Microsoft’s $17.5 B Sovereign AI Cloud Gamble in India (2026-29)

Global tech firms have courted India before—IBM’s 1969 computer leasing and GE’s 1989 engineering centre—but few bet this magnitude upfront. Like the U.S. transcontinental railroad subsidies of the 1860s or Japan’s METI-backed semiconductor push of the 1970s, the deal weds state ambition (Digital India, data-localisation rules) with private capital seeking new growth as China becomes politically riskier. It reflects two structural currents: the ‘cloud nationalism’ wave, where nations demand in-country data control, and the scramble for AI-grade electricity, talent and users amid a plateauing Western consumer base. Whether this moment matters a century hence hinges on execution: if India converts cheap spectrum, a young workforce and regulatory heft into indigenous AI platforms—avoiding the enclave-style foreign factories of colonial 1900s—it could shift the centre of gravity of computation southwards; if not, the spend may echo the 1990s dot-com data centres that became stranded assets after the bubble. Either way, the scale signals that the geopolitical contest over where the world’s models ‘live’ has decisively expanded beyond the U.S.–China axis.