India is entering the next phase of its AI buildout. The government confirmed this week that the IndiaAI Mission now backs 15 indigenous foundation models, with compute available at ₹65 an hour to startups and researchers. Meanwhile, the Centre is moving to take a stake in Sarvam AI — a decision that could reshape hiring patterns across India’s entire IT services industry. Here’s what’s actually happening on the ground.
🏛️ 1. Centre to Take a Stake in Sarvam AI — and It Could Reshape India’s IT Hiring Pyramid
According to a Business Standard report published today, the Centre plans to acquire a small equity stake in Sarvam AI as part of its continued backing under the IndiaAI Mission’s Foundation Model pillar. The move, first reported by The Economic Times last week, is intended to strengthen India’s domestic AI capabilities — but its second-order effects may be just as significant.
A new report from staffing firm Quess Corp suggests Indian IT firms are increasingly hiring for AI governance, data residency, and sovereign cloud architecture roles — not just model-building skills. As sovereign AI investment scales, the demand is shifting from “people who can train models” to “people who can deploy, govern, and secure them in production.” That’s a meaningfully different talent pyramid than the one India’s IT sector has built over the last two decades.
📌 Why it matters: If you’re hiring or upskilling for AI roles in India right now, governance and deployment expertise — not just model training — is where demand is heading. Worth factoring into recruitment and L&D planning over the next two quarters.
⚡ 2. 15 Indigenous Foundation Models Now Backed Under the Mission
A government update published today as part of the broader Digital India 2.0 push confirmed that the IndiaAI Mission, with an outlay exceeding ₹10,000 crore, is now supporting 15 indigenous foundation models — up from the 12 originally selected across the first two phases. High-performance compute remains available to startups and researchers at ₹65 per hour, among the cheapest subsidised AI compute rates anywhere in the world.
The announcement was framed as part of India’s broader digital transformation story: Digital India has now completed 11 years, with over 102.86 crore internet connections and UPI processing 24,162 crore transactions annually, valued at ₹314 lakh crore. AI, semiconductors, and quantum technologies are positioned as the next growth engines — with 12 semiconductor projects worth ₹1.65 lakh crore already approved alongside the AI push.
📌 Why it matters: Three more foundation models joining the cohort means more sovereign options on the table for India-specific applications. Worth tracking which new entrants get named in the coming weeks.
🔌 3. The Honest Gap: Compute Targets Are Outrunning India’s Power Infrastructure
A detailed sovereign AI status review notes a structural tension that doesn’t make it into most headlines: GPU integration demands 7–8 times higher power density than traditional server racks, and most existing Indian data centres cannot handle that load without major retrofits. The IndiaAI Mission’s compute targets and India’s power infrastructure are, in the report’s words, “not currently on compatible trajectories.”
There’s also the chip dependency question. India’s entire compute base runs on US-designed and manufactured hardware — the same category of hardware the US Commerce Department has shown willingness to restrict elsewhere. The February 2026 US-India interim trade framework includes language protecting India’s access to advanced AI chips, but as the report points out, that protection exists inside a trade agreement, not a permanent guarantee.
On the funding side, a Rajya Sabha reply from MeitY confirmed that only about ₹400 crore of the ₹10,371.92 crore outlay has actually been released over two years — ₹21.79 crore in 2024–25 and ₹379.15 crore in 2025–26, with nothing yet released for 2026–27. Meanwhile, Amazon and Microsoft have each committed over $50 billion to India’s cloud and AI infrastructure — far exceeding public spending, and raising real questions about who ends up controlling India’s compute pipeline.
📌 Why it matters: The infrastructure story is genuinely impressive, but the power-grid and fund-release gaps are real constraints — not just commentary. Worth keeping in mind before treating GPU growth numbers as the whole picture.
📋 IndiaAI Mission at a Glance — June 30, 2026
| Metric | Current Status | Priority |
|---|---|---|
| Indigenous Foundation Models | 15 backed under the mission | 🔴 Critical |
| Sarvam AI Stake | Government finalising equity stake | 🔴 Critical |
| Compute Access | ₹65/hour subsidised, 38,000+ GPUs deployed | 🟡 High |
| Fund Releases | Only ~₹400 crore released of ₹10,371.92 crore outlay | 🟡 High |
| Power Infrastructure | Misaligned with GPU density requirements | 🟠 Medium |
| Startups Global Cohort | Applications closed June 28, 2026 | 🟠 Medium |
✍️ Editor’s Take
There’s a useful word for where the IndiaAI Mission stands today — nascent (अंकुरित / ankurit, meaning newly sprouted or just beginning to take shape). The compute numbers, the model count, the equity stake in Sarvam — all of it signals real intent and real progress. But the power-grid mismatch and the ₹400-crore-of-₹10,371-crore fund release gap are reminders that infrastructure ambition and infrastructure delivery are two different timelines.
The talent shift toward AI governance and sovereign cloud architecture roles is arguably the most underrated story in today’s update. It suggests Indian IT services firms are reading the sovereign AI moment correctly — not as a model-building race they need to win outright, but as a deployment and governance opportunity they’re well positioned to own. That’s a more realistic, and more durable, strategy than chasing frontier model parity.
Q: Is the government taking a stake in Sarvam AI?
A: Yes. According to reports from The Economic Times and Business Standard, the Centre plans to acquire a small equity stake in Sarvam AI as part of its continued support under the IndiaAI Mission’s Foundation Model pillar.
Q: How many indigenous foundation models does the IndiaAI Mission now support?
A: As of late June 2026, the mission is backing 15 indigenous foundation models, up from the original 12 selected across the first two funding phases.
Q: How much of the IndiaAI Mission’s budget has actually been released?
A: Per a Rajya Sabha reply from MeitY, only about ₹400 crore of the total ₹10,371.92 crore outlay has been released over two years — ₹21.79 crore in 2024–25 and ₹379.15 crore in 2025–26, with no funds yet released for 2026–27.
Q: What is India’s power infrastructure challenge for AI?
A: GPU clusters require 7–8 times higher power density than traditional server racks. Most existing Indian data centres cannot handle this load without significant retrofits, creating a mismatch between compute ambitions and grid readiness.
Q: What is the current subsidised compute rate under the IndiaAI Mission?
A: Startups, researchers, and academic institutions can access high-performance AI compute at ₹65 per hour, among the most affordable subsidised AI compute rates globally.
Follow our AI4Planet Weekly News and AI News pages for updates every week. Have a story tip? Drop it in the comments below — we read every one.
Tags: IndiaAI Mission June 30 2026, Sarvam AI government stake, IndiaAI 15 foundation models, IndiaAI compute power infrastructure, IndiaAI fund release Rajya Sabha, India sovereign AI talent
