OpenAI Told Investors Revenue Is Near $50 Billion, Not $70 Billion

OpenAI is said to have told investors that its annualized revenue is approaching $50 billion, about $20 billion below a roughly $70 billion figure that was widely reported in September, according to reporting published on 8 October 2026. TechCrunch, citing the Financial Times, said the higher number came from attempts by OpenAI's own investors to produce a direct comparison with Anthropic's annualised revenues. Axios reported the same $50 billion figure and described the earlier $70 billion number as a gross-up meant to mirror Anthropic's method of counting cloud-partner sales. Bloomberg Law likewise said the run rate was an uptick from the prior year but less than some recently reported estimates, and said the Financial Times earlier reported the new estimate.

What the $50 billion figure does and does not mean

Annualized revenue, or ARR-style run rate, multiplies recent sales into a twelve-month pace; it is not the same as audited yearly revenue. TechCrunch noted that Anthropic counts sales made by its cloud partners in its annualised figure, while OpenAI does not, which is why investor-side adjustments produced a higher OpenAI number for comparison. Axios reported that the $70 billion ARR figure was an attempt to put the two labs on equal footing using Anthropic's methodology, and that the discrepancy relates to how the companies account for sales made through cloud partners. Bloomberg Law dated its account to 8 October 2026 at 6:22 PM UTC and attributed the new estimate to people familiar with the matter after the FT report.

The distinction matters for anyone who treated the September headlines as a head-to-head scoreboard. A methodology alignment exercise that grosses up partner-mediated revenue is a different fact pattern from a demand shock. None of the 8 October stories reviewed here claim OpenAI restated historical GAAP revenue; they describe how an annualized, investor-facing metric was constructed and how a lower figure was later reported.

Why investors care about the comparison with Anthropic

OpenAI has been raising large sums while still spending heavily relative to revenue. TechCrunch recalled a $122 billion March funding round and leaked 2025 financials that showed about $13 billion of revenue against much larger spending, and said an IPO once rumoured for 2026 has been pushed toward early 2027. In that context, a $70 billion run-rate headline was useful shorthand for bulls arguing that consumer and enterprise demand could fund the infrastructure buildout; a $50 billion figure on OpenAI's own accounting is still enormous, but it narrows the rhetorical gap with Anthropic less aggressively.

Axios emphasised that public-market investors usually prefer actual revenue over early-stage constructs, a reminder that ARR comparisons between private AI labs are negotiated artefacts as much as operating truth. PromptCrates has previously covered OpenAI's capital story, including reporting on large raises and long-term compute commitments and business-user momentum versus Anthropic. Those pieces are about financing and customer mix; Thursday's news is specifically about how a widely repeated run-rate was adjusted once the Anthropic-comparison overlay was stripped back.

What secondary outlets added on 8 October

Coverage clusters around the same core claim with small differences in surrounding colour. TechCrunch stressed that it had reached out to OpenAI for comment and that the company is under pressure to justify infrastructure investment. Axios focused on the accounting bridge to Anthropic. Bloomberg Law underlined that the $50 billion pace is still higher than the prior year even if it undershoots the inflated comparison. Axios also cited accounting professor Francine McKenna, who said Anthropic's method lets it record the full amount a customer pays through a cloud provider as top-line revenue and list the provider's cut as an expense, while OpenAI records only its share of certain partner sales; Axios added that both companies are GAAP compliant.

Readers should treat any further growth statistics cautiously unless OpenAI publishes them. The multi-sourced point for 8 October is the report that OpenAI's investor-facing annualized revenue is approaching $50 billion on its own methodology, versus a previously circulated comparison figure near $70 billion.

How to read AI lab run rates from here

The episode is a warning about scoreboard journalism in the foundation-model market. When two companies define annualised revenue differently, a single headline number can embed a silent translation layer. Buyers, reporters and secondary investors who need a clean comparison should ask whether cloud-partner gross-ups are included, which months are being annualized, and whether the figure is company-stated or investor-reconstructed. Until OpenAI is public and reporting audited segments, run-rate snapshots will keep arriving through FT, Axios and similar channels rather than through a standard 10-Q.

For competitive context, Anthropic's inclusion of partner sales is not inherently improper; it is simply a different choice. OpenAI's choice not to include those sales makes its $50 billion figure harder to line up beside Anthropic without footnotes. The executive turnover and governance churn PromptCrates has covered separately remains a different story line; it does not change the arithmetic of Thursday's revenue report, but it does explain why capital markets watch every OpenAI metric so closely. The news dated to 8 October 2026 is that the reported gap is about methodology and comparison design, not a sudden $20 billion collapse in customer spend. Finance teams comparing term sheets should also ask which month's sales are being annualized and whether enterprise and consumer segments are blended, because a single top-line run rate can hide very different growth rates underneath. Those questions do not change Thursday's core report, but they determine whether a $50 billion headline is usable in an internal model or only as a media shorthand.

• TechCrunch: OpenAI's revenue is reportedly $20 billion less than previously projected
• Axios: OpenAI annualized revenue $20 billion less than previously reported
• Bloomberg Law: OpenAI's revenue run rate nears $50 billion, less than reported
• PromptCrates: OpenAI raise and long-term compute commitments
• PromptCrates: OpenAI gains on Anthropic in business users
• PromptCrates: OpenAI executive exodus and Brockman day-to-day role