Sunday 1 November 2026International EditionRSS
The financial daily of the startup economy
Startups

OpenAI trims its asking price to $1.25tn as it races Anthropic for the same investors

A $30bn round led by Coatue and Gulf money would still be the richest private mark ever. The discount from $1.4tn says more about the market than the headline does.

· · 5 min read

OpenAI is working to raise up to $30bn at a pre-money valuation of $1.25tn, according to people familiar with the talks who spoke to CNBC on Wednesday, October 22nd. Coatue is reported to be leading, with investors from the United Arab Emirates alongside, and the equity is being negotiated together with a separate commercial arrangement that would give OpenAI more computing capacity in the UAE. OpenAI declined to comment, and every outlet that has followed the story stresses that terms could still change.

If it closes on those numbers, the round would set a record for the most valuable private company ever. Yet the more telling figure is the one that has gone missing. At the end of September Bloomberg reported that OpenAI was seeking the same $30bn at a valuation of $1.4tn. Within roughly three weeks the asking price has fallen by about $150bn, or close to 11%. For a company whose valuations have only ever moved in one direction, that is a notable change of tone, and it arrives at a moment when its closest rival is preparing to test public investors directly.

From $1.4tn to $1.25tn

The round has been in motion since early October. On October 5th Bloomberg reported that OpenAI was in talks with several UAE investment funds, including Abu Dhabi's MGX, to anchor a raise of about $30bn, with the Gulf syndicate discussing as much as $10bn in total and BlackRock in discussions to join. According to a summary by AI Weekly, the $1.4tn figure was presented to BlackRock and MGX as a fixed pre-money price. By October 22nd the reported number was $1.25tn, and Bloomberg separately confirmed the lower level.

Two pieces of context explain most of the gap. The first is revenue. On October 8th Bloomberg, following a report in the Financial Times, said OpenAI was on track for annualised revenue of roughly $50bn on current performance, well below reports a month earlier that it was already close to $70bn. The difference was attributed to investors seeking to compare OpenAI's numbers more directly with Anthropic's, including how sales made through cloud providers are counted. According to Reuters, as summarised by investingLive, OpenAI has told investors it expects annualised revenue of about $70bn by the end of the year.

The second is the calendar. OpenAI's last priced round, in March, valued it at $852bn post-money. The new pre-money figure is about 47% higher than that, a large step for seven months but a smaller one than the $1.4tn ask implied. At $1.25tn the company would be valued at about 25 times the $50bn run rate, or roughly 18 times the year-end target. Those are multiples that leave little room for another revenue surprise.

A bridge, not a destination

OpenAI filed confidential paperwork for an initial public offering in June. Since then it has, according to reports, pushed its own listing back to at least next year, and TechCrunch has described the new money as an opportunistic round raised instead of pressing ahead with a flotation. That makes the transaction a bridge in substance if not in name: capital to carry the company through a period of very heavy spending until public markets can be approached on better terms. AI Weekly reports that OpenAI plans to spend $600bn on compute by 2030, a commitment that no private round, however large, can fund on its own.

The structure of the investor group fits that reading. Coatue is a crossover investor that typically wants a position ahead of a listing. The UAE funds bring scale and, in the parallel compute negotiation, something closer to a strategic partnership than a passive stake. SoftBank, already a major shareholder, was also negotiating to take part, according to Reuters. Tying equity to a capacity deal is efficient for both sides, but it also means that the headline valuation is not a clean arm's length price. Part of what the Gulf investors receive is access and alignment that does not show up in the share count.

The dilution, by contrast, is modest. Even if the full $30bn is raised, the new shares would represent only a little over 2% of the post-money company. For existing holders, the question is less about how much they are giving up than about what mark the round puts on their books.

Anthropic sets the benchmark

The timing is not accidental. Anthropic raised $65bn at a valuation of $965bn in May, filed confidentially for an IPO in June and, according to CNBC, is seeking a valuation of between $1.5tn and $2tn when it lists. Reuters reports that its debut is expected in mid-November, and that OpenAI is rushing to secure commitments before then because the two companies draw on much the same pool of large institutional investors. AI Weekly reports that OpenAI has told investors it aims to secure signatures by the end of next week.

That sequencing creates an unusual dynamic. Until now, OpenAI has set the reference price for frontier AI and its rivals have been valued against it. If Anthropic prices a public offering above $1.5tn while OpenAI closes privately at $1.25tn, the order of the two companies would be reversed, and the reversal would be set by a public order book rather than a negotiated term sheet. Some investors, Reuters reports, see OpenAI's push as an attempt to capture capital that might otherwise go into Anthropic's offering. Conversely, a cooler reception for Anthropic would make the lower OpenAI price look prudent rather than defensive.

What changes for private markets

For the wider venture market, the significance lies less in OpenAI's own numbers than in the precedent. A private company of this size has accepted a lower price than it first sought, after investors asked for more comparable disclosure on revenue. That is how a functioning market is supposed to behave, and it is a reminder that even the most sought after assets face diligence once the sums involved approach the scale of public equity offerings.

It also illustrates how the largest AI rounds are now built: a crossover lead, sovereign capital, a strategic counterparty and a commercial agreement running in parallel. Founders further down the chain should expect late-stage investors to borrow the same tools, notably closer scrutiny of how revenue is defined and a growing preference for valuations that can be defended against a listed comparable.

Our view

The $1.25tn headline will be read as a triumph, and on any historical scale it is one. We think the more useful signal is the $150bn that disappeared between the first ask and the second. Investors have shown that they will pay a great deal for frontier AI, but they now insist on comparing like with like, and they are pricing OpenAI with one eye on Anthropic's prospectus. That is healthy. It also means the private mark is only provisional: within weeks a public market will offer its own view of what a leading AI laboratory is worth, and every late-stage portfolio holding AI exposure will be revalued against it. Founders should take note that the definition of revenue has become a negotiating point at the very top of the market. What to watch: whether the round is signed at the full $30bn and on the reported terms, how the UAE compute agreement is structured, and where Anthropic prices. A smaller close, or a slip beyond the target date, would suggest that even $1.25tn was a stretch.

Sources

Filed under Startups. Written by InvestTimes Editorial under our editorial standards. Commentary and analysis for information only, not investment advice.

More in Startups

All Startups analysis