OpenAI Lowers Its Price to $1.25 Trillion as Investors Push Back

Weeks after pitching a fixed $1.4 trillion price, the ChatGPT maker is closing up to $30 billion at a lower mark, racing Anthropic’s IPO as AI stocks slide and revenue draws scrutiny.

Key Takeaways

• OpenAI cut its pre-money target to $1.25 trillion from $1.4 trillion after some investors balked, a price that still stands about 47% above the $852 billion valuation set in March.
• Annualized revenue near $50 billion, not the $70 billion widely cited, puts the new price at about 25 times current sales and leaves investors underwriting an ambitious forecast.
• OpenAI wants commitments before Anthropic files publicly, as a Nasdaq sell-off, Meta’s negative free cash flow and SoftBank’s halved share price harden attitudes toward AI capital.

The Price Comes Down

For most of its commercial life, OpenAI has set the terms on which others could own a piece of it. This week, the terms moved. The ChatGPT maker is close to finalizing a round of up to $30 billion at a $1.25 trillion pre-money valuation, led by investors from the United Arab Emirates and Coatue, people familiar with the deal told CNBC on Thursday. OpenAI has told prospective backers it wants signatures by the end of next week, though the final size remains open, and it is describing the money as an “opportunistic” bridge to a public listing. The more telling detail is where the number began. Earlier this month, Bloomberg reported that OpenAI was presenting $1.4 trillion as a fixed price before it had chosen a lead investor, reversing the customary sequence in which investors propose terms and compete for the right to set them.

According to the Financial Times, Morgan Stanley and Dragoneer took that figure to the bank’s private wealth clients, and some declined to commit at that level. One person close to the process told the newspaper that earlier OpenAI raises had produced a frenzy, while this one landed as “a bit of a thud.” OpenAI declined to comment. The concession deserves a careful reading. At $1.25 trillion, OpenAI would still be valued roughly 47% above the $852 billion post-money mark it set on March 31, when it closed $122 billion in committed capital anchored by Amazon, Nvidia and SoftBank, a round SoftBank’s own disclosures show was priced at $730 billion before new money. This is a lower asking price, not a down round. What has shifted is the balance of negotiating power, which at this scale of fundraising is no small matter.

Revenue, Restated

The pressure on price traces back to a disclosure earlier this month. On October 8, the FT reported that OpenAI had told investors its annualized revenue was approaching $50 billion at the end of September. The widely cited $70 billion was an investor construction: OpenAI’s numbers restated on Anthropic’s basis, which books sales made through cloud partners at their gross value, whereas OpenAI records only its own share. The business did not change overnight, yet a $20 billion gap in a headline metric is the kind of discrepancy that sends late-stage investors back to their models. The trajectory itself remains impressive. A person familiar with an investor presentation told CNBC that total run-rate revenue grew 77% in the third quarter and the enterprise run rate rose 107%, and Bloomberg reported that OpenAI expects to finish the year at no less than $70 billion on its own accounting.

That acceleration followed a softer second quarter. The Wall Street Journal reported that revenue rose 18% sequentially to $6.7 billion while the operating loss, including stock-based compensation, widened to $12.3 billion from $9.3 billion. In the same quarter, Anthropic’s revenue reached $11.6 billion, the first time it had outsold its older rival. Set against those figures, $1.25 trillion prices OpenAI at about 25 times its September run rate and about 18 times its year-end target, so new investors are underwriting the forecast rather than the business as it stands. The bull case rests on mix: OpenAI said in March that enterprise customers already accounted for more than 40% of revenue and were on track to match consumer by the end of 2026. The bear case rests on cost. The company told investors in February that it is targeting roughly $600 billion in total compute spend by 2030.

A Race Against a Prospectus

Timing explains much of the urgency. Reuters reported on October 19 that OpenAI was racing to secure commitments before Anthropic’s expected public debut in mid-November, meeting long-only and sovereign wealth funds in the same week its rival hosted an investor day. One person said OpenAI wanted to finish before Anthropic’s disclosures invite side-by-side comparison, and some investors read the effort as an attempt to lock up capital that might otherwise flow to Anthropic’s offering. The two companies are drawing on the same finite pool of institutional appetite. Anthropic arrives with momentum. It raised $65 billion at a $965 billion post-money valuation in May, has told investors its third-quarter revenue reached $16.8 billion, up 45% from the prior quarter, and could make its prospectus public as soon as this week.

Under SEC procedures, a company that submits a draft confidentially must file publicly at least 15 days before its roadshow, which places Anthropic’s figures before the market just as OpenAI’s round closes. A draft reviewed by the Wall Street Journal showed first-half revenue of $16.3 billion against a net loss of $80.7 billion, most of it a non-cash charge tied to financing instruments. Neither side is assured of a flattering comparison. Reuters reported that some investors now prefer an Anthropic IPO valuation nearer $1.5 trillion than the $2 trillion it had sought, while OpenAI has questioned Anthropic’s practice of booking cloud-partner sales gross, which Anthropic says follows established accounting. OpenAI submitted its own confidential draft S-1 in June, yet Sam Altman told Fortune in September that listing this year would be an “ill-advised moment,” citing safety work. Until OpenAI files publicly, Anthropic’s audited accounts will set the benchmark for frontier AI economics, which is why OpenAI would prefer its own price settled first.

The Market Sets the Mood

Private valuations rarely move in isolation, and this one has not. The Nasdaq Composite fell about 4.5% in the week to October 16, its worst week since late June, as hot inflation data, debt-funded data center spending and Chinese memory competition weighed on the AI trade. The Philadelphia Semiconductor Index lost 11.4% that week, and CoreWeave shed roughly a quarter of its value as yields on its 2031 bonds topped 16%, Reuters reported. On Thursday the Nasdaq fell 2.6%, its steepest session since June, and Meta dropped nearly 7% after reporting negative third-quarter free cash flow of $2.12 billion, raising the floor of its 2026 capital spending range to $135 billion and creating nonvoting shares to help finance growth.

OpenAI’s largest backers absorb that sentiment directly. SoftBank completed a final $10 billion tranche on October 1, lifting its cumulative investment in OpenAI to $64.6 billion for an ownership interest of about 13%, and funded the payment with foreign-currency senior notes. Its shares, which dropped as much as 7.3% on October 9 after the FT’s revenue report, fell 6.9% in Tokyo on Thursday to half their June 2 record, Bloomberg reported, as rising borrowing costs drew attention to its leverage. The FT also cited funding fatigue among investors already heavily committed to AI. When the shareholders of OpenAI’s strategic partners are questioning returns on AI capital, a private price set without reference to public markets becomes hard to hold, even if $1.25 trillion would still value SoftBank’s stake well above its cost.

What a Lower Price Buys

A lower price buys OpenAI time for its commercial expansion to catch up with its valuation, and that expansion has not slowed. On October 21, the company announced that its frontier models and Codex agentic workflows are coming to Google Cloud’s Model Garden, sold through Google Cloud Marketplace and eligible to draw down customers’ existing spend commitments, with general availability planned in the coming weeks. Microsoft remains its primary cloud partner, OpenAI said, but with Amazon Bedrock already carrying its models, OpenAI now reaches enterprises through all three of the largest US cloud platforms. A day later, Uber began rolling out ChatGPT Work, OpenAI’s agent for multistep knowledge work, to more than 30,000 employees, with GPT-6.1 Sol as the default model. Alongside the equity, CNBC reported, OpenAI is negotiating a commercial agreement to secure additional computing capacity from the UAE, pairing the capital with the capacity it is meant to fund.

In effect, OpenAI has exchanged $150 billion of headline valuation for speed and certainty, before Anthropic’s prospectus resets the terms of comparison across the sector. The coming weeks will test that judgment on three counts: whether signatures arrive by the end of next week at $1.25 trillion and the full $30 billion, whether the $70 billion year-end target is met on OpenAI’s own accounting, and how investors read Anthropic’s filed numbers. Chief Financial Officer Sarah Friar has told employees that OpenAI will be a public company in 2027, and public investors will not accept a fixed price, a contested revenue definition or a forecast taken on trust. By meeting the private market rather than dictating to it, OpenAI has received an early lesson in the discipline a listing will demand, and has received it at a valuation most companies would still envy.