Microsoft is doubling down on artificial intelligence with one of its largest infrastructure commitments yet, unveiling $23 billion in new global AI investments, anchored by a $17.5 billion expansion plan in India.

CEO Satya Nadella said the four-year spending push, starting in 2026, would be Microsoft’s largest investment in Asia and would turn India into the company’s largest cloud footprint outside the United States. The move builds on a $3 billion commitment announced earlier this year and reflects Microsoft’s growing focus on markets where demand for cloud services and AI capacity continues to surge.

India has become a strategic priority for U.S. tech giants, backed by nearly one billion internet users, a deep pool of software talent, and rising enterprise demand for AI services. Large-scale data centers are a central part of that play, especially as India looks to secure a foothold in the AI economy despite limited domestic chip manufacturing.

Microsoft’s India plans sit alongside an aggressive global spending cycle. In January, the company said it expects to invest roughly $80 billion in fiscal year 2025, primarily to build and upgrade data centers for AI model training and cloud services such as Copilot.

The company has followed a similar playbook across Europe. In February, Microsoft announced a €3.2 billion investment in Germany over the next two years, its largest commitment to the country in four decades, with AI infrastructure and developer capacity at the center. Over the past year, Microsoft has committed at least $6 billion across multiple countries, including Poland, alongside large-scale AI training efforts. The company has pledged to upskill two million people in India’s smaller cities and one million in the UK.

“Thank you, PM @narendramodi ji, for an inspiring conversation on India’s AI opportunity. To support the country’s ambitions, Microsoft is committing US$17.5B—our largest investment ever in Asia—to help build the infrastructure, skills, and sovereign capabilities needed for India’s AI first future. @PMOIndia,” Nadella said in a post on X, sharing a photo with Indian Prime Minister Narendra Modi.

Microsoft CEO CEO Satya Nadella and Indian Prime Minister Narendra Modi

Nadella is in India for a three-day visit tied to Microsoft’s AI conferences, with stops in New Delhi, Bengaluru, and Mumbai. The announcements land at a tense moment, as New Delhi and Washington remain at odds over tariffs and a stalled trade agreement, adding a geopolitical layer to the rush for AI capacity.

Microsoft’s expansion comes as rivals make similarly bold moves. In October, Google said it would invest $15 billion over five years to build an AI data center in Andhra Pradesh, its largest commitment in the country. Industry-wide, U.S. cloud providers are expected to spend more than $400 billion this year on data centers that power services such as ChatGPT, Copilot, and Gemini. That pace has raised concerns around lofty valuations and a shortage of clear productivity gains, fueling debate over whether the AI spending cycle is overheating.

In India, Microsoft said a new hyperscale data center region in Hyderabad will become its largest in the country, with operations planned for mid-2026. Existing regions in Chennai, Hyderabad, and Pune will be expanded, and the company has doubled a prior pledge to equip 20 million Indians with basic AI skills by 2030. Real estate consultancy Colliers expects India’s total data center capacity to more than triple to roughly 4.5 gigawatts by the end of the decade.

Beyond India, Microsoft outlined fresh capital plans for North America. The company said it will invest more than C$7.5 billion in Canada over the next two years, part of a broader C$19 billion commitment between 2023 and 2027. New cloud capacity tied to that effort is expected to come online in the second half of 2026. Microsoft is pairing that investment with partnerships, including a collaboration with Canadian AI startup Cohere to offer Cohere’s models on Azure, Reuters reported.

Microsoft now employs more than 22,000 people in India and about 5,300 in Canada. Recent commitments include $10 billion in Portugal and $15 billion in the United Arab Emirates, reinforcing the company’s view that AI infrastructure will shape economic growth over the next decade.

Taken together, the spending signals a clear bet: control the pipes that feed AI workloads, even as questions linger over how fast real returns will arrive.