Amazon's India AI Bet Reaches $48B Total

Amazon CEO Andy Jassy traveled to New Delhi this week to announce an additional $13 billion investment in India's AI infrastructure through 2030, bringing the company's cumulative commitments in the country to $48 billion. The funds go toward expanding AWS data centers in Mumbai and Hyderabad, and the timing is deliberate: global tech giants are racing to lock in computing capacity across Asia-Pacific before the best sites and power contracts are claimed.

The rounds stack up fast. Amazon committed $12.7 billion to AWS India in 2023, added $35 billion more in December 2025, and now tacks on another $13 billion just six months later. The competitors are moving just as aggressively: Microsoft has pledged $17.5 billion in India by 2029, Google has committed $15 billion for data center and AI hub expansion, and Indian conglomerate Reliance Industries has laid out a $110 billion AI investment plan of its own. Adani Group has announced $100 billion. Across the board, India is absorbing a surge of capital that dwarfs most nations' entire digital infrastructure budgets.

Why India became the new front in the AI infrastructure race

Three things make India particularly attractive right now. First, land: the country has space for the kind of gigawatt-scale campuses that AI training and inference require. Second, energy: India's power grid is expanding fast, and the government has been aggressively adding renewable capacity. Third, policy: India offers tax exemptions to foreign cloud providers on overseas services, with some arrangements extending through 2047. For AWS, Azure and Google Cloud, locking in data center locations now is a bet on where AI demand will be concentrated a decade from now — and India, with 1.4 billion people increasingly online, is a strong candidate.

Geopolitics adds another layer. Restrictions on US technology companies in certain markets have made some regions complicated for aggressive AI buildouts. India, with its close technology partnerships with US firms and a government actively courting foreign cloud investment, offers a different calculus. Jassy's meeting with Prime Minister Narendra Modi in New Delhi was as much a diplomatic signal as a business announcement.

What this means if you build with AI

Infrastructure investment at this scale has a direct and practical downstream effect: it makes AI cheaper and faster to use. More AWS data centers in South Asia means lower latency for API calls coming from that region — your app responds faster, costs less to run. And the competition itself matters. When Amazon, Microsoft and Google are each spending tens of billions to win the same markets, the pricing pressure on AI services stays fierce. That global infrastructure arms race is a big part of why the cost per token for frontier models has dropped so dramatically over the past two years, and why it is likely to keep falling.