A look at differing revenue calculations of Anthropic and OpenAI, as Anthropic books gross sales through cloud partners, while OpenAI records only its net share
When measuring the race between artificial intelligence leaders OpenAI and Anthropic PBC, investors have run into a problem …
Context & Ripple Effects
OpenAI and Anthropic already dominate the revenue pool among leading AI startups, according to a May analysis of startup revenue concentration. But their reported figures have been assessed alongside very different cost profiles: investor materials examined in April separated profitability calculations with and without training costs and put inference costs above half of revenue for both companies.
The comparison problem is therefore not just which company sells more AI services; it is whether cloud-partner sales are counted on a gross or net basis before investors use revenue to judge growth, margins, and valuation.
First-order effects
- Anthropic’s reported revenue is immediately larger than a net-revenue presentation of the same cloud-partner transaction, while OpenAI’s reported figure excludes the partner’s share.
- Investors can no longer treat Anthropic and OpenAI revenue estimates as like-for-like scorecards without adjusting for the two companies’ recognition methods.
Second-order effects
- Cloud partners become material to Anthropic’s headline growth measure, tying its reported scale more closely to partner-mediated distribution than OpenAI’s net reporting does.
- Analysts assessing OpenAI’s reported first-quarter revenue and negative adjusted operating margin must separate revenue-recognition differences from underlying demand and cost performance.
Third-order effects
- As private AI leaders rely on cloud distribution and compute-intensive delivery, revenue definitions are becoming a central valuation input rather than a back-office accounting detail.
- If investors press for reconciled gross and net metrics, AI-company comparisons will shift toward unit economics and cash costs, reinforcing scrutiny already raised by profitability projections that separated training costs.
The trend: AI infrastructure commercialization is making partner economics and revenue recognition central to how investors compare model providers’ growth.