India 2028:
The Intelligence Export Crisis

The Consequences of Abundant Intelligence: An Indian Postscript

India's unemployment rate among urban graduates printed 23.4% this morning. The Nifty 50 is down 33% from its September 2026 highs. The rupee has breached 108 to the dollar. And this quarter, for the first time in thirty years, India posted a current account deficit exceeding 5% of GDP.

The American version of this crisis is about a consumer economy losing its consumers. The Indian version is about something more existential: an entire nation's economic model, built over three decades on the arbitrage of human intelligence, rendered obsolete in under two years.

India didn't just participate in the global knowledge economy. India was the global knowledge economy's back office, its help desk, its codebase, its audit team, its radiologist on the night shift. When the cost of machine intelligence fell below the cost of an Indian engineer, the arbitrage collapsed. And with it, the rupee, the real estate market, the banking system, and the aspirations of 200 million middle-class families.

How It Started: The Arbitrage Dies

India's IT services industry was built on a simple equation: an Indian software engineer costs $25,000–$40,000 a year. An American one costs $150,000–$250,000. For decades, the math was unassailable. The wage gap financed Whitefield's apartment towers, Gurgaon's malls, Hinjewadi's gated communities, and the entire aspiration of India's post-liberalization middle class.

By late 2025, a different equation had taken hold. An AI coding agent, running on a cluster that could be anywhere on the planet, could produce the output of a mid-level developer for under $500 a month. Not $25,000. Not $150,000. Five hundred dollars.

The Indian engineer was never competing with the American engineer. That competition ended a decade ago; India won. The Indian engineer was now competing with electricity and silicon. That competition was over before it started.

TCS Q2 FY27: NET NEW DEAL TCV FALLS 34% YOY; MANAGEMENT CITES "STRUCTURAL SHIFT IN CLIENT DELIVERY MODELS"; ANNOUNCES VOLUNTARY SEPARATION SCHEME FOR 45,000 EMPLOYEES; ATTRITION PARADOXICALLY FALLS TO 6% AS INDUSTRY HIRING FREEZES | Bloomberg, October 2026

The first signal was not mass layoffs. Indian IT firms don't do mass layoffs, culturally or legally. The signal was the bench. By mid-2026, Infosys reported that 28% of its workforce was on bench, allocated to no billable project, drawing salary, waiting for work that wasn't coming. At TCS, the number was similar. Wipro quietly stopped all lateral hiring.

What had changed wasn't that the Indian firms were bad at their jobs. They were excellent. The problem was simpler: their clients no longer needed the jobs done by humans at all.

The Global Capability Centers, the in-house offshore arms of Goldman Sachs, JPMorgan, Google, and Microsoft, told the same story but faster. GCCs had grown to employ over 1.9 million people across India by 2025, becoming the white-hot centre of Bengaluru and Hyderabad's real estate booms. Their parent companies, under immense pressure to show AI-driven efficiency gains, began converting GCC headcount to AI agent capacity. A GCC that employed 4,000 people in 2025 now ran with 1,200 people and a fleet of AI agents that handled everything from code review to financial modelling.

The displacement didn't arrive as a single dramatic event. It was a slow suffocation: a thousand contract renewals that came back 30% smaller, ten thousand requisitions that were never opened, a hundred thousand engineers who went from "on project" to "on bench" to "on voluntary separation" to "on Swiggy delivery."

The Current Account Collapse

To understand why this hit India harder than almost any other economy, you need to understand one number: $245 billion.

That's what India's IT and business services sector exported in FY2025. It was the single largest line item in India's current account. It was what financed India's chronic goods trade deficit: the oil it imported, the electronics it bought, the gold its households accumulated. Without the services surplus, India's external accounts don't balance. Full stop.

India imports roughly 85% of its crude oil. In a good year, the oil import bill runs $150–180 billion. Electronics imports add another $70 billion. Gold another $40 billion. These are structural, not discretionary. India needs dollars to run its economy, and the overwhelming source of those dollars was the labour exported (virtually, via fibre optic cable) by millions of engineers and analysts sitting in Bengaluru, Hyderabad, Pune, Chennai, and Noida.

RBI MONTHLY BULLETIN: SERVICES EXPORTS DECLINE 22% YOY IN Q1 FY28; CURRENT ACCOUNT DEFICIT WIDENS TO 4.1% OF GDP; FOREX RESERVES DRAW DOWN $38B IN 90 DAYS | Reserve Bank of India, August 2027

The math was brutal and immediate. As AI agents replaced Indian service delivery, the dollars stopped flowing. Not all at once (these are multi-year contracts) but the trajectory was unmistakable. New deal signings collapsed. Existing contracts were renegotiated downward at renewal. The revenue that did survive shifted from labour-intensive (high headcount, high dollar inflow) to platform-based (low headcount, lower dollar value).

The RBI's response was textbook: defend the rupee using forex reserves. India entered 2027 with approximately $725 billion in reserves, comfortable by any historical standard. By Q1 FY28, $32 billion had been spent in ninety days, and the rupee had still fallen from 91 to 99. By June 2028, reserves stood at $580 billion and the rupee had breached 108. The psychological barrier of 100 ruptured in March with the velocity of a dam breaking.

A weaker rupee, in theory, should have helped exporters. But India's exporters were the very IT firms whose business was evaporating. A cheaper rupee doesn't help when your client has replaced your team with machines. Instead, the weaker rupee inflated India's import bill, particularly for crude oil, creating a vicious cycle: services income falls, rupee weakens, import costs rise, current account widens, rupee weakens further.

India's external vulnerability, dormant for a decade thanks to the IT boom, had returned with a vengeance. The "fragile five" label from 2013 was back in the Financial Times. The IMF's "preliminary discussions" with New Delhi, mentioned in the Citrini memo, were not about a bailout (India is far too large and far too proud for that) but about a precautionary credit line. The mere fact of those discussions accelerated capital flight.

The Bengaluru Reckoning

Bengaluru is not just a city. It is a financial instrument: a leveraged bet on the permanent demand for Indian software talent, priced into every apartment, every commercial tower, every school fee, every restaurant lease.

In 2025, Bengaluru's residential real estate market was valued at approximately ₹8 lakh crore. Whitefield, Sarjapur Road, Electronic City, Marathahalli: these localities had seen 60–80% price appreciation between 2021 and 2025, driven almost entirely by IT/GCC salary growth. A two-bedroom apartment in Whitefield that cost ₹60 lakh in 2020 was selling for ₹1.1 crore by 2025.

The buyer was typically a 28-year-old software engineer earning ₹18–25 lakh per annum, borrowing 80% of the purchase price on a 20-year mortgage, confident that their salary would grow 12–15% annually because it always had. The bank underwrote the loan against that trajectory. The developer priced the next project against that demand. The state government counted stamp duty revenues against those transactions.

Every link in this chain assumed the IT salary would keep growing.

KNIGHT FRANK INDIA: BENGALURU RESIDENTIAL TRANSACTIONS FALL 38% YOY IN Q4 FY27; UNSOLD INVENTORY RISES TO 42-MONTH SUPPLY; WHITEFIELD AND SARJAPUR ROAD REPORT 18-22% PRICE CORRECTIONS FROM PEAK | Knight Frank, January 2028

The sequence was predictable. Hiring freezes led to fewer new buyers. Voluntary separations led to distress sales. Falling prices led to negative equity, where the loan was worth more than the apartment. Negative equity led to defaults. Defaults led to tighter credit. Tighter credit led to further price declines.

The same dynamics played out in Hyderabad's HITEC City corridor, Pune's Hinjewadi, Chennai's OMR, and Gurgaon's Cyber City belt. Each city had its own version of the same story: a real estate market built on the assumption of permanent IT demand, now discovering what happens when that demand evaporates.

India's housing market is structurally different from America's. There is no equivalent of Fannie Mae or Freddie Mac, no government-sponsored enterprise backstopping mortgages. The exposure sits directly on bank balance sheets, primarily at HDFC Bank (post-merger, India's largest mortgage lender), SBI, ICICI, and the housing finance companies like LIC Housing, PNB Housing, and Bajaj Housing Finance.

The RBI's Financial Stability Report from December 2027 flagged a troubling pattern: non-performing assets in the housing loan book had risen from 1.8% to 3.9% in twelve months, with the deterioration concentrated entirely in IT-corridor PIN codes. The rest of the country's housing loans remained healthy, but the geographic concentration was its own problem. These corridors represented a disproportionate share of total outstanding mortgage value.

The NBFC Tremor, and the Bajaj Doom Loop

India's banking system, to its credit, entered this crisis far healthier than it entered the 2018 NBFC crisis or the 2015 NPA cycle. Gross NPAs had fallen to a decade-low of 2.8% by March 2026. Capital adequacy was strong. The RBI had been conservative throughout the post-COVID period.

But the banking system is not where India's financial fragility lives. It lives in the shadow banking system: the Non-Banking Financial Companies and Housing Finance Companies that have proliferated over the past decade.

NBFCs had grown to account for roughly 25% of total credit in the economy by 2026. Many of them had concentrated exposure to precisely the segments now under stress: personal loans to IT professionals, loans against property in IT corridors, developer financing for projects in the same areas, and (most dangerously) unsecured consumer credit extended on the assumption of continued IT salary growth.

The fintech lending boom of 2022–2025 had layered additional risk. Apps like Slice, KreditBee, MoneyTap, and dozens of others had extended unsecured credit to young IT professionals with minimal underwriting, the assumption being that anyone with an offer letter from TCS or Infosys was money good. When those offer letters started being rescinded and those salaries started being cut, the delinquency rates in fintech portfolios exploded.

RBI DIRECTS THREE NBFCS TO HALT FRESH UNSECURED LENDING; CITES "MATERIAL DETERIORATION IN ASSET QUALITY" AND "CONCENTRATED GEOGRAPHIC AND SECTORAL EXPOSURE"; PAYTM LENDING PARTNERS AND TWO MID-SIZE NBFCS AFFECTED | Economic Times, March 2028

Bajaj Finance became the case study in reflexive doom loops.

By 2026, Bajaj Finance was India's largest non-bank consumer lender with an AUM exceeding ₹3.5 lakh crore. Its loan book was a near-perfect cross-section of India's urban aspirational class: personal loans to salaried professionals, consumer durable financing, two-wheeler loans, home loans via Bajaj Housing Finance, and a rapidly growing unsecured digital lending book. Over 40% of its new customer acquisition was in IT-heavy metros.

When IT-corridor stress hit, Bajaj's playbook was rational. Delinquencies rose in its personal loan and consumer durable segments, concentrated in Bengaluru, Hyderabad, and Pune PIN codes. Credit costs climbed from 1.6% to 2.8% in three quarters. The stock fell 35%. Analysts downgraded the sector. Bajaj's cost of borrowing from the bond market widened 80 basis points.

Bajaj's response was the same as every other company facing margin compression: accelerate AI adoption. The company announced "Project Velocity" in Q3 FY28, a ₹2,400 crore investment in AI-driven operations. AI agents replaced 8,000 collections staff. Underwriting that once required a team of credit analysts was handled by models that processed applications in seconds. Customer service centres in Pune were consolidated from four floors to one. Internal estimates suggested Project Velocity would save ₹1,800 crore annually by FY29.

The savings were real. But they created a second-order problem that no one on the board had modelled.

The 8,000 collections staff who lost their jobs at Bajaj were themselves borrowers. Many held Bajaj personal loans. Many had EMIs on consumer durables financed through Bajaj. Their families lived in apartments with mortgages from Bajaj Housing Finance. The people Bajaj fired to protect its margins were the same people whose loan repayments constituted its revenue. Bajaj was, in a very literal sense, firing its own customers.

Multiply this across the NBFC sector. Every lender was simultaneously cutting headcount and watching its borrower base deteriorate. Each company's individual cost-cutting was rational. The collective result was a system that was cannibalising itself.

BAJAJ FINANCE Q4 FY28: CREDIT COSTS RISE TO 3.4%; GNPA CLIMBS TO 2.1% FROM 0.9% A YEAR AGO; MANAGEMENT GUIDES FOR "ELEVATED STRESS IN SALARIED PROFESSIONAL SEGMENTS" DESPITE ₹1,200CR SAVINGS FROM AI-LED EFFICIENCY PROGRAMME | BSE Filing, April 2028

The IL&FS crisis of 2018 showed how NBFC stress transmits to the wider economy. The mechanism is the same now but the scale is larger. NBFCs borrow from banks and mutual funds. When their asset quality deteriorates, their borrowing costs rise, their ability to roll over commercial paper weakens, and the entire credit chain tightens. In 2018, the contagion from a single NBFC nearly froze the entire corporate bond market.

This time, the stress is not concentrated in a single entity but distributed across dozens of lenders with correlated exposure to the same sector. The RBI has been proactive, imposing lending restrictions and requiring accelerated provisioning, but proactive regulation cannot solve a revenue problem. When borrowers genuinely cannot pay, no amount of regulatory strictness prevents the losses from materialising.

The Market Repricing: Equities, Debt, and the Flight to Gold

The Nifty 50 peaked at roughly 27,000 in September 2026. As we write this, it sits at 18,000. The drawdown of 33% makes it the worst equity market decline since 2008, and unlike 2008, there is no obvious catalyst for recovery.

The mechanics of the sell-off operated through three distinct but reinforcing channels.

First, the foreign investor exit. Foreign Portfolio Investors (FPIs) had been steady buyers of Indian equities through 2025 and early 2026, drawn by the "India structural growth story." But FPI returns are denominated in dollars. An FPI that bought Indian equities in January 2026 at a Nifty of 25,500 and a rupee of 91 now faces a Nifty at 18,000 and a rupee at 108. The equity loss is 29%. The currency loss adds another 19%. Combined, the dollar-denominated return is roughly negative 42%. The hurdle rate for new foreign capital entering India has become punitive. Why take the equity risk in a structurally impaired economy and the currency risk of a weakening rupee when US treasuries yield 4%?

FPI outflows from Indian equities totalled $42 billion in FY28, the largest annual outflow on record. Each wave of selling weakened the rupee further, which made the dollar returns worse, which triggered more selling. The reflexivity was textbook.

NSDL DATA: FPI NET OUTFLOWS FROM INDIAN EQUITIES REACH ₹3.6 LAKH CRORE IN FY28, SURPASSING PREVIOUS RECORD BY 3X; DEBT OUTFLOWS ADD ANOTHER ₹89,000 CRORE; "INDIA ALLOCATION UNDER REVIEW" AT MULTIPLE GLOBAL FUNDS | Livemint, April 2028

Second, the SIP slowdown. This is the channel that surprised everyone, because SIPs were supposed to be India's structural advantage. Monthly SIP flows had crossed ₹31,000 crore by early 2026, powered by a generation of young, salaried professionals who had been conditioned to invest systematically through every dip. The mutual fund industry celebrated this as evidence of a permanent shift in Indian household savings behaviour.

The problem is that a disproportionate share of SIP contributors worked in IT and professional services. When those salaries were cut or eliminated, the SIPs didn't just slow. They stopped. Then, as markets fell further, the redemptions began. A 28-year-old engineer who lost her Infosys job and had ₹8 lakh in a small-cap mutual fund didn't think about long-term compounding. She needed to make rent.

Monthly SIP flows fell from ₹31,000 crore to ₹18,000 crore by Q4 FY28. Net equity mutual fund flows turned negative for the first time since 2020, as redemptions outpaced fresh inflows. The domestic institutional bid that had supported Indian markets through every FPI sell-off for years was no longer there. The floor had disappeared.

Third, the consumption-to-earnings transmission. Slowing consumption from displaced IT workers hit corporate earnings across the board. Not just IT companies, but the banks that lent to them, the consumer companies that sold to them, the real estate developers that built for them. The Nifty's forward earnings estimates were cut 22% between September 2026 and March 2028. Markets were falling on lower multiples and lower earnings simultaneously, the dreaded double compression.

Debt markets told a grimmer story. Corporate bond spreads widened sharply, particularly for NBFCs and real estate developers. Several mid-tier housing finance companies saw their commercial paper rates spike 200–300 basis points above pre-crisis levels. The mutual fund industry, scarred by the Franklin Templeton episode of 2020, began proactively reducing exposure to lower-rated credit, which tightened conditions further for the borrowers that needed liquidity most.

Government bonds, paradoxically, rallied as the RBI cut rates and capital sought safety. But the flight to safety had an unexpected primary beneficiary.

Gold.

Gold has always occupied a unique place in the Indian economy. It is simultaneously a cultural tradition, a store of value, and the informal banking system of rural India. In a crisis defined by rupee depreciation, equity destruction, and credit stress, gold became the only asset class that worked.

In rupee terms, gold rose 40% between September 2026 and June 2028, driven by the combination of global dollar prices holding firm and the rupee losing 19% of its value. A middle-class family that held ₹20 lakh in gold in 2026 now held ₹28 lakh. A family that held ₹20 lakh in the Nifty now held ₹13 lakh. The wealth divergence was staggering, and it flowed along familiar lines: rural India, which held gold, preserved wealth. Urban IT India, which held equities and real estate, saw it evaporate.

MUTHOOT FINANCE Q4 FY28: GOLD LOAN AUM SURGES 34% YOY TO ₹1.12 LAKH CRORE; MANAGEMENT NOTES "UNPRECEDENTED DEMAND FROM URBAN SALARIED SEGMENT"; NEW GOLD LOAN CUSTOMERS UP 48% IN BENGALURU AND HYDERABAD | BSE Filing, May 2028

The gold loan market exploded. Muthoot Finance and Manappuram, India's two largest gold loan NBFCs, posted record growth as distressed households pledged family gold to meet EMIs, school fees, and living expenses. Gold-backed lending, long considered a sleepy, semi-rural business, became the fastest-growing segment in Indian financial services. The irony was bitter: gold, the asset that economists had spent decades trying to wean Indian households away from in favour of "financial savings," turned out to be the only functioning safety net for the new urban poor.

Gold imports, which had been declining as a share of the current account, surged again as households and institutions sought safety. This, of course, made the current account deficit worse, adding yet another loop to the spiral.

The Demographic Dividend Inverts

For two decades, India's most celebrated macroeconomic story has been the "demographic dividend": a young, growing workforce that would power consumption and growth for decades to come. Every sell-side deck, every multilateral report, every government white paper opened with the same slide: India's median age is 28. One million people enter the workforce every month. By 2030, India will be the world's largest labour force.

That story assumed the labour force would have somewhere to go.

India produces approximately 1.5 million engineering graduates every year. In the good years, the IT services industry alone absorbed 400,000–800,000 of them as freshers. GCCs took another 150,000–200,000. Startups, e-commerce, fintech, and the broader digital economy absorbed much of the rest. The pathway from engineering college to IT company to middle-class life was so well-worn it had become a cultural institution, an entire civilizational aspiration compressed into a campus placement season.

In the 2028 placement season, that pathway collapsed.

NASSCOM: INDIAN IT INDUSTRY NET HEADCOUNT DECLINES FOR FIRST TIME IN HISTORY; SECTOR SHEDS 340,000 JOBS IN FY28 AFTER ADDING 5.2M OVER PRIOR DECADE; FRESHER HIRING FALLS 85% FROM FY25 PEAK | NASSCOM, April 2028

The tragedy is not just the current unemployed. It is the pipeline. India's engineering colleges, over 3,500 of them, had scaled to produce talent for an industry that no longer needs talent at that scale. The students entering their third and fourth years in 2028 are training for jobs that will not exist when they graduate. Their families have invested ₹10–20 lakh in their education. The return on that investment, which was supposed to be a ₹4–8 lakh starting salary and a trajectory into the middle class, has evaporated.

The demographic dividend has become a demographic burden. A million young people entering the workforce every month is an engine of growth when there are jobs. It is an engine of social instability when there aren't.

The historical safety valve for Indian unemployment has been the informal economy. And indeed, the same pattern the Citrini memo described in the US (displaced white-collar workers downshifting into gig and service work) has played out in India, but with a critical difference. The Indian gig economy was already saturated. There were already too many Swiggy delivery riders, too many Ola drivers, too many Zepto warehouse workers, even before the IT exodus. The informal sector, which employs roughly 80% of India's workforce, has limited capacity to absorb millions of highly educated, formerly well-paid workers.

The result is not just unemployment. It is a collapse in the economic aspirations of an entire generation: the generation that was supposed to be the dividend.

The Consumption Shock

India's domestic consumption story, the other pillar of the bull case, was always more fragile than the headline GDP numbers suggested. India's per-capita income at purchasing power parity was roughly $10,000 in 2025. The consumption boom of the past decade was driven not by broad-based income growth but by a relatively thin layer of 150–200 million urban, formally employed, digitally connected consumers. These were the people buying iPhones, eating at Zomato restaurants, taking MakeMyTrip holidays, and financing Hyundais on EMI.

A disproportionate share of these consumers worked in IT, financial services, professional services, and the startup ecosystem. Precisely the sectors being disrupted.

India's consumption pyramid is steep. The top 10% of households account for roughly 40–45% of total consumer spending. The IT/professional services class sits squarely in this bracket. When their incomes are impaired, the consumption hit is wildly disproportionate to their share of the population.

AVENUE SUPERMARTS (DMART) Q4 FY28: SAME-STORE SALES GROWTH SLOWS TO 2% FROM 14% A YEAR AGO; MANAGEMENT NOTES "VISIBLE DOWNTRADING" IN BENGALURU, HYDERABAD, AND PUNE CLUSTERS; PREMIUM CATEGORY SALES DECLINE 18% | BSE Filing, May 2028

The consumption unwind moved in waves. First, discretionary spending on travel, dining, and electronics softened. Then, the downtrading began: families switching from branded to unbranded, from restaurants to home cooking, from private schools to government schools. Then, the EMI stress: missed car payments, defaulted personal loans, credit card delinquencies spiking in IT-heavy PIN codes.

The startup ecosystem, which had been the other engine of urban consumption and employment, was hit from both sides. Its customers (urban consumers) were spending less. Its funding model (venture capital predicated on growth metrics) collapsed as investors repriced every business that depended on human intermediation. The Zomato delivery rider who used to be a Cognizant engineer was now delivering to fewer restaurants, earning less per delivery, in an economy where the customer on the other end was also cutting back.

India's premium consumption economy (the Nykaa, Trent, Titan story that had powered the Nifty's rally) turned out to be thinner and more concentrated than anyone had admitted. When that narrow base cracked, it took the market narrative with it.

What Didn't Break

It would be dishonest to tell only the story of collapse. Parts of India have proven remarkably resilient, and some have outright benefited.

Agriculture and rural India, ironically, have been the stabiliser. The two-thirds of India's population that never participated in the IT boom has been largely insulated from the IT bust. Rural consumption, driven by a normal monsoon season and continued government transfer payments (PM-KISAN, MGNREGA), has held steady. The MSP procurement system, whatever its inefficiencies, has maintained a floor under rural incomes.

UPI and India's digital payments infrastructure have continued to scale. Transaction volumes crossed 25 billion per month in 2028. Unlike the US, where agentic commerce began routing around card interchange via stablecoins, India's zero-cost UPI rails meant there was no interchange rent to arbitrage away. India's payments infrastructure, built as public utility rather than private toll booth, has proven structurally more resilient to agentic disruption.

Physical manufacturing (the PLI-backed sectors of electronics assembly, pharmaceuticals, and defence) has continued to grow, partly because these sectors benefit from a weaker rupee and partly because they were never dependent on the same arbitrage. The Apple assembly ecosystem in Tamil Nadu, the pharma clusters in Hyderabad and Ahmedabad, the defence corridor in UP: these are creating jobs, just not at the scale needed to absorb the IT displacement.

India's AI startups have, paradoxically, thrived. Companies building AI-native products, as opposed to providing human labour to Western clients, have attracted significant investment. India's cost advantage has shifted from labour to infrastructure: running AI inference on Indian data centres is cheaper due to lower real estate and energy costs. But this new ecosystem employs tens of thousands, not millions. The math doesn't come close to replacing what was lost.

The Policy Labyrinth

The Indian government's response has been a mixture of genuine innovation and performative chaos, which is to say, exactly what you'd expect.

The RBI has been the most competent actor. Governor Das's successor has managed the rupee decline with a controlled float, using reserves strategically rather than defensively, and has introduced targeted long-term repo operations to provide liquidity to stressed NBFCs. The rate-cutting cycle that began in early 2027 has brought the repo rate from 6.5% to 4.75%, though the transmission to lending rates has been slow, as it always is in India.

The Finance Ministry's response has been more complicated. The February 2028 Budget, widely seen as the government's definitive policy statement on the crisis, contained several notable measures:

The state governments have been less coherent. Karnataka, which derives roughly 25% of its tax revenue from IT-related economic activity, faces a genuine fiscal crisis. Stamp duty collections in Bengaluru have collapsed. GST revenue from the services sector is declining. The state's ambitious infrastructure programmes, including the perpetually under-construction metro, face funding gaps.

The most provocative policy debate has been around data localisation and AI sovereignty. A faction within the government has pushed for mandatory local deployment of AI models serving Indian customers, arguing that India should capture the value of AI inference the way it once captured the value of IT services. Critics point out that forced localisation would raise costs for Indian businesses and slow adoption. The debate remains unresolved, which is perhaps the most Indian thing about it.

The Intelligence Export Premium Unwind

India's economic model since 1991 has rested on a single, extraordinary insight: Indian brains are as good as Western brains but dramatically cheaper. Liberalisation didn't just open India's markets. It opened a pipeline through which human intelligence could flow from India to the world at a fraction of the cost of domestic production.

That arbitrage built modern India. It created a 200-million-strong middle class. It financed the current account. It built the cities. It funded the banks. It powered the stock market. It gave a billion people a reason to believe that education was the ticket to prosperity.

What we are witnessing now is the unwind of the intelligence export premium. Not because Indian engineers became less capable (they remain among the best in the world) but because the cost of the alternative fell below any human wage, anywhere on earth. India competed on cost and lost to something with a cost approaching zero.

The American crisis, as Citrini described, is about a consumer economy losing its consumers. The Indian crisis is about an export economy losing its export. Both are painful. But the Indian version is, in a specific sense, more structurally threatening, because the US can at least print the currency its consumers need. India cannot print dollars.

But (and this is where the scenario permits some carefully guarded optimism) India has been here before. The 1991 crisis, triggered by a balance of payments emergency, led to the very liberalisation that created the IT industry in the first place. India has a historical pattern of using existential economic crises as the catalyst for structural reform that would otherwise be politically impossible.

The question is whether the current political system can execute the same clear-eyed pragmatism that Manmohan Singh and Narasimha Rao demonstrated. The reforms needed this time are harder. In 1991, the solution was to open up, to let market forces do what they do. In 2028, the challenge is to redistribute the gains of a technology that concentrates wealth by its very nature. That requires a state that is simultaneously pro-innovation and pro-redistribution, a combination that virtually no government in history has managed.

But you're not reading this in June 2028. You're reading this in February 2026.

The Nifty is near all-time highs. TCS just posted decent quarterly results. The campus placement season, while soft, has not collapsed. The rupee is at 91. The current account is manageable.

The canary is still alive. But it's coughing.

India has time. Perhaps more than any other country, it has the human capital, the institutional resilience, and the historical muscle memory to navigate a structural transition. The question is not whether India can adapt. It always has. The question is whether it will adapt fast enough, and whether the gains of adaptation will be shared broadly enough to hold the social contract together.

For now, the arbitrage that built modern India is narrowing by the quarter. Every procurement negotiation, every contract renewal, every GCC headcount review is a small test of whether human intelligence retains any premium over machine intelligence. The answers are coming in. They are not encouraging.

The export model that financed India's rise is under existential threat. The next model, whatever it is, has not yet been built. We are in the gap.

The gap is where the story gets written.