▶️ LISTEN TO THIS ARTICLE

In July 2025, Tata Consultancy Services announced 12,000 layoffs, the first mass layoff in the company's history. TCS wasn't struggling. Revenue was fine. The problem was simpler: one AI-powered platform could now do the work of five engineers. Infosys had already eliminated 26,000 positions in fiscal 2024. Wipro shed 24,500. India's six largest IT firms added just 3,847 jobs in Q2 2025, a 72% drop from the previous quarter.

This is the paradox at the center of India's AI story. The country has the second-largest AI talent pool on Earth, the world's highest year-on-year growth in AI hiring at 33.4%, and ranks third globally in Stanford's AI competitiveness index. It's also watching AI hollow out the $283 billion outsourcing industry that made it a technology power in the first place.

The Investment Gap Nobody Can Close

The numbers tell a stark story. India has attracted $11.29 billion in cumulative private AI investment since 2013. The United States: $470.9 billion. China: $119.3 billion. Even that understates the gap, because in 2024 alone the US poured $109.1 billion into AI, nearly the entirety of India's cumulative total in a single year.

India produced only 74 new AI startups in 2024, compared to 1,073 in the US, 116 in the UK, and 98 in China. The government's own investment, $1.25 billion through the IndiaAI Mission, is a fraction of what competitors are spending. France committed EUR 109 billion. The UAE pledged $148 billion. Saudi Arabia pledged $100 billion for Project Transcendence. Canada put up $2.4 billion. India's $1.25 billion isn't nothing, but in the context of a country with 1.4 billion people, it works out to less than a dollar per citizen.

The market projections are enormous. Fortune Business Insights estimates India's AI market will grow from $13.05 billion in 2025 to $130.63 billion by 2032, a 39% CAGR. But projections aren't products, and the gap between India's research output and its ability to capture commercial value from AI remains wide.

The IndiaAI Mission: Compute First, Everything Else Later

The Cabinet approved the IndiaAI Mission in March 2024 with a five-year outlay of INR 10,372 crore (about $1.25 billion). Nearly half goes to compute infrastructure: 18,693 GPUs deployed through public-private partnerships, with eligible users accessing resources at up to 40% reduced cost. Another $240.7 million targets deep-tech startups.

India's greatest AI threat is also its greatest AI opportunity. One AI platform can now do the work of five engineers.

The mission's highest-profile move came in April 2025, when the government selected Sarvam AI to build India's first sovereign LLM. Collaborating with AI4Bharat at IIT Madras, Sarvam is developing three model variants covering advanced reasoning, real-time interaction, and edge deployment. The models support 10 Indian languages including Hindi, Tamil, Telugu, and Bengali. They received access to 4,000 GPUs for six months to train.

Meanwhile, Krutrim, founded by Ola's Bhavish Aggarwal, achieved unicorn status in record time with LLMs capable of working in 10 Indian languages. These companies are betting that multilingual AI built for India's linguistic diversity will carve out a market that OpenAI and Google can't easily serve. Whether that bet pays off depends on whether Indian enterprises actually buy domestic models instead of defaulting to GPT-4.

The Outsourcing Reckoning

India's $283 billion IT services industry is facing the most serious threat in its history. The sector employs over 5 million people, and AI is compressing what used to require large teams into automated workflows.

The damage is already showing. India's big four outsourcers have essentially stopped hiring. Analysts warn that 400,000 to 500,000 IT jobs could disappear over the next two to three years. Mid-level professionals with four to twelve years of experience are most vulnerable. India's top five IT firms lost over $150 billion in market value in the first nine months of 2025 alone.

The industry's response has been to retrain aggressively. TCS is training 25,000 engineers on Microsoft's Azure OpenAI tools. NASSCOM projects that the industry can reskill 8-10 million professionals for AI-augmented roles by 2030. Companies argue they're not replacing jobs but "recasting talent into higher-order roles." That framing is comforting. It also ignores that higher-order roles require fewer people by definition, a reality the IMF's workforce displacement warnings and the AI coding productivity paradox both underscore.

The irony is sharp: India's greatest AI threat is also its greatest AI opportunity. If Indian IT firms successfully pivot from body-shopping to building AI-powered services, they could command higher margins and deeper client relationships. EY's 2025 report found that 47% of Indian enterprises now have multiple AI use cases live in production, with 58% of Global Capability Centers already investing in agentic AI. But the transition window is narrow, and the competition is moving fast.

Talent: Abundant, Unevenly Distributed, Leaving

The U.S. invests in AI every five weeks what India has attracted in total since 2013.

India's AI talent story cuts both ways. The country's AI talent pool is expected to grow from 600,000-650,000 to over 1.25 million by 2027, according to a Nasscom-Deloitte report. India ranked second globally in AI skill penetration between 2015 and 2024, just behind the US. It contributed 19.9% of all AI projects on GitHub in 2024, second only to the US.

But the demand-supply gap is real. Roles like ML Engineer, Data Scientist, and Data Architect show 60-73% demand-supply disparity. And India's IMD World Talent Ranking dropped to 63rd out of 69 economies in 2025, down from 58th the year before. The ranking measures talent investment, appeal, and readiness, not just volume.

The brain drain compounds things. India produces world-class AI researchers, and many of them leave, a pattern South Korea knows intimately, ranking 35th out of 38 OECD countries in AI talent retention. Creating domestic opportunities that compete with Google, Meta, and OpenAI on compensation and research access remains the hardest part of India's talent strategy. The IndiaAI Mission's compute investments are partly aimed at this: give researchers infrastructure worth staying for.

What India Actually Needs

India doesn't need another market projection. It needs three things that projections can't provide.

First, real venture capital. The $11 billion India has attracted since 2013 is roughly what the US invests in AI every five weeks. Until Indian AI startups can raise Series B and C rounds domestically instead of relocating to Singapore or San Francisco, the startup pipeline will keep leaking.

Second, a credible answer to the outsourcing disruption. The "reskilling" narrative buys time, but 400,000-500,000 jobs don't reskill themselves. India needs specific, funded programs that move displaced IT workers into AI engineering roles at scale, not just corporate training press releases.

Third, multilingual AI that actually works at population scale. Sarvam and Krutrim are promising, but they're competing against companies with 100x their compute budgets. If India's sovereign LLM initiative produces models that government agencies and enterprises actually deploy, it creates a domestic AI industry with real defensibility. If it produces a prestige project that everyone admires but nobody uses, the gap widens.

India has more AI talent than any country except the US and China. It has a billion potential users for AI applications in agriculture, healthcare, and financial services that Western companies aren't building. And it has a $283 billion IT industry that either transforms or shrinks. The ingredients are all there. What's missing is the capital, infrastructure, and urgency to combine them before the window closes.

Sources

Research Papers:

Industry / Case Studies:

Commentary:

Related Swarm Signal Coverage: