"In 2024 alone, private U.S. investment in artificial intelligence reached roughly $109 billion, according to Stanford’s AI Index Report 2025, and major firms are now committing hundreds of billions more to AI infrastructure and data centers in 2025. For comparison, researchers estimate that ending homelessness nationwide would cost on the order of $9–30 billion a year; clearing the public-transit repair backlog would require about $140 billion; and cancelling student debt for millions of Americans could cost anywhere from $300 billion (for a $10,000 per-borrower plan) to over $870 billion. These are not impossible sums; they are simply directed elsewhere. The problem is not that we invest in technology — it’s that we do so to avoid investing in one another. We could rebuild public transit, deliver universal healthcare, cancel student debt, end homelessness — all projects of collective possibility — but instead we feed the circus. We reward those who promise salvation through algorithms while punishing those who simply demand dignity through policy. And if AI spend is today’s growth engine, it is also today’s concentration risk — a stimulus routed through a handful of firms, data centers, and supply chains.

The defenders of this frenzy will tell us to look forward — to think of the productivity gains that will surely justify today’s excess. It is, they insist, short-sighted to critique the future before it arrives. But history is not a ledger that balances itself. Productivity gains are not moral gains; they are distributed, like everything else, according to power. The Industrial Revolution doubled output but also deepened inequality. The internet democratized information yet concentrated ownership. Efficiency rose; bargaining power fell."

AI’s Grand Circus

We inflate markets, deflate possibility, and call it progress.

(democracyatwork.substack.com)

#AI #GenerativeAI #Capitalism #AIBubble #AIHype