Meta Weighs 20 % Workforce Cut to Fund $600 B AI Build-Out

Tech giants periodically pivot by trading payroll for capital expenditure: IBM’s 1993–95 shift toward services saw 35,000 cuts while pouring billions into mainframes; General Motors’ 1980s robotisation cost 40,000 jobs even as it spent $40 b on flexible factories. Meta’s mooted 2026 downsizing echoes that pattern—jettisoning human overhead to finance a new production means, this time clusters of AI accelerators instead of stamping presses. It reflects a longer arc in which digital firms, inflated by pandemic-era advertising booms, are now unwinding ‘just-in-case’ hiring as generative AI promises higher output per coder. Whether the $600 b figure is bravado or real, the gesture signals a structural move: from labour-heavy social media moderation toward capital-intensive computation, mirroring the century-long trend of automation reallocating productivity gains from wages to infrastructure owners. On a 100-year horizon, such consolidations may mark the early labour displacement phase of large-scale AI, comparable to mechanisation waves of 1810–1910; the long-term question is not this round of 16,000 jobs, but whether repeated cycles will entrench a few data-rich platforms as the railroads of the 21st century or prompt antitrust and social push-back that reins in their capital-first logic.