The macro is well known and over-written: Treasuries at their highest since 2023, Japanese yields above 3% for the first time since 1996, Brent above $100. I will not add to that pile. The one number to carry is India’s own: between June and August, the RBI raised $127 billion of three-to-five-year dollars through FCNR(B) deposits at 5.5–7.1%, and wrote the currency hedge for free. That is the marginal price of foreign money for a country that imports 90% of its crude and all of its potash.
Why the American cycle is unavoidable — and reversible
Two things are true about the United States at once, and the tension between them is the whole story.
The first is that the AI build-out cannot be switched off. The five largest cloud companies will spend north of $600 billion on capital expenditure this year — a figure that has been revised up at every earnings call — and are borrowing to do it, with $1.5 trillion of debt issuance projected over the next few years. Data centres are power-constrained, multi-year, contracted, and half-built. That capex competes directly with the US Treasury for the world’s savings, and it is the structural reason long yields will stay high whatever the Fed does.
The second is that interest-rate policy can reverse overnight. The Fed sits at 3.50–3.75%, inflation is running near 3.5%, and markets price a 60% chance of a hike on September 16. On the merits, that hike is correct: with oil above $100 you tighten and let the term premium do the work. But the midterm elections are eight weeks away, the President, Vice-President and Treasury Secretary have all publicly urged the Fed not to raise, the administration has drawn a line at 5% on the ten-year, and the Treasury has tripled its buybacks to defend it. Whatever the Fed does in September, the political pressure after November runs one way.
For an Indian investor, the practical translation is this: the capex is a decade-long tailwind for everything that feeds it; the rate policy is a trade that can flip with an election. Position for the first; stay nimble on the second. The businesses that feed the build-out — power equipment, transformers, cables, engineering exporters into US data-centre supply chains — are long-duration assets with a contracted customer, which is the one kind of long duration worth owning now.
The duration paradox
The objection to my own argument: equity is the longest-duration asset there is. A deposit reprices in a year; a ten-year G-sec has a duration of about seven; an equity index, valuing cash flows to infinity, is twenty-five years or more. If rising yields punish far-out cash flows, every stock should lose.
Two things break that logic. Equity cash flows are nominal and grow — inflation lifts the discount rate but also the revenue line, which is why the Nifty rose 4% in 2022 while long gilt funds lost money. And the duration that matters is not the equity’s but the mismatch inside the company you own. Equity is the residual claim, the shock absorber for the gap between how fast a firm’s assets reprice and how fast its liabilities do. Own the equity of companies whose assets reprice fast and whose liabilities are long, cheap and in rupees. Sell the reverse.
What it does to the index is simpler: multiple compression eats earnings growth. In 1994, S&P 500 earnings grew 20% and the index returned 1%; the Sensex went nowhere from 1995 to 1999 while profits grew. Expect the aggregate flat and the dispersion underneath enormous. The money is in the dispersion.
Dollar earners — three tiers, one mismatch
The best mismatch in the market right now is a rupee cost base against dollar revenue. The rupee is down 7% this year and every percentage point is margin. But the three groups that qualify are not equal.
IT services has the rupee tailwind and two headwinds: the same US rates that weaken the rupee cut American discretionary budgets, and AI is eating the labour-arbitrage model from the inside. Services exports still grew 8% in FY26; the question is whether the growth belongs to the incumbents. Treat IT as a currency hedge, not a growth holding.
IP-driven exporters — CDMO and CRDMO — are the cleaner version. Pharma exports reached $31 billion in FY26, with over 60% going to regulated markets. The FY26 growth of 2% flatters nobody, but it is an artefact: American buyers pulled about $1.6 billion forward into FY25 on tariff fears. Underneath, the contract-development pipeline has the China-plus-one flow that IT never had, a rupee cost base, and a customer who cannot switch suppliers in a quarter. That is the sector for which “dollar earner” is a structural description, not a currency bet. It is also the one dollar earner that survives the rate reversal: if the Fed folds after November and the rupee recovers, IT loses its tailwind and the assembler loses its margin, but a five-year development contract with a regulated-market customer does not reprice. Owning the exporter whose revenue is contracted rather than transactional is how you hold the dollar trade without betting on the Fed.
Manufacturing exporters are having their genuine moment — the one part of the export story where “golden time” is not hyperbole. Electronics exports hit $47 billion in calendar 2025, $30 billion of it smartphones, eleven times the 2015 figure; engineering goods, chemicals and auto components set records inside a $863 billion total. Two caveats separate the winners. First, merchandise exports to the US grew under 1% in FY26 against reciprocal tariffs, and the President threatened last week to halt trade with surplus countries — the dollar earner with a single-market dependence is a policy bet. Second, an assembler that imports 70% of its components in dollars has far less rupee leverage than a CDMO whose costs are chemists in Hyderabad. Rupee costs, dollar revenue, diversified buyers: score every exporter on those three.
Domestic winners and losers — in one pass
Winners. Banks in the repricing lag — PSU banks returned 71% in 2022 against 22% for the Bank Nifty, though they booked treasury losses first; year one is margin, year three is deposit cost. Upstream oil and gas, subject to the windfall-tax reflex. Urea, not DAP: urea is cost-plus and passes gas costs to the subsidy, while phosphate importers eat the gap under a fixed nutrient subsidy. Life insurers on non-participating margins. Gold lenders, holding the shortest-duration collateral in Indian finance.
Losers. Oil marketers and city-gas distributors, who buy at world prices and sell at administered ones — the ₹1.4 lakh crore of oil bonds from 2005–10 is still being serviced. Anything whose raw material is a barrel. Vehicle and housing financiers with fixed-rate assets and floating liabilities: IL&FS and DHFL in 2018 failed their funding test, not their credit test. Debt-funded developers and REITs. Long-gestation capex where a 200-basis-point move erases the equity return. Aviation, the double hit of dollar fuel and dollar leases. And small caps — the 2022 result, Smallcap 250 down 5% while the Nifty rose 4%, was about who needed the QIP.
How the portfolio should behave
Shorten equity duration without leaving equities: banks in the lag window, upstream, urea, and the exporters with rupee costs. Run the mismatch test on every holding — promoter pledges, short-tenor NCDs, dollar leases and pending QIPs are short liabilities against long assets, and that is what the market punishes first. On debt, stay short and floating until the turn, then extend hard: the equity bottoms of August 2013 and June 2022 came the month the central bank stopped surprising. Three signals for that turn: crude below $80, the India–US spread back above 250 basis points, and the RBI shifting from absorbing liquidity to injecting it. A fourth, specific to this cycle: the first post-midterm Fed meeting.
India’s 2022 says this market can outperform when global yields rise — a banks-heavy index, low external debt, ₹25,000 crore a month of SIP money, and now $127 billion of fresh dollars, a fifth of reserves raised in one summer. That cushion is real. It is also what a 210-basis-point spread and $100 crude are testing, and the RBI has added $127 billion of par swaps to a forward book already short $107 billion. Buying time at 7% is legitimate. It is not free.
I run a long-only equity portfolio; every incentive I have says equities always win. History says they don’t. The mismatched ones don’t.
The bond market has repriced the price of money for the world, and the RBI has repriced it for India. The only duration you still control is your own portfolio’s.
Disclaimer
Nakul Sarda is Founder and Partner of ProfitGate Capital Services LLP, a SEBI-registered Portfolio Management Service (SEBI PMS Registration No. INP000008233). He writes on Indian equities, geopolitics, and the long structural shifts most market commentary misses. He is on X at @nakul_sarda.
The views expressed in this article are personal and are intended for general information only. They do not constitute investment advice, a recommendation to buy or sell any security, or a solicitation of any kind. Readers should consult their own financial advisors before acting on any information contained herein.