The variance arises from the different methods the two firms use to account for revenue earned via third-party cloud services. Anthropic counts sales secured via partners like Google Cloud and Amazon Web Services (AWS) toward its revenue figures, whereas OpenAI handles certain partner-driven sales differently.
Anthropic compensates its cloud partners with roughly 16% of every dollar generated via these pathways. A Reuters analysis revealed that these particular sales made up half of the company’s revenue over the past year. Consequently, these contrasting accounting practices complicate straightforward comparisons unless the data is adjusted to a uniform standard.
While OpenAI declined to comment when contacted by Reuters, the *Financial Times* initially disclosed the updated projection, drawing on financial data provided to investors.
OpenAI expands as competition intensifies
OpenAI kicked off 2026 with an annualized revenue run rate sitting at USD 20 billion, up from USD 6 billion in 2024. Even with this expansion, the firm encountered heightened competition from Anthropic over the course of the second quarter.
OpenAI reported, “Quarterly revenue of USD 6.7 billion in the period, while Anthropic generated USD 11.5 billion. Anthropic’s annualised revenue run rate crossed USD 65 billion in July, and sources previously told Reuters that it could reach USD 100 billion by the end of 2026.”
These metrics highlight a fiercely competitive race for enterprise clients as corporations increasingly integrate generative AI tools into their operations. Both organizations are striving to scale their commercial ventures while balancing the heavy financial burdens tied to computing infrastructure and developing AI models.
Investors seek clarity ahead of potential listings
This adjustment to the estimates arrives while both OpenAI and Anthropic ready themselves for possible initial public offerings. Going public could grant investors clearer insights into their respective financial health, operating expenses, and revenue accounting practices.
Rapidly expanding tech enterprises frequently rely on the annualized revenue run rate metric, which multiplies a single month’s sales by 12 to project yearly performance. Nonetheless, this metric falls short of actual full-year revenue and risks giving investors a distorted picture if monthly sales figures are volatile.
Also Read: OpenAI Safety Alert: 3 Fired Researchers Urge AI Monitoring
