Citrini Research's "2028 Global Intelligence Crisis": Should Indian Investors Panic?

Did a Wall Street research note labelled as 'Citrini Research' trigger an ₹84,000 crore single-day wipeout in Indian IT stocks? Here’s what it says and what it gets right and wrong.

On February 24, 2026, Indian IT investors had a bad morning. Nifty IT had already been bleeding for weeks. Then a report from a Substack-based firm called Citrini Research started circulating across Twitter and investment inboxes, and within hours the sector’s combined market cap fell by ₹84,000 crore (not just due to this report).

TCS lost ~3.5%. Infosys dropped ~3.5%, and Wipro fell 2.6%. Nifty IT also hit a 2-year low.

The report’s title: "The 2028 Global Intelligence Crisis."

Headlines started comparing the destruction in market value to the entire GDP of Venezuela and Sri Lanka. That’s a dramatic frame for what is, at its core, a Substack essay, but the market didn’t seem to care about the distinction.

So, is this report credible? And should anyone with Indian IT stocks actually be worried? let's find out

Who Wrote It, and Why Did Anyone Pay Attention?

The report was co-authored by Citrini Research, Alap Shah, and James van Geelen.

Alap Shah is worth understanding because he’s the reason this landed differently from the usual internet doom-posting. Harvard economics background, portfolio manager at both Viking Global and Citadel, co-founder of Sentieo (which raised $70 million before AlphaSense acquired it). He now runs Littlebird and Lotus Technology Management. He spent 15 years building AI-powered financial research tools before writing a report about AI potentially destroying the industry that relies on humans doing financial research manually.

That context cuts both ways. He understands AI capabilities deeply. He also benefits from AI adoption accelerating. Both realities can be true at the same time.

But here’s what got lost in the panic: the report is explicitly a scenario, not a forecast. Shah’s framing describes it as a structured stress test. The entire piece is written as a fictional retrospective from June 2028, describing how things supposedly went wrong over the prior two years.

Most people sharing it didn’t read that part.

What the Report Actually Says

The core argument is that AI could break the economic feedback loop that keeps growth stable.

Companies cut white-collar headcount to expand margins. Displaced workers spend less. Demand softens. Companies invest more in AI to cut further. Repeat until GDP looks stable on paper while real wage income quietly collapses. The report calls this "Ghost GDP."

The Fictional Timeline

• Late 2025: Agentic AI coding tools improve sharply. SaaS pricing collapses as enterprises build internally.
• October 2026: S&P 500 touches 8,000, Nasdaq 30,000. Peak euphoria.
• Late 2026: ServiceNow announces a 15% workforce cut. Stock drops 18%.
• 2027: Weekly US jobless claims reach 487,000. Mortgage delinquencies rise in tech-heavy cities. Zendesk defaults on debt.
• November 2027: Markets crack.
• June 2028: S&P 500 down 38%. US unemployment at 10.2%. Indian IT in freefall. The IMF is in preliminary talks with New Delhi.

The narrative is vivid and internally coherent. It reads more like a financial thriller than an equity research note, which is exactly why it spread so fast.

The India Angle

The India argument hinges on one claim: Indian IT’s export model depends on labour cost arbitrage. If the marginal cost of an AI coding agent collapses to near the cost of electricity, pricing power disappears.

In the scenario, contract cancellations at TCS, Infosys, and Wipro accelerate through 2027. Procurement teams demand 30% renewal discounts. The rupee falls 18% against the dollar. IMF discussions begin.

It’s unsettling to read. The real question is how much of it reflects current data.

The Actual Numbers

Indian IT exported $224 billion in FY2025. The US accounts for 54% of that, which is roughly $103 billion. North America contributes 58% of Infosys revenue and about 50% for TCS. The exposure is real and concentrated.

But here’s what matters: AI currently contributes only 5–6% of revenues at TCS and Infosys.

TCS runs 620 AI engagements, with AI revenue growing 17%+ quarter-on-quarter, roughly $1.8 billion per quarter against a $30 billion annual base. Infosys disclosed AI services at 5.5% of quarterly revenue across 460+ projects. Fast growth. Small base. Both statements are true.

Motilal Oswal estimates 9–12% of IT services revenue could be eliminated over the next 3–4 years. That implies 2–3% annual revenue pressure, which is meaningful but far from collapse.

Jefferies, the most bearish major brokerage, projects 6% earnings CAGR for large-cap IT through FY2028. Their concern is application managed services — maintenance, testing, routine development — which account for 22–45% of revenues and face pricing pressure from AI coding tools.

That concern is legitimate. The idea of complete sector destruction by 2028 is not supported by current evidence.

What the Market Had Already Priced In

The selloff didn’t begin with the Citrini report.

• February 4: Anthropic launches Claude Code. Nifty IT falls 6%.
• February 16–17: Khosla’s comments intensify concerns.
• February 23: Jefferies downgrades six Indian IT stocks.
• February 24: Citrini report spreads. 

By then, Nifty IT had already fallen around 21% in the month — its worst February in 23 years. The sector has lost over $100 billion in market cap over the past year. TCS alone is down roughly $40 billion from its peak.

The Nifty IT P/E has compressed to 23.6x versus a 1-year average of 26.8x, a 12% discount. 

The Historical Context

Indian IT has faced existential predictions before.

Y2K was supposed to be a crisis; it became an opportunity. The dot-com bust hurt Western startups but not Indian IT’s core enterprise clients. After 2008, cost pressure drove more outsourcing. During the cloud wave, firms repositioned as migration partners. When RPA threatened BPO work, Indian IT became automation implementers.

For 25 years, every technological shift has triggered “this kills Indian IT” headlines. The sector repeatedly adapted.

Is AI different? Possibly. The scale and speed are unprecedented. But dismissing decades of adaptability requires stronger evidence than a fictional stress test.

What Should Investors Do?

The bear case has substance. Application managed services face real pricing pressure. Revenue erosion in certain segments is likely already underway. Firms slow to adopt AI-augmented delivery models will lose contracts.

The IMF bailout scenario has no current macroeconomic support. There are no indicators pointing to 10%+ US unemployment or an 18% rupee depreciation by 2028. The authors themselves describe the report as a stress test.

For long-term investors, valuation compression to a 12% discount versus recent averages historically represents reasonable entry levels. Balance sheets remain strong, cash flows healthy, and revenues dollar-denominated.

For traders, volatility will persist. Every major AI announcement or contract renegotiation will trigger sharp moves.

Two metrics matter most in upcoming quarterly results:

• Order book TCV (Total Contract Value)
• AI deal win rates

Those numbers will reveal real disruption far more clearly than viral essays.

Where This Leaves Us

An ex-Citadel portfolio manager wrote a fictional 2028 collapse scenario. It went viral. It spooked markets. It accelerated an ongoing correction.

Some concerns are legitimate. The labour-arbitrage model must evolve. Certain revenue streams will compress. AI coding tools will pressure pricing.

The catastrophic narrative, however, lacks present-day evidence and ignores a 25-year record of adaptation. In fact, Indian IT has been declared dead before. But it adapted then, and it’s adapting now. The only real question is whether it adapts fast enough — and that answer will appear in quarterly numbers, not in a Substack scenario.