absolute survival.

The number nobody puts in the headline

Net interest margin — how much a bank actually earns on every rupee of assets — came in at 3.26%.

Before the HDFC Limited merger three years ago, HDFC Bank was earning above 4% on this metric. That single gap — from 4%-plus to 3.26% — is the entire post-merger story in one number. And three years in, the recovery is still slow.

This matters because NIM compression isn't a one-quarter problem. It's structural. It tells you the bank absorbed a large, low-margin mortgage book through the merger and still hasn't fully repriced or optimised around it.

And the funding mix just moved against the bank

Here's the part that should worry you more than the headline miss.

• Retail deposit share slipped from 82% to 80%

• CASA ratio sits at 32.3%

• Time deposits — the expensive kind — grew 17.4%

Translate that out of banking jargon: the bank is paying more for the money it lends out. Cheap deposits (current and savings accounts, which pay depositors little to nothing) are growing slower than expensive deposits (fixed/time deposits, which pay depositors a lot more).

Why does this matter right now, specifically?

Crude oil is elevated. Inflation pressure is building back up. The RBI's next move on rates is more likely to be up than down.

When rates rise, deposits reprice fast — banks have to raise what they pay to keep depositors from walking. Loans reprice slower, especially long-tenure ones like mortgages. The bank with the weakest funding mix — the one leaning more on expensive time deposits — feels that squeeze first and hardest.

The market didn't need the RBI to actually move. It just needed to do the math on who's most exposed if it does. That's what Monday's price action looks like.

The fairness point the bears are skipping

Here's where I'll push back on the pessimism, because a good analysis holds both sides at once.

Last year's same quarter included a one-time gain of ₹6,949 crore from the HDB Financial stake sale. That's not a repeatable, structural source of income — it's a windfall. Strip it out, and this quarter's real profit growth is closer to 9.8%, almost double the 5% headline everyone's reacting to.

And zoom out to the balance sheet, and this is not a shaky institution:

• Capital adequacy ratio: 19.6%, against a regulatory requirement of 11.9%

• Gross NPA: 1.17% — genuinely excellent asset quality

• 9,694 branches, roughly half in semi-urban and rural India — a distribution moat almost nobody else in the sector can replicate

This is not a weak bank. This is a strong bank moving through a weak phase. Those are different diagnoses, and conflating them is how you either panic-sell a quality franchise or over-romanticize a genuine structural issue.

The overhang that has nothing to do with this quarter's numbers

There's a second story running underneath all of this, and it's arguably doing more damage to the stock than NII growth ever could.

The chairman resigned suddenly in March. The stock is down roughly 26% year-to-date in 2026. The RBI has stated it found no material governance concerns.

But markets price uncertainty before they price answers. A governance question mark sitting on top of a soft quarter gets punished twice — once for the numbers, once for the doubt. That compounding is why the reaction feels sharper than the results alone would justify.

The actual lesson

Strip away the ticker and the headline, and here's what this quarter is really teaching:

Profit growth is not the same as business growth. One is an outcome; the other is the engine producing it. You have to check which one you're looking at.

Provisions can manufacture a good quarter. A 79% drop in provisioning will always flatter a headline number — that's exactly why it's the first place to look, not the last.

Funding mix decides who survives a rate cycle, not who looks good in a rate-cut environment. Cheap deposits are a competitive advantage that only shows up when rates move against you.

And governance clarity is worth more than any single NIM print. Numbers you can model. Uncertainty about leadership, markets can't price cleanly — so they discount first and ask questions later.

The stock is now roughly 25% below its October high. Pessimism is doing exactly what pessimism does: creating prices that optimism never offers.

Whether that gap is opportunity or a trap isn't something this quarter answers. It's something the next two will.

If you made it till here, thank you for your time.

Subscribe

Share

Comment

But this week, the petroleum ministry published a long, formal defence of E20. And buried inside it was a sentence that had nothing to do with fuel, mileage, or engines at all.

It was about banks.

The ministry’s lead argument: public sector banks have lent nearly ₹1 lakh crore annually to ethanol production and infrastructure. Reverting to E10 would put those investments at risk.

Read that again. A consumer fuel policy is being defended by the size of the loans already sunk into it.

When the strongest argument for a policy is the money already spent, the policy has stopped being about the consumer. It’s about protecting a balance sheet.

Three admissions that were unthinkable a year ago

Buried in the same statement are three concessions that simply did not exist in the government’s messaging twelve months ago.

1. Mileage loss is now official. “In some vehicles there may be a 3–5% reduction in fuel economy,” the ministry said. Compare that to August 2025, when this same ministry called mileage concerns misplaced and pointed to tyre pressure and driving habits instead. The concession is new. Worth noting.

2. The savings story is dead. At $70 crude, E20 costs more to produce than pure petrol — by the ministry’s own admission. The old pitch (E20 saves you money) is officially gone. What’s replaced it is an insurance argument: ethanol protects India if crude spikes to $120.

3. There is still no consumer choice — and the reason given doesn’t hold up. The ministry refused to let consumers choose between pure petrol, E10, and E20 at the pump, citing logistics. But India already runs regular and premium petrol through the exact same supply chain. If E20 could win on merit, offering choice would cost nothing. The fact that choice isn’t offered tells you the real fear: that consumers wouldn’t pick it.

A policy that cannot survive voluntary adoption is telling you something about its underlying economics.

Follow the price, not the press release

Ethanol prices in India are administered by the government — and they move in one direction only. ₹57.97 per litre for C-molasses. ₹71.86 for maize. Farmers have to be kept whole; that’s the political contract underpinning the entire ethanol programme.

So when crude sits near $70 and ethanol still costs more to produce than petrol, someone has to absorb that gap.

It’s either the oil marketing companies buying ethanol at fixed, administered rates — or it’s you, at the pump.

There is no third option.

Why the exit door is welded shut

Here’s the actual signal buried under all the mileage debate: E20 is not getting reversed.

Run the incentive map:

• Distilleries get a guaranteed buyer at administered prices.

• Banks get their ₹1 lakh crore in loan exposure protected.

• Farmers get remunerative, government-fixed rates.

• Consumers get lower mileage and a fuel that costs more to make.

Four stakeholders in this chain. Three of them win by the policy continuing exactly as is. The fourth absorbs the cost, silently, at every fill-up.

Every incentive in this map points away from reversal. Farmers, cooperatives, distillery owners, and PSU bank balance sheets are all locked in together. Once ₹1 lakh crore of public capital is deployed against a policy, that policy is no longer just a policy — it’s a liability the system has to keep defending.

The fight, from here, doesn’t move back to E10. It moves to price.

The crude connection nobody’s saying out loud

By the government’s own math, ethanol only beats petrol on cost once crude climbs to $120–130 a barrel.

Sit with that for a second. E20’s economics are now, functionally, a bet on oil staying expensive.

If crude stays where it is — or falls — the gap between ethanol’s administered cost and petrol’s market cost doesn’t close. It has to be paid by somebody, every single day, at every single pump, indefinitely.

The gap between the statement and the spreadsheet

The official statements talk about scientific evidence, consumer interest, and energy security.

The numbers talk about administered prices that only rise, ₹1 lakh crore of protected bank exposure, and a consumer denied the one thing that would prove the policy’s merit — choice.

This was never just a Citroën problem, or a fuel-quality problem, or an engine-tolerance problem. It’s a financial architecture problem — one where the exit costs now exceed the political will to walk through the door.

E20 isn’t being reversed because the mileage numbers are fine. It isn’t being reversed because ₹1 lakh crore of bank capital, farmer income guarantees, and distillery balance sheets are all sitting on the other side of that decision.

The fuel tank was never the real story. The loan book was.

If you made it till here, thanks for your time!

Subscribe now

Share

Leave a comment

There’s no contradiction. The explanation is one line in the update — and almost nobody stopped to read it properly.

The Surface Story

Revenue: Rs 5,666 crore, up 19% YoY. Store count: 1,312, with 20 net additions this quarter and 19 of them: Zudio.

Read at face value, this is a growth machine running exactly as advertised. Nothing here should move a stock down 11%.

The Line That Actually Matters

Revenue per square foot fell 12.2% year on year.

Sit with that for a second, because it changes the entire read of the quarter. Every rupee of growth this quarter came from new stores. The existing store base — the stores that were already open and already had a customer base — earned less per square foot than they did a year ago.

There are only two ways to grow a retail business: sell more through the stores you already have, or open more stores. On a headline income statement, both look identical.

Underneath, they are not the same business decision at all. One compounds — same fixed cost base, more revenue, expanding margins. The other dilutes — you’re paying for new real estate, new fit-outs, new working capital, just to keep the top line moving.

Trent’s growth right now is entirely the second kind.

Why This Is Happening

Zudio is now at 982 outlets and expanding aggressively into Tier II and Tier III towns.

Smaller catchments, thinner disposable incomes, and — critically — new stores increasingly

opening in markets that overlap with existing ones.Citi flagged this directly: cannibalisation. New Zudio stores aren’t just capturing fresh demand, they’re pulling footfall away from older Zudio stores nearby. Falling revenue per square foot is that entire dynamic distilled into a single number.

This is the tell that separates unit economics from unit count. A retailer can keep opening stores and keep growing revenue for a long time even as the underlying business quietly gets worse per store. The top line hides it. Revenue per square foot doesn’t.

The Valuation Problem

Trent has already corrected 47% from its high of 5,674. Even after that fall, it still trades near 91x earnings.

A multiple like that isn’t pricing in “good company.” It’s pricing in a specific assumption: earnings compounding above 25% annually, sustained for years, with limited execution risk.

This quarter is the first hard data point suggesting the engine is slowing rather than accelerating — and the market’s initial reaction was to reprice that assumption immediately.

Where’s the MOAT?

Value fashion, as a category, has essentially no switching costs. There’s no brand loyalty tax a customer pays to stay with Zudio over an identical t-shirt sold 50 rupees cheaper two blocks away. And that competitive pressure isn’t theoretical — V2 Retail, VMart, Reliance, and ABFRL are all fighting for exactly the same price point, in exactly the same Tier II/III geographies Zudio is now expanding into.

Growth by store count works until it runs into a ceiling of competitive intensity. Falling revenue per square foot, in a category with zero switching costs, is an early signal that ceiling is closer than the multiple assumes.

The Behavioural Trap

An 11% single-day fall feels like a discount. That’s the instinct most investors will act on “quality business, temporary dip, buy the fear.”

On a stock trading at 91x earnings, an 11% fall is not a discount. It’s a repricing. Trent would need to fall by roughly half again just to reach a more defensible 45x — and that math still assumes the growth story holds together. If revenue per square foot keeps declining, even 45x becomes generous.

The Takeaway

Store counts make headlines. Revenue per square foot pays shareholders.

Quantity of growth is easy to manufacture — open enough stores and the top line takes care of itself for a while. Quality of growth is what determines whether that growth actually compounds into shareholder returns, or just dilutes the base it’s built on.

Watch the facts buried in the disclosures. Not the headline growth number management chooses to lead with.

This is not investment advice. Do your own research, or speak with a licensed advisor before making investment decisions

if you made it till here, thanks for your time!

Subscribe

Share

Comment

The number that matters

USD/JPY is trading at 161.96. That’s the weakest the yen has been against the dollar in four decades — levels last seen in 1986. To put that in context, this isn’t a routine drawdown. This is a currency at a structural breaking point, and everyone from Tokyo to Wall Street knows it.

The driver is straightforward on paper: a massive interest rate gap. The Fed is holding rates around 3.50–3.75%, while the BOJ sits near 1.00%. A gap that wide makes holding dollars far more attractive than holding yen, and capital flows accordingly — out of yen, into dollars. That gap is the engine behind this entire move.

Why the BOJ can’t just watch this happen

Here’s the part most coverage skips: a weak yen isn’t just a “currency story” for Japan — it’s an inflation story.

Japan imports the overwhelming majority of its energy and a large share of its food. When the yen weakens, those imports get more expensive in yen terms, and that cost gets passed straight through to consumers. So a falling yen directly feeds domestic inflation.

The problem is, inflation is already the BOJ’s single biggest concern right now — it’s the exact reason they’ve been hiking rates over the past couple of years, breaking decades of near-zero policy to do it. A weakening yen actively works against everything they’ve been trying to engineer. Every day USD/JPY drifts higher, it’s quietly undoing the BOJ’s own policy work.

That’s why this isn’t a “maybe they’ll act” situation. It’s a “they almost have to” situation.

The lever Japan actually has: selling Treasuries

When the BOJ does intervene, the playbook is well established — they did it less than two months ago. The mechanism is:

1. Sell US dollars from reserves

2. Buy yen with those proceeds, propping up its value

3. To generate the dollar liquidity for this at scale, Japan typically sells US Treasury holdings

This is the part that turns a regional currency story into a global one. Japan is one of the largest foreign holders of US government debt on the planet. This isn’t a peripheral player making a symbolic move — it’s one of the biggest holders of US Treasuries deciding to become a net seller, even temporarily.

When a holder of that size sells Treasuries at scale, basic supply-and-demand kicks in: bond prices fall, and yields rise. And rising yields are not a contained event — they ripple through nearly every other asset class:

• Higher Treasury yields → higher borrowing costs across the economy

• Higher yields → equity valuations compress, because future earnings get discounted more harshly (this hits growth and tech names hardest)

• Higher yields → bonds become more competitive with stocks for capital, pulling some flows out of equities

So the chain is: yen weakness → BOJ intervention → Treasury selling → yield spike → equity pressure. That’s domino one.

Domino two: the yen carry trade unwinds

This is the mechanism that actually worries global markets more than the Treasury selling itself — and it’s worth understanding properly, because it’s one of the most important structural trades in global macro right now.

What is the yen carry trade?

For years, Japan has had near-zero interest rates while the rest of the world — especially the US — offered meaningfully higher returns. That created a simple, hugely profitable trade for institutional investors:

1. Borrow yen, because it’s essentially free to borrow

2. Convert that yen into dollars (or other currencies)

3. Use the proceeds to buy higher-yielding assets — US equities, US Treasuries, gold, even crypto

4. Pocket the spread between what you’re paying to borrow yen and what you’re earning on the assets you bought

This trade only works smoothly under one condition: the yen stays weak (or keeps weakening). If the yen is falling, your borrowing cost in real terms is shrinking even further, and your trade gets more profitable over time. It’s a trade built on a falling currency.

What breaks it?

A sudden yen strengthening event — exactly what BOJ intervention is designed to produce. The moment the yen starts appreciating:

• Loans taken out in yen become more expensive to repay (in the currency terms that matter to the borrower)

• The arbitrage that made the trade profitable starts shrinking or reversing

• Leveraged investors are forced to unwind the trade — which means selling the assets they bought with the borrowed yen to raise cash and repay those loans

That selling pressure doesn’t discriminate by asset class. It hits equities, metals, and crypto simultaneously, because all three have been popular destinations for carry-trade capital. This is exactly the kind of cross-asset, correlated selloff that catches markets off guard — not because the fundamentals of stocks or gold changed, but because a funding mechanism underneath all of them got disrupted.

This is also why carry trade unwinds tend to be sharp and fast rather than gradual. Leverage doesn’t unwind politely — it unwinds in a scramble, because everyone is trying to close the same trade at the same time.

The receipts: Japan already tried this, and it barely worked

This isn’t a hypothetical. Japan has already spent real money defending the yen this cycle.

Between late April and late May, Japanese authorities deployed a record ¥11.73 trillion — roughly $72.4 billion — in direct intervention, after the yen first broke past the ¥160 level. That intervention very likely included exactly the mechanism described above: selling Treasuries to fund yen purchases.

And the result? Once that support faded, the yen resumed its slide. It’s now down more than 12% against the dollar over the past year, and sitting at a fresh 40-year low.

That’s the uncomfortable reality here: Japan has already spent a record-breaking sum, and the underlying pressure — the interest rate gap with the Fed — hasn’t gone away. One intervention bought time, not a solution.

The warning shot

Today, Japan’s Finance Minister Satsuki Katayama didn’t soften the language. Asked about the currency, she told reporters:

“We stand ready to take appropriate action whenever necessary.”

In FX policy circles, this is about as close to a direct warning as officials get without naming a specific level or date. Japanese officials have also separately referenced “decisive action, as confirmed between Japan and the US,” and the Chief Cabinet Secretary has said the government wants to build an economy “less vulnerable to foreign-exchange volatility” while staying ready to step in.

Translation: this isn’t a “if” anymore. It’s a “when, and how big.”

So what’s actually downstream of this?

Putting the full chain together:

Yen weakness → BOJ under pressure to intervene → dollar selling + Treasury selling → US yields rise → equity valuations pressured

Simultaneously: yen strengthens from intervention → carry trade unwinds → forced selling of equities, metals, crypto to repay yen loans

Both dominoes point the same direction: pressure on risk assets, broadly, not narrowly. This is why macro traders are watching USD/JPY far more closely than the average retail investor — it’s not really a “Japan trade,” it’s a global liquidity trigger sitting one policy decision away from activating.

What to actually watch from here

• USD/JPY price action — particularly whether it breaks meaningfully past 162, a level multiple analysts have flagged as a pressure point

• Any signs of direct BOJ intervention — sudden, sharp yen moves intraday are usually the first tell, before it’s officially confirmed

• US Treasury yield moves, especially on days yen volatility spikes — that’s the tell for Treasury-selling-driven intervention

• Risk asset correlation — if equities, gold, and crypto start selling off together without an obvious fundamental trigger, the carry trade unwind is a prime suspect

This is a setup worth tracking over the coming days and weeks, not just today. Intervention, when it comes, tends to be sudden — and the assets connected to it don’t wait for a press conference to react.

If you made it till here, thanks for your time!

Subscribe now

Share

Leave a comment

There are two reasons. And they're both happening at the same time.

---

1. The US Fed Angle: Bear Flattening

Start with the yield curve. Since Kevin Warsh was named the next Fed chairman, the curve has been bear flattening.

In plain terms: the yield curve compares short-term and long-term interest rates. Normally short-term rates sit below long-term ones. "Flattening" means that gap is shrinking — and "bear" flattening specifically means it's shrinking because rates are rising, with short-term rates rising faster than long-term ones.

What that signals: liquidity in the banking system is tightening, right now. Not a growth-fear signal — a tight-money signal.

Tightening liquidity is generally bullish for the Dollar. Right on schedule, the Dollar just hit a fresh 52-week high.

A strong Dollar plus tightening liquidity is bad news for risk assets broadly — but metals and Bitcoin are always first to take the hit. If this bear flattening continues, the Dollar keeps grinding higher, global liquidity keeps tightening, and that becomes a drag on growth everywhere.

---

2. The PBOC Angle: A Bigger Story

This is the part most people are missing — and it's the bigger driver.

China has been the single largest marginal buyer of gold globally. Why? Because real estate and equities aren't where Chinese capital wants to sit right now. So liquidity that would otherwise chase those assets flows into gold instead.

This wasn't passive. It's been a deliberate, quiet strategy on two fronts:

- Devaluing the Yuan against gold — a way of working out of a domestic debt trap.

- Reducing their stake in US Treasuries — a slow-motion attempt to weaken the Dollar and chip away at dollar dominance globally.

It worked. For a while.

Then, in the first week of March 2026, PBOC suddenly stopped expanding liquidity. Possibly tied to the Iran war, with central banks needing gold reserves freed up to pay for oil.

So here's the real timeline: gold's weakness didn't start this week. It started in March, the moment China stepped back.

---

Putting It Together

PBOC pulled back in March. Now the Fed's bear-flattening curve and a roaring Dollar are piling on top of an already-weakened market. Two separate forces, same direction, same victim: metals.

My take: a bear-flattening curve doesn't actually fix anything — it just makes US debt worse over time. This looks more like posturing than a durable policy stance.

On the PBOC side, the liquidity switch can flip back on anytime. When it does, that's likely your sustainable bottom in metals with a fundamental standpoint.

---

Share this with someone who's probably crying over their silver and gold positions right now. Hope it helps. If you made it till here, thank you for your time.

Subscribe

Share

Comment

Infosys down 8%. TCS down 6%. HCLTech down 5%. Wipro down 4%. All in one morning.

The story being sold across every business channel and brokerage note this morning is simple: AI is eating IT services.

The facts say something messier.

What actually happened

Accenture — the global bellwether every Indian IT stock gets benchmarked against — fell 17% overnight. Indian IT opened red across the board on the read-through.

But here's what Accenture actually reported for the quarter ended May 31:

- Revenue: $18.72bn, up 6% in dollars

- EPS: $3.80, a beat

- Operating margin: 17%, expanded 20bps

Read that again. Revenue grew. Earnings beat. Margins expanded. This is not a company falling apart. So what spooked the market into wiping 17% off the stock in a single session?

The three real reasons, in order of how much they matter

1. Bookings fell: New work came in at $19.3bn versus $19.7bn a year ago. That's the leading indicator — it tells you what revenue looks like two or three quarters from now, and right now it's slowing.

2. Guidance got trimmed: FY26 revenue guidance had its top end cut from 5% to 4%. Not a collapse. A trim.

3. US federal contracts are getting cancelled: Spending cuts under DOGE are dragging roughly 1% off growth.

Notice what's not on that list. "AI replaced our developers" is not why Accenture guided down. In the company's own commentary, AI is framed as a demand driver that simply isn't large enough yet to offset weak discretionary spend and delayed deal closures.

That's a demand cycle problem with an AI question sitting on top of it. Those are not the same thing, and conflating them is exactly how a sector ends up overcorrecting.

Why Indian IT fell harder than Accenture

This is the part that should bother you if you're holding these names.

The federal contract cancellations are a US government spending story. But Indian IT majors aren't insulated from US federal exposure the way the "it's an American problem" framing suggests — a meaningful share of their client base is large American enterprises, and softness in US corporate spending (federal cuts feeding into broader budget caution) flows through indirectly even without direct government contracts on their books.

Still, that's a difference in degree of exposure, not a reason for Infosys to fall harder than Accenture itself. And that's exactly what happened. Brokerage notes naming Infosys as a preferred sector pick landed the same morning Infosys was the worst performer in the index, down 8%.

That's not analysis. That's sentiment dressed up as analysis.

Separating fact from assumption from guesswork

If you're holding these stocks, here's the honest breakdown:

Fact: Bookings and guidance are soft right now. Demand is genuinely weak.

Assumption: This is mostly cyclical and recovers as rate cuts return and corporate spending normalizes.

Open question: How much of the margin pressure is permanent AI-driven substitution versus a normal down cycle.

Nobody has a real answer to that third one yet. Anyone telling you they do is guessing — and you should price their conviction accordingly.

Why the AI question still matters, even though today's move is overdone

A 6-8% single-day move on a sector-wide read-across isn't noise in the sense of "ignore it and move on" — these stocks were already trading at valuations that priced in continued growth, and on a PEG basis they still look expensive relative to a slowing growth trajectory. That's a real repricing question, not just sentiment.

But the AI question is the one worth watching over years, not days. If AI genuinely compresses the billable-hours model that the entire industry is built on, the recovery looks structurally different from every past IT down cycle. The old playbook — demand returns, headcount returns, margins follow — breaks if the same revenue can be delivered with meaningfully fewer billable hours.

That's the scenario that would justify a permanent multiple compression, not just a cyclical dip.

What to actually watch from here

Accenture's exposure mix — heavier on consulting, heavier on US federal — isn't identical to the offshore-heavy Indian majors. Same direction of travel, different magnitude of impact.

The real signal isn't today's stock move. It's what TCS, Infosys, and the rest say in their own guidance over the coming weeks — whether deal pipelines are genuinely shrinking, or just slipping a quarter.

Watch the guidance. Not the stock price.

---

If you made it till here, thanks for your time

Subscribe

Share

Message me

He said the Bank of Japan is now more concerned about inflation running too hot than it is about the economy slowing down.

Read that again.

This is a central bank that fought deflation for so long that an entire generation of Japanese policymakers built their careers around the singular obsession of getting prices to go up. Zero rates. Negative rates. Yield curve control. Unlimited bond purchases. Decades of extraordinary accommodation to solve one problem: prices weren’t rising.

Now they are. And the BOJ is officially more worried about that than anything else.

That shift — quiet, understated, delivered in the measured language of a central bank press conference — may be the single most consequential macro development of 2025.

The Inflation Picture

The numbers aren’t ambiguous.

Japan’s CPI is projected to hit 2.8% this fiscal year and is expected to rise above 3% for a period. The Producer Price Index came in at 4.9% year-on-year in April — the highest reading in nearly three years.

These aren’t surface-level price moves. Upstream pressure is already spreading into plastic products, construction, transport, and food. The pipeline is full and flowing downward into the real economy.

Wages are rising at 5% for the third consecutive year — structurally, not cyclically. Inflation expectations among both firms and households are moving higher, which is the part that should genuinely concern policymakers. Once expectations become embedded, the job of controlling prices gets exponentially harder.

Ueda made the BOJ’s reaction function explicit: if upside risks to prices continue to outweigh downside risks to the economy, they will raise rates.

That’s not ambiguity. That’s a roadmap.

Where Rates Stand — And Why It Matters

Despite three rate hikes since 2024, real interest rates in Japan are still negative. The policy rate currently sits at 0.75%.

A move to 1% at the June 15–16 meeting is now widely expected.

In isolation, 0.75% to 1% sounds trivial. A quarter-point hike from a rate most developed economies would consider emergency-level accommodation.

But Japan is not in isolation. And this is where the story stops being about one country’s inflation problem and becomes about every portfolio, every asset class, and every market on earth.

The Carry Trade: A $4 Trillion Coiled Spring

For decades, the architecture of global capital flows has had one quiet foundation: Japanese interest rates near zero.

Here’s how it worked. Investors — hedge funds, prop desks, institutions — borrowed money in Japan at near-zero cost. They took those yen and deployed them into higher-yielding assets elsewhere in the world: US equities, US tech stocks, emerging market bonds, crypto. The spread between what they borrowed at and what they earned on the other side was the profit.

The trade has one condition for survival: Japanese borrowing costs must stay low.

When the BOJ raises rates, the calculus breaks. The cost of borrowing in yen goes up. The spread compresses or disappears. Investors are forced to unwind — to sell the US stocks, sell the crypto, sell the EM bonds — and convert everything back into yen to repay the original loan.

The unwinding doesn’t happen on an orderly schedule. It happens fast, across every asset class, simultaneously. Not because anyone wants to sell. Because they have to.

The global carry trade built on near-zero Japanese rates is estimated at over $4 trillion.

August 2024: The Preview

We already have a case study.

In August 2024, the BOJ raised rates by just 0.15%. Fifteen basis points.

Within 48 hours, the Japanese stock market posted its single largest one-day crash in history. Global equities sold off. Crypto sold off. Emerging markets sold off. The violence was simultaneous and indiscriminate.

That was 0.15%.

The expected June move is from 0.75% to 1.00% — a larger absolute move, in a global market that is already carrying significantly more stress than it was in August 2024. The US equity market is more stretched. Geopolitical risk premiums are higher. The margin for error is thinner.

If August 2024 was the tremor, June 2025 may be something else entirely.

The Liquidity Trap Inside the Decision

There is one more layer to the June 15–16 meeting that deserves attention and isn’t getting enough of it.

At the same meeting where the BOJ is expected to hike rates, it will also conduct an interim assessment of its bond purchase reduction plan and announce a new guideline for 2027.

Here is the tension embedded in that: if the BOJ raises rates and simultaneously pauses or slows its quantitative tightening — its reduction of bond purchases — that is a direct admission that Japan’s bond market cannot yet stand on its own.

Tighter monetary policy through the rate channel. Continued life support through the bond purchase channel.

Both at once.

That combination tells you exactly how fragile the underlying system is. Japan’s government debt-to-GDP is the highest in the developed world. The BOJ owns a staggering share of the JGB market. Letting rates rise while simultaneously acknowledging you cannot withdraw bond market support is a high-wire act with very little margin below.

What You’re Actually Watching

The June 15–16 BOJ decision is not a Japan story.

It is a global liquidity event dressed in the language of a routine central bank meeting.

Three questions will determine the market impact in the weeks that follow:

Does the BOJ hike? A hold would be a surprise and a relief. A hike confirms the trajectory Ueda signalled this morning.

How much? 25 basis points to 1% is the consensus. A larger move would be a shock. Even the expected move carries significant tail risk given August’s precedent.

What do they say about QT? If they pause bond purchase reduction alongside a rate hike, read the subtext: they know the bond market needs them. That is not reassurance. That is acknowledgment of structural fragility.

Japan spent thirty years trying to make inflation happen.

Now it’s happening. And the tool they’re reaching for — higher interest rates — is the same tool that unraveled $4 trillion in global carry trades the last time they used it.

Watch June 15th.

Not because you’re invested in Japan. But because the money that’s been borrowed there is invested in almost everything else.

If you made it till here, I thank you for your time.

Subscribe now

Share

Leave a comment

The Merger That Changed Everything (And Not in the Way They Said It Would)

When HDFC Ltd merged into HDFC Bank in July 2023, it was sold as a transformational event. A behemoth getting bigger. The best liability franchise in India absorbing the best mortgage franchise. HDFC Bank bulls were triumphant.

What they underweighted — or chose to ignore — was the structural arithmetic of what that merger actually delivered to the bank’s balance sheet.

HDFC Ltd was a mortgage-heavy NBFC. It funded long-duration home loans largely through bonds, debentures, and wholesale borrowings. HDFC Bank, on the other hand, is a deposit-funded institution governed by CRR, SLR, and PSL norms that NBFCs simply don’t face. The moment those NBFC assets landed on a banking balance sheet, the rules of the game changed entirely.

The result? A credit-to-deposit (CD) ratio that went from a comfortable ~85–87% pre-merger to above 110% post-merger — a number that would make any RBI examiner uncomfortable, and did. The banking regulator had already been signalling that system-wide CD ratios were running too hot. For HDFC Bank, this wasn’t a mild overshoot. It was a structural imbalance that the bank now had to unwind, slowly, at cost.

The CD Ratio Problem: More Painful Than It Sounds

A high CD ratio isn’t just a regulatory discomfort. It is a business constraint that plays out in three dimensions simultaneously.

First, you have to slow loan growth. If you cannot fund new loans with new deposits at an acceptable cost, you either shrink the loan book or raise expensive bulk deposits. HDFC Bank chose to slow disbursements — and the data shows it. Loan growth, which once clocked 18–22% CAGR at the old HDFC Bank, has decelerated meaningfully. A bank growing its loan book at 8–10% in an economy where nominal GDP is running at 10–11% is not compounding wealth for shareholders — it is treading water.

Second, the deposit mobilisation drive that followed was not cheap. HDFC Bank needed to raise retail deposits at scale, quickly. That meant competitive rates, higher branch-level incentives, and aggressive term deposit offerings. The cost of funds went up. And it went up at exactly the wrong time — when the broader interest rate cycle was peaking and loan yields were not expanding proportionally. This is the direct mechanism behind NIM compression.

Third — and this is the part nobody talks about enough — slowing disbursements is not a neutral act. It means you are ceding market share. PSU banks, which are flush with government-driven deposits and have been on a credit quality improvement cycle, stepped in. So did private peers. HDFC Bank, the perennial market share gainer, became a market share ceder. That is a qualitative shift that does not fully show up in one quarter’s numbers but has compounding consequences over years.

NIM Compression: Death by a Thousand Basis Points

Net Interest Margin for HDFC Bank has been under persistent pressure since the merger. The merged entity came in carrying HDFC Ltd’s relatively lower-yielding mortgage book at scale. Home loans, structurally, are low-margin assets — secured, long-duration, and priced competitively because every bank wants them for their risk profile. When your largest incremental loan category yields 8.5–9% and your cost of funds is approaching 5–5.5%, the spread is acceptable but not spectacular — and it certainly doesn’t look like the 4%+ NIMs that HDFC Bank used to post when its portfolio was more skewed toward high-yield retail and SME loans.

The NIM trajectory has essentially been a slow grind lower. Management has repeatedly guided for NIM stabilisation. That guidance has been repeatedly pushed forward. Bulls have repeatedly said “the worst is behind us.” Investors who believed that in Q1 FY25, Q2 FY25, Q3 FY25, and Q4 FY25 have had to revise their models each quarter.

There is no obvious near-term catalyst for NIM reversal. RBI rate cuts — if they come — will compress lending yields before they materially reduce deposit costs (given the lag in deposit repricing). The mortgage book doesn’t disappear. The cost of deposit mobilisation doesn’t fall overnight.

Asset Quality: Not a Crisis, But Not Clean Either

HDFC Bank’s asset quality is not blowing up. Let’s be clear about that. The gross NPA ratio is not alarming by Indian banking standards. But the direction of travel matters as much as the absolute level — and the direction has been quietly deteriorating.

Slippages in the unsecured retail book — personal loans, credit cards — have been creeping up. This is not unique to HDFC Bank; the entire industry has seen stress build in the post-COVID over-extension of unsecured credit. But HDFC Bank, which was aggressively growing its credit card and personal loan business as a high-yield lever to offset the low-margin mortgage drag, is now sitting on a vintage of loans originated in 2022–2023 that is seasoning poorly in some segments.

Credit costs — provisions as a percentage of loans — have been running slightly above what the old HDFC Bank would have tolerated. Not alarmingly so. But in a stock where the premium valuation was always justified by the argument that it is the cleanest, tightest-risk-managed bank in India, any deterioration in credit costs deserves scrutiny, not dismissal.

The concerning part is structural: the bank is simultaneously navigating slower growth (which means fixed costs are being absorbed over a smaller revenue base), margin compression, and rising credit costs. These three vectors moving in the same direction at the same time is a recipe for operating leverage working in reverse — something HDFC Bank’s valuation history has never had to price in before.

Decade-Low Valuations: A Value Trap Wearing a Value Stock’s Clothes

HDFC Bank currently trades at roughly 1.9–2.1x price-to-book, depending on the day. For a bank that spent most of the last decade at 3.5–4.5x P/B, this looks cheap. Every fund manager presentation includes a chart showing the valuation mean-reversion opportunity. “Trading at a discount to its own history” is practically a pitch deck template at this point.

But cheap relative to history only means something if the business is the same business it was when it commanded those valuations. It is not.

Pre-merger HDFC Bank was a deposit-funded, retail-heavy, high-NIM franchise with industry-leading loan growth and pristine asset quality. It deserved a premium. Post-merger HDFC Bank is a much larger institution with a structurally different loan mix, a challenged CD ratio, lower NIMs, slower growth, and the operational complexity of integrating what was effectively a parallel financial institution.

When the underlying business changes, the historical P/B multiple ceases to be a meaningful anchor. You are not buying a “cheap” version of the old HDFC Bank. You are buying a new bank that happens to share the same name and needs to prove it can earn the right to re-rate.

The re-rating will require one of two things: a sustained pickup in loan growth (not happening near-term given the CD ratio constraint), or a significant NIM expansion (unlikely without either rate cycle tailwinds or a meaningful shift in loan mix). Neither is visible in the next one to two quarters. Which means the stock can stay “cheap” for longer than most models assume.

That is the definition of a value trap — not a stock that is fraudulent or fundamentally broken, but a stock where the fundamental improvement required to justify re-rating keeps getting pushed further into the future.

Why Mutual Funds Won’t Say This Out Loud

HDFC Bank is the single largest holding in almost every large-cap and flexi-cap mutual fund in India. Its weight in the Nifty 50 sits around 13–14%. Its weight in the Nifty Bank Index is even higher. This creates a structural conflict of interest that shapes how fund managers talk about the stock.

If you are a fund manager who holds 8–10% of your AUM in HDFC Bank, you have a strong incentive to frame the stock as “attractively valued” regardless of the fundamental picture. The alternative — acknowledging that a decade-long compounder is facing structural headwinds that may suppress returns for two to three years — creates redemption pressure, benchmark tracking concerns, and uncomfortable client conversations.

The language they use is carefully chosen. “Best-in-class franchise.” “Cyclical headwinds, not structural.” “Compelling risk-reward.” “The patient investor will be rewarded.” These are not analysis. These are holding pattern narratives designed to keep clients comfortable while the manager waits for the thesis to play out.

Notice what they do not say: what is the expected loan growth CAGR for the next three years? What NIM do you expect at steady state, and when do you expect the bank to reach it? What is your credit cost assumption, and what is your base case for RBI regulatory action on CD ratios? How do you think about market share dynamics if PSU banks continue their current trajectory?

When fund managers say HDFC Bank is cheap, ask them those questions. The discomfort in the answers will tell you more than the valuation multiple.

The Honest View

HDFC Bank is not a broken bank. It is not going to blow up. The management is competent, the liability franchise remains formidable, and the long-term structural demand for financial services in India remains intact.

But a good bank and a good investment are different things, especially at the wrong point in the cycle.

The honest view is this: HDFC Bank is a bank that needs three to four years of patient execution to work through the structural consequences of its own mega-merger. During that time, loan growth will be sub-optimal, margins will be under pressure, and credit costs will be elevated relative to the bank’s own history. The stock may not go to zero. It may not even fall from here. But the expected return profile for the next 18–24 months does not justify the “no-brainer buy” consensus that has calcified around it.

Decade-low valuations are compelling when the business is temporarily impaired. They are dangerous when the impairment is structural and the duration of recovery is uncertain.

Right now, with HDFC Bank, the honest answer is: we don’t know how long this takes. And in a market full of businesses growing at 18–25% with cleaner balance sheets and no integration overhangs, that uncertainty should demand a far more critical eye than the one currently being applied.

The market labels it cheap. The smarter question is: cheap relative to what future?

Disclaimer: This is not a buy/sell recommendation. Please consult your financial advisor before taking an action. I will not be liable for any profits or losses you incur after taking an action on the basis of this content project.

If you scrolled till here, thank you for your time.

Subscribe now

Share

Leave a comment

The Shanghai Composite Index is 33% below its 2007 peak.

Not 33% below an all-time high from last year. From 2007. Eighteen years ago. A time when the iPhone had just launched, when Lehman Brothers was still standing, when India’s Sensex was at 20,000.

In the same eighteen years, the S&P 500 has gone up over 500%. The Sensex has delivered over 400% returns. Even European markets, perpetually written off as sclerotic and slow, have delivered meaningfully positive returns.

China’s main index has gone backwards.

And this week, Beijing made sure it stays that way.

What Actually Happened

On May 27th, 2026, the China Securities Regulatory Commission launched what analysts are already calling its most aggressive cross-border enforcement action ever.

The targets: Futu Holdings, Tiger Brokers, and Longbridge — the three platforms that tens of millions of Chinese retail investors use to access US and Hong Kong-listed stocks. Think of them as China’s version of Zerodha or Groww, except they gave Chinese investors something Zerodha can’t — a doorway out.

The CSRC’s order was immediate and total:

• Mainland clients are banned from adding new capital or buying new positions on these platforms, effective immediately

• Users can only sell existing holdings

• After a two-year wind-down period, these platforms must completely shut down all mainland-directed apps, websites, and data infrastructure

Eight separate government departments issued a joint statement in support of the crackdown simultaneously. This wasn’t a rogue regulator overstretching. This was a coordinated, state-level decision approved at the highest levels of the Chinese government.

The market’s reaction was instant. ¥1.44 trillion was wiped from Chinese stocks in a single session. The Shanghai Composite fell 1.21%. The Shenzhen Component fell 0.88%.

The Real Reason This Happened

Here’s what the official statement won’t tell you.

An estimated $1 trillion in capital left China last year through these exact platforms as Chinese retail investors moved their savings offshore. Not institutional money. Not foreign hot money. Ordinary Chinese people, voting with their savings, choosing to put their money into Apple and Alibaba Hong Kong rather than A-shares.

That should have been a signal to Beijing. A clear, unmistakable message that its own citizens don’t trust its own markets enough to keep their money there.

Beijing’s response was not to fix the markets. Not to improve corporate governance. Not to strengthen property rights or limit arbitrary regulatory intervention.

The response was to lock the exit door.

Instead of asking “why are people leaving?” — they asked “how do we stop them from leaving?”

That distinction matters enormously. One is a country trying to become more competitive. The other is a country trying to become more captive.

Instagram

The Brutal Irony: The Economy Is Actually Fine

Here’s what makes this genuinely maddening from a markets perspective.

China’s economic fundamentals right now are not bad. They’re actually good.

• Industrial profits grew 18.2% in the first four months of 2026 — accelerating from 15.5% in Q1

• GDP expanded at 5% in Q1 2026

• Manufacturing output is strong, exports have held up despite tariff pressures, and domestic consumption is recovering

On any conventional macroeconomic scorecard, China’s economy is outperforming most peers right now.

None of it mattered. Not for a single trading session.

Because in China, the government’s policy decisions override every economic datapoint. You can have the best earnings season in five years. You can have accelerating industrial profit growth. You can have 5% GDP. And a single regulatory announcement from eight coordinating ministries can undo all of it before lunch.

This is the fundamental problem. Not the economy. The unpredictability of the state.

The Pattern That Never Breaks

If you’ve watched Chinese markets for any length of time, this week feels familiar. Because it is.

The playbook repeats with almost mechanical regularity:

1. Macro data improves

2. Retail and foreign investor sentiment recovers

3. Markets build genuine upward momentum

4. Government intervention resets everything to zero

5. Repeat

In 2021, it was the EdTech crackdown — Beijing wiped out an entire sector overnight. Companies like TAL Education and New Oriental lost 90%+ of their market cap in weeks. Tens of billions of dollars of foreign investment were destroyed.

In 2021-22, it was the tech crackdown — Alibaba, Tencent, Didi, Meituan. Again, coordinated regulatory pressure that turned global darlings into uninvestable stocks.

In 2023-24, it was property sector instability and deflationary pressure that paralysed domestic consumption and market confidence.

And now, in 2026, it’s the cross-border brokerage crackdown.

Every. Single. Time. Chinese markets start building real momentum, the same regulatory reflex activates and resets the counter.

The Shanghai Composite isn’t 33% below its 2007 peak because of bad luck. It’s there because of a structural, repeated, systemic pattern of government intervention that makes long-term equity investing in China a fundamentally different — and fundamentally riskier — proposition than anywhere else in the world.

What This Means for Indian Investors

You might be wondering why this matters to you.

A few reasons.

First, capital flows. When Chinese retail investors lose access to offshore markets, some of that capital looks for alternative channels. India, as one of the few large emerging markets with open capital account access, liquid markets, and genuine rule of law around equity investing, becomes relatively more attractive to global allocators who are reducing China exposure. This is a slow-burn tailwind for Indian equities that most retail investors aren’t pricing in.

Second, the comparison benchmark. Every time China does this, it reminds global institutional investors why India’s market structure — despite its inefficiencies, despite SEBI’s occasional overreach — is fundamentally more investable. Markets need predictability above all else. Indian investors complain about SEBI often. But SEBI has never woken up one morning and decided to shut down Zerodha in 48 hours. That distinction is worth more than most people realise.

Third, the geopolitical read. This crackdown didn’t happen in a vacuum. It happened in a context where the Chinese government is tightening capital controls more broadly — limiting dollar outflows, restricting offshore investment channels, and reinforcing the message that Chinese savings belong inside China’s financial system. That’s not the posture of a government confident in its economic trajectory. That’s the posture of a government managing a slow-moving confidence crisis.

The Bottom Line

China has a genuinely strong economy. It also has an uninvestable stock market.

Those two things are not contradictory. They are the direct result of a system where the government can change the rules overnight, where investors hold no real property rights, and where the state’s short-term political priorities will always override market fundamentals.

The $1 trillion that left China last year through Futu and Tiger Brokers wasn’t irrational. It was Chinese people making a perfectly rational calculation: their own government’s markets are too unpredictable to trust with their savings.

Beijing’s response to that rational calculation was not to become more trustworthy. It was to make it harder to leave.

That’s not regulation.

That’s a managed cage.

And until that changes — until China decides to compete for its investors’ confidence rather than legislate for their captivity — the Shanghai Composite will keep finding new ways to underperform everything around it.

Eighteen years is long enough to call it a pattern.

If you made it till here, thank you for your time.

Subscribe now

Share

Leave a comment

Putin arrived in Beijing for a two-day summit with Xi Jinping, visiting an ally that barely had time to clear the ceremonial trappings it laid out for US President Donald Trump.

Read that slowly.

Xi hosted the American president. Artillery salutes, CEO delegations, trade frameworks. And the moment Trump’s plane left Chinese airspace — news of Putin’s forthcoming trip arrived one day after Trump departed China following the first presidential visit to Beijing in almost a decade.

This is not scheduling. This is Xi telling Washington — in the clearest diplomatic language available — exactly where China stands. You are one guest. Russia is the other. I host both. I need neither exclusively.

The significance of the Xi-Putin summit runs much deeper — as does its timing.

II. WHAT ACTUALLY GOT SIGNED — THE POWER OF SIBERIA-2

This is the headline that changes the global energy map permanently.

Russia’s state-owned energy company Gazprom announced that a legally binding agreement had been signed for the construction of the massive Power of Siberia-2 gas pipeline, which Moscow has been trying to push off the drawing board for years. The new pipeline will supply 50 billion cubic metres of gas annually from western Russia to northern China.

50 billion cubic metres. Annually. Via Mongolia. This is not a letter of intent. This is a signed, legally binding infrastructure commitment that will take years to build and decades to run.

The deal is a major win for Putin, who has increasingly turned to China to replace Europe as its major gas buyer. It’s also a collective show of defiance against US President Donald Trump, who is pushing countries to cut Russian energy imports as part of his attempt to end the war in Ukraine.

Beyond the pipeline, the scale of agreements is staggering. The visit is expected to result in a substantial package of agreements — some 40 documents, 21 of which will be signed in the presence of the presidents — including a Joint Statement on further strengthening the comprehensive partnership and strategic cooperation.

Putin had signaled last week that Russia is close to a “serious” gas and oil deal with China, saying “we are at a very advanced stage of agreement on making a serious, very substantial step forward in the gas and oil sector.”

III. THE REVERSE NIXON — AND WHY IT JUST FAILED

In 1972, Nixon flew to Beijing to split China from the Soviet Union. It was the most consequential diplomatic manoeuvre of the Cold War — driving a permanent wedge into the communist bloc and reshaping the entire balance of power.

Trump’s strategic advisors have theorised for years about a Reverse Nixon — getting Russia away from China to isolate Beijing, the same way Nixon got China away from the USSR to isolate Moscow.

Putin landing in Beijing the day after Trump left is Xi’s direct, public, ceremonially staged answer to that strategy:

No.

Putin and Xi spent hours together — meeting with Mongolia’s president, sitting down for formal talks, and sipping tea at the Chinese leader’s official residence — in the latest show of solidarity between the two strongmen seeking to present a new world order.
The Reverse Nixon is dead. Xi will not trade Russia for American concessions. He doesn’t have to. He just demonstrated he can have both conversations, back to back, in the same hall.

IV. WHAT THE “NO LIMITS” PARTNERSHIP NOW MEANS

Xi’s “no limits” alliance with Putin — announced just before the full-scale Russian invasion of Ukraine in 2022 — has undercut China’s stated position as a neutral mediator.

The partnership declared in February 2022 — days before Russian tanks rolled into Ukraine — was widely dismissed in Western capitals as rhetorical. Three years later, with Putin receiving artillery salutes at the Great Hall of the People immediately after Trump’s visit, the “no limits” framing has hardened into something structurally real.

What it now encompasses:

Energy — Power of Siberia-2 locks Russia and China into a 50 BCM/year gas relationship for decades. Russia’s energy pivot from Europe to China is no longer a threat or a trend. It is signed infrastructure.

Finance — Yuan-ruble settlement mechanisms, alternatives to SWIFT, de-dollarisation architecture. Russia needs Chinese financial rails. China needs to prove those rails work under sanctions pressure.

Diplomatic cover — China and Russia both vetoed the UN Security Council resolution on the Strait of Hormuz. They vote together. They veto together. That coordination is now institutional.

Military adjacency — Not formal alliance, but joint exercises, technology transfer discussions, and shared intelligence on Western military capabilities observed in Ukraine and the Middle East.

V. THE IRAN-RUSSIA-CHINA AXIS — NOW EXPLICIT

This is the strategic picture Trump walked into in Beijing — and the one he flew home without resolving.

Three powers, each under US pressure, each with complementary leverage:

• Iran controls the Strait of Hormuz — energy chokepoint

• Russia controls European energy alternatives and Arctic routes — geographic leverage

• China controls rare earths, semiconductor supply chains, and global manufacturing — economic chokepoint

None of them need to formally ally. They simply need to coordinate enough to ensure that US pressure on any one of them faces resistance from the other two. That coordination is now visible, staged, and ceremonially affirmed — in Beijing, with artillery salutes.

They also touched on the Russia-Ukraine war, in which China is officially neutral and Xi has presented himself as a mediator. Still, Xi’s “no limits” alliance with Putin has undercut that stance.

Xi presenting himself as a Ukraine mediator to Trump while simultaneously hosting Putin with full state honours is not contradiction. It is leverage multiplication. He is the indispensable power in both conversations.

VI. WHAT THIS DOES TO THE TRUMP-XI DEAL FRAMEWORK

Trump flew home from Beijing with broad trade outlines and no resolution on Taiwan or Iran. Although Trump and Xi touted several broad trade deals, they appeared to make little public progress on key sticking points related to Taiwan or the US-Israel war on Iran.

Putin’s immediate arrival reframes what those unresolved sticking points actually mean.

On Iran: Trump asked Xi to pressure Tehran toward a Hormuz ceasefire. Xi can now credibly argue — having just hosted Putin, who has his own deep Iran relationships — that any Iran deal requires Russian buy-in too. The negotiating table just got bigger. And more expensive for Washington.

On Ukraine: Trump has been pushing for a ceasefire. Xi positioned himself as a potential mediator. Putin flying to Beijing immediately after signals that any Ukraine deal goes through Beijing — and Beijing will not deliver it without extracting significant concessions from Washington first.

On Taiwan: The Russia-China partnership strengthens Xi’s hand. Any US military action over Taiwan now risks a world where Russia provides diplomatic cover, Iran activates the Strait, and the entire Eurasian counter-bloc mobilises simultaneously.

VII. THE GLOBAL ENERGY MAP — PERMANENTLY REDRAWN

The Power of Siberia-2 pipeline is not just an energy deal. It is a civilisational infrastructure decision.

Europe spent decades building energy dependency on Russian gas. The Ukraine war broke that dependency — at enormous economic cost to European industry. Europe scrambled for LNG, built new terminals, paid premium prices, and hollowed out its industrial base in the process.

Russia’s response was to pivot east. Power of Siberia-1 was the proof of concept. Power of Siberia-2 — 50 BCM annually — is the permanent infrastructure commitment that makes the pivot irreversible.

What this means globally:

For European energy — Russian gas returns to Europe are now structurally impossible at scale. The pipeline capacity going east removes the optionality. Europe’s LNG dependency is permanent.

For global LNG markets — With Russian gas locked into Chinese contracts, global LNG supply tightens. US LNG exporters benefit. Qatar benefits. Australian LNG benefits.

For China’s energy security — China eliminates a critical vulnerability. It was exposed on seaborne energy — the US Navy, in a Taiwan conflict scenario, could theoretically blockade Chinese oil imports. Overland Russian gas via pipeline is unsanctionable and unblockable.

For the dollar — Russian energy to China will be settled in yuan and rubles. Every BCM of gas that flows through this pipeline is one more transaction leaving the dollar system.

VIII. WHAT THIS MEANS FOR MARKETS

ThemeMarket ImplicationPower of Siberia-2 signedEuropean LNG dependency confirmed — US LNG exporters, Qatar, Australia benefit long-termRussia-China energy lock-inRouble partially stabilises, yuan internationalisation acceleratesReverse Nixon failsChina risk premium stays elevated — no US-China strategic realignment priced inIran-Russia-China axisHormuz reopening becomes more complex, more expensive for Washington40 agreements signedWatch for technology transfer, agricultural, and financial services deals in the detailDe-dollarisation signalLong-term USD pressure, EM currency diversification thesis strengthensUkraine ceasefire complexityDefence stocks — any Ukraine peace deal now requires Beijing’s explicit blessing

IX. FOR INDIA — THE MOST UNCOMFORTABLE POSITION

India sits at the exact intersection of every force in play today and has no clean exit.

India buys discounted Russian oil — and just watched Russia sign a 50-year energy infrastructure commitment to China. The same China with whom India has active border disputes, a 2020 military standoff legacy, and deep strategic competition in South Asia.

India is a Quad partner — and just hosted BRICS where Iran’s FM delivered an anti-US speech. Now Putin lands in Beijing with artillery salutes, deepening the Russia-China axis that India has been carefully navigating around.

India’s strategic autonomy playbook — be indispensable to everyone, commit to no one — becomes harder to run as the blocs harden. The Russia-China axis tightening means the middle ground India occupies gets narrower.

The specific risks for India:

• Russian military technology India depends on increasingly flows through a China-aligned supply chain

• Chinese energy security via Russia pipelines frees up Beijing’s strategic attention for the Indo-Pacific

• The BRICS forum India chairs is increasingly a Russia-China platform with Indian branding

The opportunity:

• Every BCM of gas Russia sends to China via pipeline is Russian gas that doesn’t compete with Indian import alternatives

• Western capital looking for non-China, non-Russia manufacturing destinations accelerates toward India

• India’s irreplaceability in the US Indo-Pacific strategy only increases as the counter-bloc hardens

BOTTOM LINE

Trump visited Beijing and got broad outlines.

Putin visited Beijing and got a signed pipeline, 40 agreements, artillery salutes, and a public reaffirmation of a partnership without limits.

The sequencing is Xi’s answer to every question Trump asked. China does not need to choose between Washington and Moscow. It will take meetings from both. Sign deals with both. And leverage both against each other.

The Reverse Nixon failed before it was fully attempted.

The Iran-Russia-China axis is no longer theoretical — it is infrastructurally, diplomatically, and ceremonially real.

And Xi just demonstrated, in 48 hours, that he is the indispensable power in every conversation that matters.

The great game has a new centre of gravity. It is Beijing. And both Washington and Moscow just flew there to prove it.

If you made it till here, thank you for your time.

Subscribe now

Share

Leave a comment

Japan holds $1.2 trillion in U.S. Treasury bonds.

It is America’s single largest foreign creditor. For three decades, it was the quiet anchor of the global bond market — a massive, stable, long-duration buyer that helped keep American borrowing costs low, funded global risk-taking, and sat underneath the entire architecture of post-2008 financial expansion.

In Q1 2026, Japanese investors sold ¥5 trillion — $33 billion — of U.S. sovereign bonds.

The largest quarterly selloff since 2022.

This is not a trade. This is not a tactical rotation. This is the beginning of a structural exit — and the consequences run through every asset class on earth, including Indian equities.

Why They’re Selling

The reason is simple, and the logic is now irreversible.

Japan’s 10-year government bond yield has crossed 2.564% for the first time this century. By Friday, May 16, it climbed further to nearly 2.7% — the highest level since 1997. Japan’s 2-year yield surged 19 basis points in a single session.

For three decades, Japanese institutions had no reason to keep their capital at home. Domestic yields were zero — sometimes negative. So they exported it. To America, to Europe, to emerging markets. They became the world’s silent lender, deploying savings into any asset that paid more than nothing.

Now domestic yields are rising. The Bank of Japan held its short-term policy rate at 0.75% at the April meeting, but a hike is being priced in at the very next meeting. BOJ board member Kazuyuki Masu has explicitly called for rates to be raised “as soon as possible.” The OECD projects the policy rate could reach 2% by end of 2027. The BOJ has raised its inflation outlook to 2.8% for 2026. Producer prices in Japan rose 4.9% in April — nearly double market expectations.

The era of policy paralysis is over.

And the logic follows directly: when your own bonds yield 2.7% and the cost of currency hedging eats into any yield advantage from holding U.S. debt — why own American Treasuries?

Japan’s institutions are asking that question. And the answer, increasingly, is: we don’t.

Ministry of Finance data shows sustained net sales of foreign securities totaling ¥4 trillion ($25 billion) since the start of 2026, with declines visible across virtually every investor category — including government-related investors who have historically been among the most stable holders of U.S. debt.

The Yen Carry Trade

To understand why this matters to every equity investor — including in India — you need to understand the trade that built the last decade of global risk-taking.

The yen carry trade.

The mechanics: borrow in Japan at near-zero rates. Convert to dollars. Buy U.S. Treasuries, equities, emerging market bonds — any asset with a higher yield than yen funding costs. Pocket the spread. Repeat.

For as long as the BOJ kept rates at zero and the yen stayed weak, this was the closest thing to free money that existed in global finance. The trade grew to an estimated $350 to $500 billion in notional size by 2024. It quietly funded asset valuations across the world — including in India, where a meaningful portion of the FPI inflows that sustained equity markets through the 2020–2024 bull run were carry-funded.

Now the logic has reversed.

Investors who borrowed yen at 0.25% now face funding costs approaching 0.75% — with 2% on the horizon. The margin compression is severe. The trade only worked because Japan was frozen in time. Time is no longer frozen.

The unwind is mechanical, not discretionary.

Sell the foreign assets. Convert proceeds back to yen. Repay the loan.

Margin calls do not wait for good entry points. And every dollar of that unwind flows directly into U.S. Treasury selling — and equity market selling — simultaneously.

As carry trades unwind, the pressure does not stay contained to one market. It cascades: out of U.S. equities, out of U.S. Treasuries, out of emerging market positions. The repatriation of capital to Japan is both a bond market event and an equity market event at the same time.

What This Is Doing to U.S. Yields

When Japan — the single largest foreign holder of U.S. debt — reduces its purchases, the U.S. must find new buyers. It will. But only at higher yields. Every new marginal buyer demands a higher premium. And that upward pressure is now structural, not temporary.

The numbers:

• U.S. 30-year Treasury yield: 5.121% — highest in nearly a year, closing in on multi-decade highs last seen in October 2023

• U.S. 10-year Treasury yield: 4.59% — highest since February 2025

• U.S. 2-year yield: 4.09%

Markets are now fully pricing in at least one Fed rate hike by early 2027. New Federal Reserve Chair Kevin Warsh faces a deeply complicated inflation picture — with war-driven energy costs pushing CPI and PPI higher — while simultaneously dealing with a bond market under structural pressure from its largest foreign buyer stepping back.

The 30-year yield at 5.1% is not just a number. It is the rate at which the U.S. government refinances its long-term debt. It is the benchmark for corporate borrowing. It is the floor for mortgage rates.

When the primary buyer of U.S. debt starts selling, the cost of debt goes up for everyone — governments, companies, and households alike.

The U.S. is currently financing deficits running close to $2 trillion annually on a $36 trillion debt base. The loss of Japan as an anchor buyer is not an inconvenience. It is a structural shift in the cost of American borrowing.

What This Means for Equity Markets

The S&P 500 pulled back from record highs on Friday, weighed down by the yield surge and renewed inflation concerns. After a seven-week rally that pushed the index above 7,500 for the first time, the bond market is finally imposing its gravity on equity valuations.

Here is the transmission mechanism — laid out clearly:

1. Valuation compression. Higher discount rates compress what you pay for future earnings today. This is mathematics, not sentiment. Growth stocks and long-duration tech assets face maximum pressure. The AI rally runs directly into a rising yield environment.

2. Competition for capital. When 30-year U.S. Treasuries yield 5.1%, risk-free returns become a genuine alternative to equities. Capital that was forced into risk assets in the zero-yield world now has options. This rotation is not fear — it is arithmetic.

3. Mechanical selling from carry unwinds. Investors who borrowed yen to buy equities must sell those equities to repay the loan. This is forced, not discretionary. U.S. equities, EM equities, and high-yield bonds are all on the liquidation list.

4. Tighter financial conditions without the Fed moving. Corporate borrowing costs climb. Mortgage rates stay elevated. Consumer spending faces headwinds. Earnings growth projections must absorb all of this — before any Fed action.

The India Angle

India experienced ₹2.6 lakh crore in FPI outflows in calendar year 2026 — the worst annual performance in Indian capital market history.

FIIs built record short positions of 2,27,573 contracts in index futures — an all-time high. A meaningful portion of those outflows was carry-trade repatriation. Yen-funded money, going home.

As Japanese yields continue to rise and more carry-funded positions unwind, Indian equities remain directly in the crosshairs. The foreign money that helped inflate valuations through the 2020–2024 bull cycle was not permanent capital. It was leveraged carry. And leveraged carry always goes home eventually.

That reversal is not over.

Three Numbers to Watch

The entire thesis compresses into three numbers. Watch them daily.

Japan’s 10-year JGB yield — currently 2.7%. Every tick higher means more selling of U.S. debt, more carry trade margin pressure, more repatriation. A move toward 3% would be a material escalation.

U.S. 30-year Treasury yield — currently 5.12%. If it breaks convincingly above 5.5%, equity markets will react violently. That is the level where the math of equities vs. bonds tilts decisively toward bonds for institutional allocators.

USD/JPY — currently 158.5. A move below 155 — yen strengthening sharply — is your early warning signal. It means the carry trade is unwinding fast. That is when forced selling in global risk assets accelerates.

These three numbers will tell you more about what equity markets do next than any earnings report out of Wall Street.

The Bottom Line

Japan’s exit from the U.S. Treasury market is not a trade. It is a structural reallocation driven by simple economics: domestic Japanese yields now compete with foreign returns for the first time in three decades.

The $33 billion sold in Q1 2026 is a down payment on a much larger shift.

The transmission chain is clear and every link in it is now active:

Japan sells U.S. Treasuries → U.S. yields rise → financial conditions tighten globally → equity valuations compress → carry trades unwind → emerging markets face capital outflows → India bleeds.

The cheapest money in the world just got more expensive.

The entire global financial system borrowed it.

Now it has to deal with the consequences.

If you stuck around till here, thank you for your time.

Subscribe now

Share

Leave a comment