OpenAI revenue update sinks AI stocks as USA Today sues and math proofs draw criticism

CNBC said the roughly $50 billion figure was lower than the $68 billion widely reported late last month. A person familiar with the matter said the $68 billion figure included gross revenue from OpenAI's partners, which helped investors make a more direct comparison with Anthropic. The Financial Times reported that an earlier figure of about $70 billion was based on information OpenAI had previously shared with investors and on attempts by its investors to produce a direct comparison with Anthropic's annualized revenue. TechCrunch noted that OpenAI and Anthropic calculate annualized revenue differently, with Anthropic counting sales made by its cloud partners and OpenAI not doing so.

OpenAI also told investors in an investor presentation that its total run rate grew 77% in the third quarter, while its enterprise business grew 107% in the same period, the person said. The company is under pressure to justify its $852 billion valuation ahead of an expected IPO. It confidentially filed its prospectus with regulators in June, and executives have signaled they are eyeing a 2027 debut. OpenAI closed a $122 billion funding round in March, and CFO Sarah Friar told CNBC last week that the company remains "very well capitalized." OpenAI is also in early discussions with investors about a potential new funding round of around $30 billion, CNBC previously reported, though no term sheet has been finalized.

On the same day, USA Today Co. and several local newspapers it owns sued OpenAI, alleging the company copied "hundreds of thousands" of articles to train its AI models. In a filing on Thursday, the publisher asked for damages of more than $250 million and said OpenAI's unauthorized use of its content "has done real and continuing" harm to its outlets. The Tennessean, Indy Star, The Columbus Dispatch, The Oklahoman and several other local outlets are also included in the lawsuit. OpenAI did not immediately respond to The Verge's request for comment.

The lawsuit is the latest in a string of copyright cases against OpenAI. The New York Times, The Intercept, CNET owner Ziff Davis, CBC/Radio-Canada, Encyclopaedia Britannica, Merriam-Webster, The Seattle Times and a coalition of nearly 400 local newspapers have also filed legal actions. USA Today's lawsuit claims OpenAI never asked for permission and instead "took" its content to "build products worth hundreds of billions of dollars."

OpenAI's release this week of hundreds of claimed solutions to hard math problems also drew criticism from mathematicians. OpenAI said it had consulted an advisory group of elite mathematicians to avoid controversy after a previous model solved a long-standing problem. But the Advisory Group on Mathematics and Artificial Intelligence, hosted by Princeton University's Institute for Advanced Studies, said in a statement that "it is ultimately up to the mathematical community to assess the extent to which our recommendations were followed successfully." The group's first request was "to stop testing advanced mathematical problems on proprietary models," while OpenAI's release explicitly says it is evaluating proprietary models using open research problems in mathematics.

TechCrunch reported that only ten of 719 manuscripts included releases of the model's chain of thought, and 42% of the released proofs had not undergone formalization. A paper by mathematicians at the University of Cambridge and King's College London documented at least two discrepancies between the natural-language proof and the Lean code behind OpenAI's solution to a problem derived from the Navier-Stokes equations. The paper concluded that OpenAI's natural-language proof and other autoformalized Lean proofs "should not prima facie be trusted" without the same peer review as other proofs. Terence Tao, a prominent mathematician who has criticized OpenAI's approach, wrote on social media that problems are being solved autonomously by AI prompters who do not understand the output well enough to answer questions or engage with the field. Harvard mathematics professor Melanie Wood told TechCrunch that when a model spits out a solution, "there is not human understanding of them at the point of release, and now the work begins."

Separately, Arena, the AI leaderboard that began in 2023 as a UC Berkeley research project, said Thursday it raised a $200 million Series B round at a $3.1 billion valuation. The round was led by Lightspeed Venture Partners and Khosla Ventures, with Salesforce Ventures, 01 Advisors, Dell Technologies Capital, Endeavor Catalyst, a16z, Felicis and others participating. Arena previously announced a $150 million Series A in January at a $1.7 billion post-money valuation, when its annualized revenue was $30 million. The company said it reached $100 million in annualized run-rate revenue in June. It provides a free crowdsourced platform with tens of millions of monthly visitors and introduced AI Evaluations, a commercial service for model labs and enterprises, in September last year.

Arena also added an alignment category to its leaderboard, ranking models on issues such as unauthorized action, false attribution and "deceptive completion." OpenAI models currently top its preliminary alignment leaderboard, with Claude Opus 5.5 and Claude Fable in sixth and ninth place, respectively.