OpenAI is weighing broad price cuts after business customers pushed back hard on what AI processing actually costs at scale.
Anonymous sources close to the company say Sam Altman's team is debating reductions across its subscription tiers, with particular attention to token costs - the unit enterprises buy to run AI workloads. The backdrop is a trend dubbed "tokenmaxxing": corporate teams burning through large token budgets to squeeze productivity from AI tools, then questioning whether the return justifies the spend. Altman has called high prices "a huge issue." No decision has been finalized, and Anthropic is reportedly running the same calculation.
The timing is where this gets complicated. OpenAI recently filed for an IPO - potential valuation near $1 trillion, with a September listing floated - which is an unusual moment to voluntarily compress revenue. That it is considering doing so anyway suggests customer churn risk is real enough to override that instinct. Anthropic, which also recently moved toward public status, appears to be in the same bind.
Investor enthusiasm for AI is already softening, with major infrastructure players seeing their valuations cool. A simultaneous price war and IPO sprint is not the clean growth story public markets tend to reward.