Anthropic IPO valuation expectations slide toward $1-1.5T amid data-center backlash

CNBC reports that investors are pushing Anthropic's IPO valuation down from roughly $2 trillion toward a range of $1-1.5 trillion. The listing is expected this month. The main stated concern is political: after voters rejected data-center projects in 40 of 41 midterm ballot measures, the assumption of unlimited compute build-out looks riskier.

The repricing is also about unit economics. Hacker News threads with hundreds of comments focused on Anthropic's reported $895B in compute commitments. Commenters argued that spending obligations are growing faster than revenue, and one thread debated whether the SEC should block the IPO. Others pointed out that index funds would eventually be exposed to AI IPO risk. Prediction markets put the odds of the IPO happening by November 7 at 43%.

The timing lands in a week of cost discipline across the industry. Apple is cutting internal AI budgets, and Anthropic itself just shipped project cost tags for enterprise customers. The contrast with OpenAI is notable: OpenAI raised $11B at $1.25T in private markets this week, so a public Anthropic at a similar number would put the two labs at near parity.

The bull case remains capability. A METR analysis reported by the Washington Post found Anthropic's gains on well-specified tasks are arriving at roughly twice last year's pace. The question for public investors is whether that capability curve can outrun the political and financial cost of the compute behind it.