'India wants data sovereignty and we're delivering it': Microsoft's Puneet Chandok on the $17.5 billion India commitment

India has made one thing clear about its AI future—it wants to own it. Own the data, own the infrastructure, own the models. At the AI Impact Summit 2026 in New Delhi this week—the first global AI summit hosted in the Global South, with 35,000-plus delegates from over 100 countries—that message is impossible to miss.

Twelve indigenous foundation models are being unveiled, the BharatGen multilingual AI platform supporting 22 Indian languages is going public, and a Rs 10,370 crore national AI mission is backing it all.And right in the middle of that push, Microsoft—one of the world's largest technology companies—is making its biggest bet in Asia: $17.5 billion in India over four years. The message from Redmond is simple: if India wants sovereign AI, Microsoft wants to be the one helping build the pipes.Speaking to TimesofIndia.com on the sidelines of the AI Impact Summit 2026 in New Delhi, Puneet Chandok, President of Microsoft India and South Asia, laid out what that means in practice—from the thinking behind the investment, to the physical infrastructure taking shape, to why he believes India has an edge that no other large economy can replicate.

Five signals. One year. A 6x bet.

Microsoft committed $3 billion to India in January 2025. By December, that figure had grown nearly six-fold. So what happened in between?Chandok points to five things coming together simultaneously.

"There's massive infrastructure build out. There's capital coming in. There's talent—we've got 22,000 engineers in India building for the world. There's a real policy framework that's supporting this. And then there's market demand," he said. "Not many countries in the world have those five things—infrastructure, capital, talent, policy and market demand—coming together.

"It's worth noting just how competitive the landscape has become.

Google announced a $15 billion plan to build AI and data centre infrastructure in India. Amazon continues to pour billions into AWS expansion. OpenAI and Anthropic have both set up offices in the country. India, seemingly overnight, has become the most contested AI market outside the US and China.Against that backdrop, Microsoft's decision to nearly 6x its commitment in under a year makes more sense. "We looked even further ahead and said, how do we invest more in this country to not just meet the demand but also build for the future?" Chandok said.And infrastructure alone isn't the play. "There's no point building infrastructure if you're not scaling the talent," he added, pointing to Microsoft's commitment to skill 20 million Indians in AI by 2030. Through its ADVANTA(I)GE India initiative, the company says it has already trained 5.6 million people since January 2025—well ahead of its original target.

Why data sovereignty is the centrepiece of Microsoft's India play

If there's one word that defines India's AI moment, it's sovereignty.

The government's push for indigenous models, domestic data processing, and homegrown AI solutions has been unmistakable—and Microsoft is positioning itself squarely within that framework.The company is launching Sovereign Public Cloud and Sovereign Private Cloud processing for Copilot in India, making the country one of only four markets globally with in-country data processing. "Indian customers and the Indian government want data sovereignty, and we are happy to deliver that," Chandok said.The implications are significant. In-country data processing means Indian enterprise and government data stays within Indian borders—processed locally, governed locally. For sectors like banking, healthcare, and defence—where data residency isn't optional—this is a foundational requirement. For a country that has made digital self-reliance a policy priority, with the Digital Personal Data Protection Act on the books and AI governance frameworks under development, this is exactly the kind of commitment India has been pushing global tech companies to make.

Data centres, talent, and 310 million workers

A large chunk of the investment is going into physical infrastructure, with Hyderabad at the centre. Microsoft's largest hyperscale data centre region in India—with three availability zones and a footprint Chandok describes as "roughly the size of two Eden Gardens stadiums combined"—is set to go live by mid-2026, giving the company its biggest data centre capacity in the country. Expansions are also underway at existing data centre regions in Chennai and Pune.But Chandok was careful to note this isn't just a hardware story. "A lot is also going back to our talent in the country. We're building AI for the world from India with our massive engineering presence. And then building solutions for India," he said.One example he pointed to is Microsoft's partnership with the Ministry of Labour and Employment, where AI is being integrated into e-Shram and the National Career Service—two platforms that serve over 310 million unorganised workers.

The AI-powered upgrades include multilingual access, AI-driven job matching, predictive skill analytics, and automated resume tools. "We're creating job opportunities for them through AI-powered matching, skilling, and career pathways," Chandok said.It's a telling illustration of what Microsoft is trying to demonstrate in India—that AI investment isn't just about enterprise software and cloud revenue.

It's about embedding into the country's public infrastructure in ways that directly touch citizens.

DPI was the foundation. AI PI is the next chapter

Ask Chandok what gives India a genuine advantage in the global AI race, and the answer is immediate—Digital Public Infrastructure.India's DPI stack—Aadhaar (1.3 billion biometric identities), UPI (processing billions of transactions monthly), DigiLocker, and a growing ecosystem of interoperable digital platforms—is something no other large democracy has replicated at scale.

It's the foundation that enabled India to deliver direct benefit transfers to hundreds of millions during the pandemic, and it's the reason the country's digital payments story became a global reference point.Now, Chandok argues, AI is the next layer."We have a massive advantage in India, which is DPI," he said. "We have the ability to take DPI, complement that with our entrepreneurial energy, complement that with the policy framework that the government is building, and almost transition from DPI to AI PI—AI public infrastructure."It's a striking framing. Chandok argues that AI in India should be built the way DPI was built—as public infrastructure. "It should be safe, open, trusted, scalable. Imagine embedding AI into every one of our digital public infrastructure platforms. It'll make it even more effective and allow us to diffuse it to every Indian."The early signs are already visible. Thirty-eight million smallholder farmers across India received AI-driven monsoon forecasts via SMS in 2025, with analyses suggesting a return of over $100 for every $1 spent on those forecasts.

BharatGen, India's first government-funded multimodal model, is designed to work across India's 22 official languages. And platforms like Bhashini—supporting 35-plus languages with over 1,600 AI models—are being deployed across railway stations and government services.If AI can be layered on top of India's DPI with the same philosophy—open, interoperable, population-scale—India isn't just adopting AI.

It's building a model no one else can copy.

Building AI from India, for the world

Beyond infrastructure and sovereignty, Chandok stressed that India's role in Microsoft's global AI story is fundamentally shifting. The country isn't just consuming AI tools—it's building them."India is building products for the world—from India to the world. The ideas coming out of India are massive," he said. Microsoft employs over 22,000 engineers across Bengaluru, Hyderabad, Pune, Gurugram, Noida, and other cities—teams that don't just localise global products but build core AI tools like Copilot Studio, Azure AI Search, AI agents, speech and translation services, and Azure Machine Learning from scratch.

A significant chunk of the AI products used by enterprises worldwide are being engineered out of Indian offices.India is also home to over 1,800 Global Capability Centres—including more than 500 focused on AI—making it one of the largest concentrations of AI engineering talent anywhere in the world. The country ranks second globally in public generative AI projects on GitHub and is home to 16 per cent of the world's AI talent.

The demand for AI professionals in India is projected to reach 1 million by 2026.For Microsoft, the investment in Indian talent isn't just about serving the Indian market—it's about making India a core engine of its global AI product line.

The $17.5 billion is in. What comes next

The $17.5 billion makes Microsoft's intent clear. But sovereign AI infrastructure is only as good as what it enables on the ground. India's challenge now isn't attracting investment—it's making sure that investment translates into something a small business in Indore or a district hospital in Odisha can actually use.The foundation is being laid—the data centres, the sovereign cloud, the DPI-AI integration, the skilling programmes. What gets built on top of it is the part that matters. And that story is just getting started.