India AI Market 2026: 100M ChatGPT Users, $200B in Bets

India’s AI market runs on a gap you could drive a truck through. Domestic AI revenues sit at $10–12 billion in FY26 within a $315 billion technology sector, according to Nasscom’s February 2026 strategic review. Announced AI investment, meanwhile, has blown past $200 billion in commitments pledged at the India AI Impact Summit in February 2026 — Reliance alone promised roughly $110 billion over seven years. That mismatch between money committed and revenue earned is the defining feature of India’s AI economy in 2026.

What makes India distinctive is that it leads the world in usage, not spend. Sam Altman confirmed in February 2026 that India has 100 million weekly active ChatGPT users, making it OpenAI’s second-largest market after the United States (TechCrunch, February 2026). BCG found 92% of Indian employees use AI at work, the highest rate in Asia Pacific (BCG, November 2025). The players to watch: OpenAI, Google, Microsoft, Reliance Jio, Adani, the IndiaAI Mission, and sovereign-model startups Sarvam AI and Krutrim. This is the market deep-dive — for who can actually access which models, see our guide to AI access in India.

How big is India’s AI market, really?

Depends which number you trust. Nasscom’s strategic review puts AI revenues at $10–12 billion in FY26 — a small but fast-expanding slice of India’s $315 billion tech industry, which itself is growing 6.1% year on year (Nasscom, February 2026). An earlier Nasscom–BCG projection had India’s AI market touching $17 billion by 2027. Within the IT services majors, AI is already material: TCS reports an annualised AI revenue run rate of $1.8 billion, while Infosys books $275 million, about 5.5% of revenue (Nasscom strategic review, February 2026).

Set that against the capital flooding in. Commitments announced at the India AI Impact Summit in New Delhi (17–21 February 2026) exceeded $200 billion, spanning data centres, compute and skilling. The revenue-to-commitment ratio tells you India’s AI market in 2026 is an infrastructure story first and a software story second. In our Asia AI market breakdown, India is the region’s paradox: modest current spend, maximal strategic weight. It rhymes with the India gaming market — 591 million players monetising at just $1.5 billion — enormous usage, thin monetisation, and everyone betting the curve inflects.

India is the frontier labs’ biggest growth market

No frontier lab can hit its growth targets without India. The 100 million weekly active ChatGPT users Altman cited in February 2026 came after a deliberate land grab: OpenAI opened a New Delhi office in August 2025, launched a sub-$5 ChatGPT Go tier the same month, then made Go free for Indian users for a full year from October 2025 (TechCrunch, February 2026). India also has ChatGPT’s largest student user base globally, per Altman.

Google matched the play through distribution, striking a deal with Reliance Jio in 2025 to give millions of Jio subscribers free Gemini AI Pro access (TechCrunch, February 2026). The pattern is familiar from a decade of Indian consumer internet: buy usage now, monetise later. Consumers are meeting the labs halfway — BCG’s November 2025 workplace study found 92% of Indian employees using AI at work versus a 72% global average, and 58% of Indian frontline workers say leadership gives them clear AI guidance, roughly double the Asia Pacific norm. India is not a laggard market being evangelised; it is the adoption frontier being subsidised.

What is the IndiaAI Mission building?

The IndiaAI Mission is Delhi’s ₹10,372 crore (roughly $1.25 billion) five-year bet, approved in 2024, that subsidised compute plus sovereign models can keep India from becoming a pure consumer of foreign AI. Its centrepiece is the GPU tender: more than 38,000 GPUs had been onboarded through empanelled providers by Q1 2026 (IndiaAI portal data, via Startup Feed, June 2026), up from about 17,000 after the first two bidding rounds in mid-2025 (Indian Infrastructure, June 2025). The stated target is 100,000 publicly accessible GPUs by December 2026, offered to startups and researchers at ₹115–150 per GPU-hour — around 42% below market rates.

The sovereign-model track delivered its first headline result at the AI Impact Summit: Sarvam AI unveiled Sarvam 30B and Sarvam 105B, open-source models optimised for Indian languages, in February 2026 (TechBuzz, February 2026). The summit itself, attended by Altman, Sundar Pichai and Emmanuel Macron, produced the New Delhi Declaration on AI governance (PIB, February 2026) — a signal that India intends to convene the Global South’s AI agenda, not just participate in it. How durable that positioning proves depends heavily on chips, which still flow through Washington; see our analysis of the US–India tech relationship.

Which Indian AI companies matter?

The domestic ecosystem is broad but top-heavy. India counts more than 1,700 AI-focused startups (Tracxn and Zinnov estimates, via Startup Feed, June 2026), and AI startups raised $1.5 billion in Q1 2026 alone — 38% of all Indian startup funding that quarter. Three names carry most of the narrative weight.

Sarvam AI: the sovereign flag-bearer

Founded by AI4Bharat alumni Vivek Raghavan and Pratyush Kumar, Sarvam has raised roughly $54 million and has been discussed at a valuation near $1.5 billion (Startup Feed, June 2026). Its February 2026 release of the open-source Sarvam 30B and 105B models — plus speech and vision models tuned for Indian-language documents — makes it the closest thing India has to a national champion lab. Its core insight is economic: Hindi text can consume 3–4x more tokens than equivalent English, so India-tuned tokenisation directly cuts inference costs (Rest of World, 2026).

Krutrim and Neysa: the frugal challengers

Bhavish Aggarwal’s Krutrim, India’s first AI unicorn at a $1 billion valuation, trained its models on more than 2 trillion tokens across 22 Indian languages (Rest of World, 2026). Neysa, a sovereign AI cloud provider running more than 20,000 GPUs, raised $1.2 billion — the largest Indian AI round to date (Startup Feed, June 2026). Add the IT majors converting services revenue into AI revenue, and the honest picture is this: India’s AI companies are strongest in infrastructure, language adaptation and deployment — not frontier research.

The data-centre land grab is the real money story

Follow the concrete. India’s operational data-centre capacity stood at roughly 1.5 GW in 2025, with 4–5 GW planned by 2030 (Ellenox AI infrastructure report, June 2026). The commitments stacked up fast in 2026. Reliance announced ₹10 trillion (~$110 billion) over seven years, anchored by a multi-gigawatt Jamnagar facility with 120+ MW expected online in H2 2026 — Mukesh Ambani framing it bluntly: “The biggest constraint in AI today is not talent or imagination. It is scarcity and high cost of compute” (TechCrunch, February 2026).

Google broke ground in April 2026 on a $15 billion, nearly 1 GW AI hub in Visakhapatnam, built with AdaniConneX and Airtel’s Nxtra and fed by three new subsea cables (Google Cloud press release, April 2026). Microsoft committed $17.5 billion over 2026–2029, including a new Hyderabad hyperscale region and a pledge to train 20 million Indians by 2030 (Microsoft, December 2025). Adani Group separately flagged $100 billion for AI data centres by 2035 (Yahoo Finance, February 2026). Cheap land, surplus renewables and 100 million users to serve locally: the hyperscalers have concluded India is where the next tranche of global AI capacity gets built.

Talent is India’s real AI export

Strip out the announcements and India’s most bankable AI asset is people. The country’s AI/ML workforce reached roughly 2.75 million professionals in 2025–26, up 55% year on year (Ellenox, June 2026), and Stanford’s AI Index 2026 ranks India as home to the world’s second-largest AI talent pool — while also leading the world in AI brain drain (The Print, 2026).

The clearest expression of talent-as-export is the global capability centre. India hosts 2,117 GCCs generating $98.4 billion in revenue and employing 2.36 million professionals — including roughly 250,000 AI specialists — with more than 1,200 GCCs now running embedded AI/ML work and 250+ dedicated AI centres of excellence (Nasscom–Zinnov, July 2026). In effect, Fortune 500 AI roadmaps are being executed from Bengaluru and Hyderabad. The sting in the tail: Nasscom reports Indian tech headcount grew just 2.3% in FY26 even as revenue rose 6.1% — AI-driven efficiency is decoupling growth from hiring in the very industry that employs India’s AI talent (Nasscom, February 2026).

India AI Market 2026: FAQ

How big is India’s AI market in 2026?

Domestic AI revenues are an estimated $10–12 billion in FY26 within India’s $315 billion technology sector (Nasscom, February 2026), with an earlier Nasscom–BCG projection of $17 billion by 2027. Announced AI investment commitments exceed $200 billion, so infrastructure spend dwarfs current revenue.

How many ChatGPT users does India have?

India has 100 million weekly active ChatGPT users as of February 2026, per Sam Altman — OpenAI’s second-largest market after the US, and its largest student user base globally (TechCrunch, February 2026).

What is the IndiaAI Mission?

A ₹10,372 crore (~$1.25 billion) government programme, approved in 2024, subsidising AI compute and sovereign model development. Over 38,000 GPUs were onboarded by Q1 2026 at ₹115–150 per GPU-hour, with a target of 100,000 GPUs by December 2026 (IndiaAI portal data, 2026).

Which are India’s leading AI startups?

Sarvam AI (open-source Sarvam 30B/105B models, ~$1.5 billion valuation discussions), Krutrim (India’s first AI unicorn at $1 billion), and Neysa ($1.2 billion raised for sovereign AI cloud) lead a field of 1,700+ AI startups (Startup Feed, citing Tracxn/Zinnov, June 2026).

Who is investing in Indian AI data centres?

Reliance (~$110 billion over seven years, including gigawatt-scale Jamnagar), Google ($15 billion Visakhapatnam AI hub), Microsoft ($17.5 billion, 2026–2029) and Adani ($100 billion by 2035) — against operational national capacity of roughly 1.5 GW in 2025, headed for 4–5 GW by 2030 (Ellenox, June 2026).