By Ramachandran Rajeev Kumar — 2026-01-21

By Ramachandran Rajeev Kumar

Google is investing $15 billion in Visakhapatnam to build a gigawatt-class AI data center. It's one of the largest foreign direct investments in Indian technology infrastructure—ever. The announcement has been met with the predictable chorus: India is becoming a global AI hub; foreign tech giants believe in India's future; this validates our digital economy.

All true. And all beside the point.

Before we celebrate too enthusiastically, let's ask two questions that rarely get asked when FDI announcements dominate headlines.

First: Why is Google doing this?

Second: Is this what India should actually be striving for?

The answers are less flattering than the press releases suggest.

Why Google Is Really Here

Google isn't investing $15 billion in Vizag because of India's infrastructure (which is adequate but not exceptional), its power grid (which is strained), or its regulatory environment (which remains Byzantine). Google is investing because it needs to hedge against China.

The writing is on the wall. US-China technology decoupling is accelerating. American AI companies face growing restrictions on operating in China—and the probability of being expelled entirely is no longer theoretical. TikTok's fate in America mirrors what could happen to Google, Microsoft, and Meta in China.

When that door closes—and it will close—Google needs AI infrastructure positioned to serve Asian markets. India, with 1.4 billion people and a growing digital economy, is the obvious alternative. Vizag isn't charity. It's insurance.

This context doesn't make the investment bad for India. Foreign capital is foreign capital. Jobs will be created. Infrastructure will be built. Indian engineers will gain experience operating world-class AI facilities. All good things.

But let's not pretend Google is doing India a favor. Google is doing Google a favor. India happens to be useful.

What India Actually Gets

A data center, even a gigawatt-scale one, is not an AI industry. It's a building full of servers.

Data centers create relatively few direct jobs—perhaps a few hundred operational staff for a facility this size. The high-value work—designing AI models, developing applications, capturing the economic value of artificial intelligence—happens elsewhere. Google's AI researchers sit in Mountain View, London, and Zurich. Vizag will host the compute. The brains remain abroad.

This is the fundamental limitation of FDI in infrastructure: it creates presence without capability. India becomes a location where AI runs, not a place where AI is invented.

Compare this to what India actually needs: indigenous AI companies that design models, build applications, and capture global markets. An Indian OpenAI. An Indian Anthropic. An Indian Nvidia that designs AI accelerators manufactured by foundry partners worldwide.

We have the engineering talent—arguably the world's deepest bench of AI/ML engineers. We have the domestic market—1.4 billion potential users of AI applications. We have the cost advantages that should make India the natural home for AI development.

What we don't have is the ecosystem that transforms these inputs into Indian companies.

The Bureaucracy Tax

Why doesn't India produce Googles and Nvidias?

The question answers itself for anyone who has tried to start a company in India, raise capital for it, hire employees, acquire land, obtain permits, or navigate the tax system.

India's penchant for bureaucratic red tape—the forms in triplicate, the inspectors with discretionary power, the regulations that change without warning, the corruption that greases every wheel—has left us struggling behind countries with fewer resources, smaller talent pools, and less obvious advantages.

Vietnam is building a semiconductor ecosystem. Indonesia is attracting electric vehicle manufacturing. Bangladesh has become a garment powerhouse. These are not countries with India's scale, India's engineering colleges, or India's diaspora networks. They are countries where it is simply easier to do business.

The World Bank's Ease of Doing Business rankings, before they were discontinued, placed India at 63rd globally—an improvement from 142nd a decade earlier, but still behind Rwanda, Kenya, and Kazakhstan. The improvement reflected genuine reform. The remaining gap reflects genuine dysfunction.

Every Indian entrepreneur has a story: the license that took eighteen months, the tax notice that arrived years after the fact, the inspector who demanded payment for approvals that should have been automatic. These frictions don't appear in FDI announcements. They appear in the startup graveyards—the companies that never scaled, the founders who gave up, the engineers who took jobs at Google instead of building competitors to Google.

The Real Benchmark

India's benchmark should not be "How much FDI did we attract?"

It should be: "How many Indian companies are capturing global markets in strategic technologies?"

By that measure, we are failing.

In AI, India has promising startups but no global champions. In semiconductors, we have ATMP facilities but no fabless design companies competing with MediaTek or Qualcomm. In cloud computing, we have data centers operated by Amazon, Microsoft, and Google, but no Indian hyperscaler. In electric vehicles, we have manufacturers assembling foreign designs, not companies exporting Indian-designed EVs to the world.

The pattern is consistent: India hosts. India does not build.

This isn't inevitable. It's a choice—or rather, an accumulation of policy choices that favor foreign incumbents over domestic challengers.

When Google builds a data center, it navigates bureaucracy with armies of lawyers and lobbyists. When an Indian startup tries to build a competing cloud service, it faces the same bureaucracy with a fraction of the resources. The playing field isn't level. It's tilted toward those who already have scale.

What "Success" Would Actually Look Like

Imagine an alternative 2026 headline: "Indian AI Startup Raises $5 Billion to Build Domestic Cloud Infrastructure."

Imagine an Indian company—founded in Bangalore, funded by Indian capital, employing Indian engineers—building AI data centers across the country. Imagine that company competing with Google, Amazon, and Microsoft for Indian enterprise customers, then expanding to Southeast Asia, Africa, and the Middle East.

This is not fantasy. This is what China did with Alibaba Cloud, what Korea did with Samsung, what Taiwan did with TSMC. These countries didn't just attract foreign investment. They built national champions that compete globally.

India's policy should be oriented toward this outcome—not because FDI is bad, but because FDI alone is insufficient. Google's $15 billion is welcome. But the goal should be an Indian company capable of investing $15 billion.

The Policy Shift Required

Creating conditions for indigenous Googles and Nvidias requires more than slogans. It requires:

Regulatory simplicity: A startup should be able to incorporate, hire, and operate without touching a government office. Digital-first, approval-free, inspection-minimal. If Estonia can do it, India can do it.

Capital access: Pension funds, insurance companies, and sovereign wealth should be investing in Indian technology ventures—not just safe government bonds. The capital exists. The permissions don't.

Government as customer: When the Indian government builds digital infrastructure, it should preference Indian providers—not as protectionism, but as anchor demand that allows Indian companies to achieve scale. Every major technology power does this. India outsources to foreign vendors and wonders why domestic companies can't compete.

Corruption elimination: Not reduction. Elimination. Every bribe demanded is a tax on Indian entrepreneurship that foreign multinationals can afford and Indian startups cannot. Anti-corruption isn't a moral issue. It's an industrial policy issue.

Patience: Indigenous technology champions take decades to build. TSMC was founded in 1987 and became dominant in the 2010s. Indian policy cycles in five-year election terms. Building Nvidias requires thirty-year horizons.

The Bottom Line

Google's $15 billion Vizag investment is good news. Take it. Build the data center. Employ the engineers. Capture whatever spillover benefits materialize.

But don't mistake hosting for building. Don't confuse a data center with an AI industry. Don't celebrate FDI as if it were the goal rather than a means.

Google is in India because China is closing and India is useful. When India is no longer useful—when labor costs rise, when regulations tighten, when some other country offers better terms—Google will optimize its portfolio accordingly. That's not cynicism. That's how multinational corporations operate.

The only durable strategy is building Indian companies that don't need to ask permission from Mountain View. The only real victory is an Indian Google, not a Google in India.

We're celebrating the latter. We should be building the former.