The AI Gold Rush Is Cover for a Class War (jacobin.com)

U.S. tech layoffs — white‑collar employment is down 1.9% since 2022 and layoffs are up 36% year‑over‑year — have been widely framed as the disruptive fallout of generative AI. CEOs warn of large-scale job displacement, but economists and studies (Yale Budget Lab, Daron Acemoğlu) find little evidence so far of sweeping productivity losses or massive short‑term unemployment; Acemoğlu estimates only ~1.5% GDP gain from AI over the next decade. Alternative explanations point to macro‑financial dynamics: the end of ZIRP left overleveraged firms exposed to higher rates (Oracle carries ~$95B in long‑term debt), while other firms raise huge sums to fund AI (Goldman: $141B new AI‑related debt in 2025; Bloomberg: $157B in tech bond issuance), even as total projected AI spending nears $1.5T this year. But the story is more political and structural than purely technological. Cash‑rich hyperscalers (Meta, Microsoft, Amazon) are funding massive infrastructure — Meta invested $14.3B and plans two data centers delivering 6 GW, using a $26B debt + $3B equity SPV — creating a closed loop where a few incumbents fund, sell to, and borrow from one another (Oracle, Nvidia, CoreWeave, SoftBank trade ~$1T in AI deals). The result is “enclosure”: returns accrue to platform owners, not productivity, and layoffs become a strategic lever to reduce labor bargaining power, intensify workloads, and normalize precarity. The piece argues that unions and political intervention — not only reskilling — are required to challenge this redistribution of power and reclaim democratic oversight over AI’s social effects.