Nadella’s Courtroom Disclosures Put $100B Microsoft–OpenAI Bet Under the Microscope

Tech history rhymes: in 1980 IBM licensed MS-DOS believing hardware would stay king; four decades later Nadella worries Microsoft could play IBM’s part if it only rents servers to OpenAI. The courtroom disclosures signal two deeper currents. First, cutting-edge AI now demands capital on oil-boom scales—far beyond nonprofit fundraising—pushing labs to hybrid or for-profit status much like Bell Labs spun into AT&T’s profit engine after the 1925 reorganization. Second, compute and model ownership are coalescing into a few megaplatforms; loosening exclusivity while locking in $250 billion of Azure consumption hints at a future where cloud oligopolies trade access for equity rather than outright control. On a 100-year arc, this moment could mark the shift from open scientific stewardship of AI to financialized infrastructure battles, with governance fought in courts rather than universities—reminiscent of the rail barons’ consolidation in the 1870s. Whether society benefits or merely witnesses another cycle of capital concentration will depend on how successor agreements balance profit motives against broad access as AI approaches general-purpose capability.