Enterprise AI spending stalls while usage climbs, challenging growth assumptions

The Ramp dataset introduces a supply-side pressure the summary gestures at but understates: customers are not just spending less, they are actively arbitraging between vendors, which means pricing power is eroding across the category, not just at one company.

This lands at the worst possible moment for Anthropic. The Decoder reported on November 6 that Anthropic's $1.5 trillion IPO is already facing investor resistance over valuation, and that story preceded any public accounting of softening enterprise spend. Layer in the earlier TechCrunch piece from October 30 on cracks in the AI spending boom, and the Ramp data starts to look less like a one-quarter blip and more like confirmation of a structural deceleration. Anthropic's own economics compound the problem: as covered in the November 3 TechCrunch piece on compute margins, inference pricing pressure is already outpacing efficiency gains. Ramp's evidence that buyers are extracting more value per dollar spent tightens that squeeze further, because it signals vendors cannot easily reprice upward to recover margin.

If Anthropic's IPO roadshow proceeds at or near the $1.5 trillion target within the next 60 days despite this spending data becoming public, it would suggest institutional buyers are discounting near-term revenue multiples in favor of long-run capability bets. A repricing or delay would confirm the valuation is load-bearing on a growth narrative the Ramp data has now weakened.