India: The First Country AI Shorted

TL;DR: India's $283 billion IT services industry — built on selling junior engineer hours to the West — is being systematically dismantled by AI. The Nifty IT index has crashed 43% from its December 2024 peak. TCS, Infosys, Wipro, and LTIMindtree have each lost roughly half their value. This isn't a market correction. It's a structural unwind of the entire labor arbitrage model that defined India's rise. The dragon that India built to replace American programmers just got replaced by a cheaper one.

James here, CEO of Mercury Technology Solutions. Hong Kong — July 9, 2026

Every morning, Shiv — 52, former Oracle engineer, 14 years at the company — opens his laptop and sends out five resumes. This ritual started in April, three weeks after Oracle laid off 12,000 people in India. He was one of them.

His family has lived in the same house for fifteen years. The rent is ₹50,000 a month. One evening, he found himself snapping at his wife for no reason. When Outlook magazine interviewed him, he said something that should hang in every business school:

"We built the technology. We learned it. We developed it. And when they were done with us, they showed us the door."

Same layoff wave. Priyanka, 25, woke up for the gym, checked her email, and learned her job no longer existed. She has two EMIs: one for an iPhone, one for a scooter. ₹20,000 a month. She's burning savings to stay in Bangalore.

This isn't a recession. This is a liquidation.

And it's not just happening to individuals. It's happening to an entire country.

The Purest AI Short in the World

If you wanted to construct the perfect trade to express "AI replaces white-collar labor," you'd go long NVIDIA and short something else. That something else trades in Mumbai.

The Nifty IT index peaked at 46,089 on December 13, 2024. By end of June 2026, it had retraced 43%. In the first half of 2026 alone, it dropped ~30% — the worst-performing sector in India, while the Nifty 50 benchmark fell only ~9%.

The four horsemen — TCS, Infosys, Wipro, LTIMindtree — are each down roughly 50% from their peaks. Combined market cap destruction across the top ten IT firms: ₹19.28 trillion, or over $200 billion. TCS alone has fallen below the ₹10 trillion threshold.

But here's what makes this story interesting: the selloffs weren't random. They were synchronized to AI product launches.

February 4: Anthropic drops a new coding tool that automates legacy system exploration and analysis. COBOL modernization — the bread-and-butter of Indian outsourcing for decades — suddenly looks automatable. The IT sector sells off 15% over the following sessions. ₹5.08 trillion evaporated.

May: OpenAI announces a $4 billion+ "pre-deployment engineer" team that embeds directly inside enterprise clients to rebuild workflows around AI. The market read the signal instantly: high-value consulting, deployment, and transformation projects — the exact projects Indian firms bill premium rates for — may soon bypass Bangalore entirely. Nifty IT drops to its lowest level since May 2023.

June: Accenture plunges nearly 18% in a single session — its largest one-day drop ever. The next morning in Mumbai, Nifty IT falls 6%. Infosys drops 8.19% to a five-year low. ₹1.35 trillion gone in one trading day. Accenture's clients are the same US and European banks, retailers, and manufacturers that Indian firms serve. When the middleman gets disintermediated, the arbitrageur dies first.

The sell-side has stopped making excuses. Jefferies warns that in a worst-case scenario, Indian IT valuations have another 30–65% to fall. Citrini Research predicts accelerating contract cancellations at TCS, Infosys, and Wipro through 2027. Nirmal Bang downgraded TCS from Buy to Sell and slashed its target price from ₹3,046 to ₹1,693.

Bloomberg data shows the combined weight of the top five IT firms in the Nifty 50 has dropped below 7.6% — the lowest since 2002. Capital is voting with its feet. Global investors are systematically shorting a nation's pillar industry.

What India Actually Sells

To understand why India is getting hit hardest, you need to understand what India's IT industry actually sells.

Not software. Not products. Not IP.

Engineer hours. Billed by the hour.

The Y2K panic gave India its first gold rush. Three decades later, the model hasn't fundamentally changed. Client in New York or London. Code written in Bangalore or Hyderabad. Same work, fraction of the cost. Labor arbitrage is the entire operating system of this $283 billion industry.

Neeti Sharma, CEO of Team Lease Digital, summarized the social contract perfectly: "You borrow four or five hundred thousand rupees for an engineering degree, get into TCS or Infosys or HCLTech, and you're set for life."

Pooja's story is the template. Grew up in a single room in suburban Kolkata. Seventy people in her building shared one bathroom. Got her diploma in 2005, moved to Gurgaon as a programmer starting at ₹7,056 a month. Today she earns ₹3.5 million a year at a top-tier IT firm.

Nasscom and Crisil research found that by 2007, every IT job created approximately four additional jobs in the broader economy — drivers, security guards, cooks, domestic workers. Housing loans as a percentage of Indian GDP grew from 0.6% in 1995 to roughly 11% today, with 35% concentrated in the IT-dense south. Bangalore and Hyderabad real estate markets are essentially levered to IT paychecks.

The problem: The product this model sells has a precise name. Repetitive labor by junior and mid-level engineers. Template code. Manual testing. Legacy system maintenance. Ticket processing.

Large language models are the perfect substitute for exactly this type of work. Marginal cost approaching zero. 24/7 availability. No visa required. No visa possible.

India spent thirty years becoming the world's largest "replacement for American programmers." Now it's being replaced by a cheaper replacement. The dragon-slayer didn't become the dragon. It got eaten by a bigger one.

The Middle-Class Decade, Ripped Up in Three Years

The collapse is already in motion.

TCS announced 12,000 layoffs in July 2025 — 2% of its workforce, the largest cut in the history of India's biggest private employer. A 45-year-old Kolkata employee told Reuters: "This is devastating news. At my age, finding new work is incredibly difficult."

Here's the absurd detail: over 500 candidates who received TCS offers with July 2025 start dates are still waiting indefinitely for onboarding. Many had already resigned from their previous jobs.

The hiring engine has stalled. India's top five IT firms reported net negative headcount of roughly 7,000 in FY2026 (ending March 2026). The year before, they had added over 12,000. Average annual gross hiring for these five firms over the past five years: ~230,000. FY2026: 170,000. TCS slashed its fresh graduate intake from an average of 40,000 over the past three years to 25,000.

Gaurav Vasu, founder of market intelligence firm UnearthInsight, estimates 400,000 to 500,000 IT workers face layoff risk in the next two to three years. 70% of them are mid-tier professionals with 4–12 years of experience — the backbone of the pyramid.

Fund manager Saurabh Mukherjea ran the larger math: India produces roughly 3 million engineering graduates annually. About 1.5 million are considered "employable engineers." Before 2020, the IT services sector absorbed nearly all of them. Over the past three years, that absorption rate has dropped to approximately zero.

Meanwhile, Azim Premji University's 2026 India Employment Report shows 40% unemployment among graduates aged 15–25.

The shockwave is propagating backward through the same channels that wealth once flowed through.

Q1 2026: India's major cities saw residential sales drop 13% year-over-year. Analysts explicitly cite IT layoffs as a primary driver. Shared apartments in Bangalore suddenly have vacancies. Landlords are blaming IT firms. Mukherjea identified a more dangerous signal: workers anticipating layoffs are rushing to take out personal loans and home loans before they lose their jobs. Some of India's recent loan growth is literally "doom borrowing."

The Exit Doors Are Welded Shut

So leave India and go to America, right?

That door is closing too.

September 2025: The Trump administration raised H-1B visa fees from $5,000 to $100,000 — a 20x increase. Two months earlier, Trump publicly told Google and Microsoft to "stop hiring in India."

In 2024, Indians received over 200,000 US work visas. Indian companies accounted for 20% of all H-1B approvals. This pipeline was the physical-world extension of the Indian IT model.

Indian IT derives roughly 60% of its revenue from the US market — nearly $135 billion. India now faces a double pincer:

1. AI gives American enterprises the first real "reshoring" option — they no longer need to send work to Bangalore.

2. Visa policy ensures Indian engineers can't easily send themselves to America.

People can't leave. Work can't enter.

And the AI liquidation is still unfolding.

India's median age is 28. For the next two decades, millions of young people will enter the workforce every year.

The demographic dividend is a check with an expiration date. Cash it, and India becomes the next great power. Miss it, and the same young people move from the asset side of the national balance sheet to the liability side.

時代的一粒灰，落在個人頭上，就是一座山。

A speck of dust from the era, when it lands on an individual, becomes a mountain.

Shiv still sends his five resumes every morning. The Bangalore office towers still glow at night. But for the first time, the people inside are seriously asking: How long will those lights stay on? And who will they be shining for?

The Pattern

This isn't about India. It's about the arbitrage model — any business built on the gap between what labor costs in one place and what it costs somewhere else.

AI doesn't just reduce that gap. It eliminates the geography entirely.

India's IT industry was the world's most efficient labor arbitrage machine. It built a middle class, transformed cities, and created a national growth story. But arbitrage is only a moat when the cost differential is durable. When technology collapses the cost of the substitute to near-zero, the arbitrageur becomes the arbitraged.

The same pattern will hit other labor-arbitrage industries. Other countries. Other job categories. The only question is sequence and speed.

The 2026 Equation: $T_{\\text{Labor Arbitrage}} < T_{\\text{AI Substitution}}$

When the time horizon of your cost advantage is shorter than the time horizon of your replacement, you're not running a business. You're running a liquidation clock.

Mercury Technology Solutions: Accelerate Digitality.