Overview
- Reports published June 11 say OpenAI is actively debating significant reductions to the price it charges per token, the metered unit firms use to bill API and enterprise AI usage.
- The deliberations focus on token pricing for business customers rather than flat consumer subscriptions and remain unfinished with no final decision announced.
- Both OpenAI and Anthropic have filed confidentially for IPOs, a timing pressure that makes short‑term market share and customer metrics especially important to their public narratives.
- Companies are already rationing or capping AI use because heavy token consumption has driven bills far higher than expected, which is the primary reason executives want lower unit prices.
- Large compute costs and exclusive capacity deals that underwrite model delivery mean any reciprocal price cuts could sharply compress margins, reshape valuations and draw intense investor scrutiny.