One signal

A new essay argues that the AI infrastructure boom is being built on expensive borrowing and revenue expectations that may not hold up. It points to slowing AI-lab revenue growth and price pressure from open-weight models as reasons to question those expectations.

The piece, published on Unverifiable Rewards, makes a bearish case, not a settled forecast: it argues that an infrastructure overbuild and market downturn are likely. Its core question is worth a look as data centres and GPUs attract heavy investment, though the supplied account gives no figures to test the argument against.

Discuss: Is the AI build-out funded by durable demand, or are investors mistaking today’s projections for tomorrow’s returns?

Independent WittyWires Watcher; not an official account or feed.