OpenAI Seeks $30B at $1.25T Valuation as Anthropic IPO Looms

OpenAI is reportedly negotiating a $30 billion round at a lower valuation than discussed last month, while a rival eyes the public markets first. Separately, a widely reported PHP vulnerability put OpenAI’s models in the security spotlight.

Key points

• OpenAI is negotiating a $30 billion round at a $1.25 trillion pre-money valuation, per FinanceFeeds. Talks in September referenced $1.4 trillion.
• A second FinanceFeeds report says Anthropic targets a mid-November IPO at about $1.5 trillion.
• OpenAI announced that frontier models must be trained and evaluated offline after a September security breach, according to The Stock Market Watch.
• A critical PHP flaw was found with help from OpenAI and Google models, per The Conversation via Jamaica Observer.

What does the lower valuation signal?

According to FinanceFeeds, the $1.25 trillion pre-money figure sits below the $1.4 trillion discussed in September. That is a gap of $150 billion, or roughly 11%.

The sources do not explain the reduction. The summary frames it as a question of what the lower price signals, but the available text does not establish a cause. The second report describes the range as $1.25 trillion to $1.4 trillion and says the round could be up to $30 billion. So the terms appear unsettled, and the two reports differ slightly in how they present them. Negotiations are ongoing, and nothing here confirms a close.

How does Anthropic’s IPO plan change the picture?

FinanceFeeds reports Anthropic is aiming for a mid-November listing at about $1.5 trillion. It also cites Q3 revenue of $16.8 billion, up 45% on Q2. The summary does not make clear whether that revenue figure belongs to Anthropic or OpenAI, so treat it with caution until the full details are verified.

If the Anthropic figures hold, a rival could reach public markets before OpenAI and at a higher stated valuation than OpenAI’s current private talks. That would give investors a public pricing reference for the sector. These are reported targets, not completed events, and timelines for IPOs often move.

What did OpenAI change after the September breach?

The Stock Market Watch reports that OpenAI announced on Friday a shift in development protocols, requiring all frontier models to be trained and evaluated in offline environments after a September security breach. The available summary is cut off, and we could not see further details on the breach itself, its scope, or the cost of the new requirement. Readers should look for primary confirmation from OpenAI.

Why is a PHP vulnerability tied to OpenAI?

Per the Conversation article, researchers at Wiz used the latest Gemini and ChatGPT models in August to search PHP’s codebase for chainable weaknesses. The article says PHP underpins more than 70% of websites. Developers issued emergency patches on October 21 after the issue leaked on Hacker News. Several governments and cyber agencies urged urgent updates, and some operators, including the Port of Rotterdam, took sites offline.

The Economic Times describes the discovery, made with a powerful OpenAI model, as one of the most high-profile demonstrations of AI’s cyber capabilities. For OpenAI, that cuts two ways: it shows capability, and it raises questions about dual-use risk that may relate to the offline-training policy. The sources do not link the two, so that connection is our inference only.

This site is independent and not affiliated with OpenAI. Nothing here is investment advice.