On February 4, 2026, the day marking the beginning of spring according to the traditional Chinese solar calendar, Anthropic, based in San Francisco, quietly issued a press release announcing the launch of an enterprise-level AI tool.

Thousands of kilometers away in Mumbai, the chart of India’s Nifty IT Index plunged like a kite whose string had suddenly snapped, dropping nearly 6% in a single day. It was the most devastating one-day decline Indian investors had witnessed since the market meltdown triggered by the COVID-19 pandemic in March 2020.

Three months later, when OpenAI announced that it would invest more than $4 billion to build a massive enterprise AI deployment team, Indian IT stocks once again reacted immediately, falling another 3.7%.

Over the past year, every major frontier AI release announced by Silicon Valley giants has sent precisely targeted shockwaves through India’s stock market.

In a market once widely regarded as a “safe haven for foreign capital,” money is now leaving with remarkable determination. During the first half of 2026 alone, more than $23 billion in foreign investment fled the Indian market, dragging foreign ownership of Indian equities down to 14.7% — a level last seen during the difficult period fourteen years ago.

The Nifty IT Index, often described as India’s version of the “Hang Seng Tech Index,” has been steadily declining for 18 months, suffering a cumulative loss of 49% — almost exactly a halving. The ten largest Indian IT giants have collectively seen more than 19 trillion rupees wiped from their market value.

That amount is roughly equivalent to 40% of India’s entire national fiscal budget last year.

But market crashes never happen without a reason.

For the past three decades, India has benefited enormously from answering global customers’ calls and maintaining the underlying code that powered the digital economy. It rode the wave of demographic dividends, yet quietly found itself standing on the opposite side of technological progress.

When the price of a single AI token became cheaper than the cost of human labor along the banks of the Ganges, the outsourcing assembly line that once supported countless Mumbai middle-class families suddenly lost much of its value in the face of cold computational power.

India — the country with the world’s largest reserve of carbon-based human labor — not only failed to capture the growth dividends of the silicon-based era, but instead became the first unfortunate victim to be harvested by the massive AI-driven disruption.

The End of the Two-Decade “Labor Arbitrage Game”

To understand India’s pain today, one must first understand how it won in the past. The rise of India’s IT outsourcing industry can ultimately be traced back to a software bug known as the “Y2K problem.”

In 1999, outdated computer systems across Western financial institutions, airlines, and power companies faced potential collapse because their legacy code recorded years using only two digits. Companies in Europe and the United States urgently needed massive numbers of programmers to inspect and modify decades-old code. The work was not particularly sophisticated from a technical perspective, but the workload was enormous — essentially the digital equivalent of manual labor.

Indian companies quickly recognized the opportunity. With three major advantages — strong English skills, low labor costs, and the willingness to work overnight shifts — Indian IT firms such as Tata Consultancy Services, Infosys, and Wipro rapidly expanded and transformed India into the “the backend of the world’s digital systems.”

According to a 2025 report by the National Association of Software and Service Companies (NASSCOM), India’s outsourcing industry has grown to an astonishing $280 billion in scale, directly supporting 5.67 million IT professionals.

If one also counts the families behind these engineers, as well as the industries built around them — restaurants, logistics, property services, and other supporting sectors — this industrial chain is tied to the livelihoods of nearly 25 million members of India’s middle class.

More importantly, this is one of India’s few pillar industries capable of generating foreign exchange on a massive scale. IT services and business process outsourcing (BPO) exports together account for nearly one-quarter of India’s total exports of goods and services.

However, once India’s outsourcing business model is examined closely, its simplicity — and fragility — become obvious: Charge by headcount. Bill by the hour.

An American programmer may earn an annual salary of $150,000, while an Indian engineer earns only $15,000 to $20,000. An American customer service representative may earn $40,000 a year, while an Indian counterpart may cost only $6,000.

Indian outsourcing firms win contracts with prices far below Western competitors, then pay local wages to employees in India. Their profit comes from the labor cost gap between the two markets.

It was a perfect arbitrage game — one India played for two decades, generating enormous wealth. Until AI arrived and flipped the entire table.

A 2025 study by researchers from Carnegie Mellon University and Stanford University delivered a devastating blow: AI agents can complete tasks 88.3% faster than humans. On the cost side, the median annual compensation for engineering and data-related positions in India in 2025 was around $22,000, while annual subscriptions for AI coding tools cost only a few hundred to a few thousand dollars.

When an AI that never gets tired, requires no social security contributions, writes code 88% faster than humans, and costs only a fraction of the price, the foundation of India’s role as the backend of the world’s digital systems suddenly collapses.

The chill has already reached the desks of Indian programmers.

On April 29, 2026, global IT services giant Cognizant officially launched a transformation initiative codenamed “Project Leap.” The company reportedly set aside $200 million to $270 million just for severance payments. Although the exact number of layoffs was not disclosed, media reports suggested that between 12,000 and 15,000 jobs would be eliminated globally, with the majority of affected employees in India.

And Cognizant is far from an isolated case. U.S. real estate technology company Opendoor shut down all of its offices in Chennai and Bengaluru, India, while French pharmaceutical giant Sanofi handed over procurement order auditing previously outsourced to Indian teams to SAP’s AI agents.

The impact is also visible in corporate earnings reports. Industry leader TCS saw its FY26 revenue in U.S. dollar terms fall to $30 billion, declining 0.5% year-on-year on a constant-currency basis — marking its first annual revenue decline in years. Wipro reported full-year revenue of only $10.5 billion, down 1.6% year-on-year on a constant-currency basis, showing signs of near stagnation.

Even Infosys, considered the most resilient among India’s IT giants, surpassed the $20 billion revenue mark for the first time, but its constant-currency growth rate was only 3.1%, far below its compound annual growth rate of 13.7% over the previous decade.

The wave of layoffs is becoming increasingly visible. In 2025, global technology companies eliminated approximately 245,000 jobs, with India ranking second worldwide with around 19,000 layoffs. This is particularly striking given that India accounts for far less than 7% of global technology employment, yet contributed 7.8% of global tech job cuts.

More alarming is the reversal of the trend itself. In the previous fiscal year (FY25), India’s five largest IT companies still recorded a net increase of 12,718 employees. By FY26, however, these five companies collectively saw a net reduction of 6,981 employees. TCS alone cut more than 23,000 positions on a net basis, with its total workforce falling from a peak of 614,000 employees to below 580,000. The last time TCS experienced a workforce reduction on such a scale was during the 2008 global financial crisis.

The country that built its middle-class dream on writing code is now being pulled back to reality by AI — with little mercy.

Why Did India Fail to Stay at the Table?

With its old source of income being disrupted, India should, in theory, have been able to turn around and compete for the new opportunities created by the AI era. After all, the country has no shortage of technically skilled engineers.

But the reality is that India never even made it to the table where the cake was being divided.

U.S. asset management firm Altimeter estimated that global AI-related net profits would reach $637 billion in 2026. The United States would capture 49% of that, while South Korea would take 35%. Together, the two countries would account for 84% of global AI profits.

The remaining 16% would be divided among regions including China’s Taiwan, mainland China, Japan, and Europe. Yet, in this long list of beneficiaries, India’s name is nowhere to be found.

Many people simply attribute India’s failure to capitalize on AI to insufficient policy investment or a shortage of computing power. But these are only surface-level explanations. The real problem is deeply rooted in the industrial trajectory India has followed over the past several decades.

Looking back at the last century, Japan, South Korea, and China all went through difficult stages of industrial upgrading. Japan’s path moved from low-cost automobiles to high-quality vehicles and then to semiconductor materials. China’s journey went from contract manufacturing to consumer electronics, and later to internet products and AI.

One thing becomes clear: every step taken by East Asian economies involved “making things.”

But what has been India’s path?

IT services, IT services, and still IT services.

India effectively skipped the industrialization stage and developed its service sector through a “leapfrog” approach.

This was not because Indians were naturally unwilling to build factories. Rather, they were locked in place by their own institutional system.

After gaining independence in 1947, India introduced a highly restrictive “license raj” system. Under this framework, anyone seeking to establish a new factory, expand production capacity, or even change product lines had to obtain approval from the central government.

The system essentially protected existing interests, making it extremely difficult for newcomers to obtain licenses and enter industries.

By the time India finally began dismantling these restrictions in 1991, East Asian economies had already divided up most of the low-end manufacturing opportunities.

During the foreign exchange crisis of the 1990s, India was effectively “forced” to turn toward global IT outsourcing.

Having missed out on the division of labor within global manufacturing supply chains, India naturally became a spectator during the massive AI infrastructure boom.

If hardware was out of reach, what about the software model race?

That path has proven equally difficult to break through.

This is actually rooted in the same logic that caused India to miss the internet era.

During the internet revolution, the United States produced companies such as Google, Amazon, and Meta. China produced Alibaba, Tencent, and ByteDance.

What about India?

A country with one of the largest pools of programmers in the world ultimately produced mainly outsourcing companies.

The core reason is that India’s domestic market has never been able to become a testing ground for product iteration.

A successful software company typically needs to go through a growth cycle: large domestic market → economies of scale → product iteration → global expansion

India appears to have 969 million internet users, but their purchasing power remains weak. Currently, India’s GDP per capita is only around $2,800, while wealth is highly concentrated, with approximately 228 million people still living below the poverty line.

The limited ability of most Indian users to pay has made it extremely difficult for Indian internet companies to replicate the growth model seen in the United States and China, where companies scale through large domestic markets.

Large technology companies in the United States and China typically validate their business models in their home markets before expanding internationally.

Take the United States as an example. SaaS giants usually first find large numbers of high-paying domestic customers before moving overseas. Large American enterprises are willing not only to purchase software but also to become early customers, helping startups continuously refine their products.
India’s situation has historically been the opposite.

A large number of Indian companies have relatively low levels of digitalization, are highly sensitive to software prices, have fragmented procurement processes, and remain more accustomed to customized development and human-based services.

As a result, Indian SaaS companies often have to compete directly in the U.S. market before they have even completed product validation at home.
This also explains a peculiar phenomenon:

Indian SaaS companies are everywhere — nearly 20,000 of them, accounting for roughly one-fifth of the global total — yet very few have grown into $10 billion-scale giants.

If a market cannot even produce major internet platforms built on relatively light assets, how could it possibly support the development of large AI models that require tens of billions of dollars in investment?

As a result, India’s technology giants simply did not invest seriously in AI.

The five largest IT outsourcing giants in India, including TCS and Infosys, have long occupied the top positions among Indian IT companies. Their combined market capitalization once exceeded $500 billion, yet their research and development spending remained extremely limited.

In fiscal year 2008–2009, TCS’s R&D expenditure accounted for only 0.2% of revenue, while Wipro’s was 0.19%. Fifteen years later, the ratio had barely changed. In fiscal year 2025, R&D spending accounted for just 1% of TCS’s revenue, while Infosys and Wipro were both at 0.5%.
By comparison, Microsoft’s R&D spending accounts for 12% of revenue, Google’s for 14%, and Meta’s as much as 25%.

After earning so much money, where did it all go?

The answer: back to shareholders.

Between fiscal years 2020 and 2025, India’s five largest IT companies returned approximately 4.8 trillion rupees to shareholders, equivalent to 87% of their combined net profits.

A payout ratio exceeding 80% is an anomaly even among global technology companies.

Behind this lies a difficult reality: these companies are squeezed between two powerful forces — industrial conglomerates and Wall Street.

Take TCS as an example. Its parent company, the Tata Group, operates a wide range of capital-intensive businesses, including steel and automobiles.

Many of these industries are highly cyclical and some have operated at losses for extended periods, yet they employ hundreds of thousands of workers who cannot simply be laid off.

As a result, TCS must maintain an extremely high dividend payout ratio, continuously transferring its dollar earnings back to the parent group to cover financial gaps elsewhere.

At the same time, Wall Street values these outsourcing companies as high-dividend, “bond-like” assets. If they dared to pour profits into uncertain areas such as large AI model development, investors would have no hesitation in voting with their feet.

Outsourcing giants earn profits, then return them to shareholders. Capital markets reward stable cash flows, pushing valuations higher and higher.
Within this self-reinforcing cycle, no one has the incentive — and no one dares — to take the risks required for genuine innovation.

Ultimately, India’s “absence” from the AI era is not an accident. It is the inevitable consequence of the industrial path it has followed.

When Cheap Labor Becomes a Burden

After all this discussion, what has India’s AI industry actually become?

As of the first half of 2026, India has only three widely recognized AI unicorns.

Sarvam AI is the only company genuinely developing foundational AI models. In June 2026, it completed a Series B funding round at a valuation of $1.5 billion. However, its FY26 revenue was only a modest $54 million.

Krutrim once declared its ambition to compete with OpenAI. But in less than two years, it removed its AI assistant from service, suspended its chip development efforts, significantly reduced its team size, and shifted its focus toward selling AI cloud services. Moreover, 90% of its revenue came from internal transactions within its parent company — essentially money moving from one pocket to another.

The third company, Neysa Networks, is focused on providing computing infrastructure rental services.

The three unicorns that India has produced through nationwide efforts to develop AI have a combined valuation of less than $4 billion.

Placed against the backdrop of the global AI race accelerating at full speed, this scale does not even qualify as truly “entering the game.”

The latest projections from the International Monetary Fund (IMF) have effectively confirmed the scale of this transformation: the organization lowered its forecast for India’s GDP growth in fiscal year 2026–27 to 5.8%, marking the largest downward revision since the pandemic.

The only core reason cited was the permanent contraction facing the country’s pillar industry — IT services exports.

This is not merely India’s story.

It is a harsh economic metaphor.

When an economy treats “cheap labor” as its core competitive advantage for decades and builds an enormous system of interest distribution around it, it is essentially betting against technological progress.

The faster technology changes, the greater the backlash it suffers.

A country holding the world’s largest pool of low-cost human labor has found that, in the AI era, its greatest advantage has instead become its heaviest burden.

The country with the largest number of programmers in the world has, in this way, become trapped in the twilight of the old era.

Editor: LQQ