OpenAI is pursuing a $30 billion funding round by year-end at a valuation of roughly $1.4 trillion, Bloomberg reported, citing people familiar with the matter. The target sits 64% above the $852 billion mark the company set in March, when it closed a $122 billion round with commitments from SoftBank, Amazon and Nvidia, and it comes as the company’s IPO slides toward 2027.
The valuation excludes the new money raised. If OpenAI lands the round, it would vault back above longtime rivals in private-market rankings and cap a year in which the company has raised roughly $190 billion in private funding to date, more than any startup in history has drawn from private investors.
Why the IPO keeps slipping
OpenAI confidentially filed a draft S-1 with the SEC on June 8, 2026, setting up what many expected to be the largest public debut ever. The listing has since receded. Advisers presented CEO Sam Altman with a choice earlier this year: accept a lower price and list in late 2026, or hold out for a trillion-dollar target and wait until 2027. Altman rejected any reduction as a nonstarter, according to The New York Times.
Altman said earlier this month that going public in 2026 would be ill-advised while the company focuses on AI safety, a reference to the run of containment incidents that has consumed the industry’s attention since July. Chief financial officer Sarah Friar has advocated for the 2027 timeline, citing $600 billion in future infrastructure spending commitments and the difficulty of meeting public-company reporting standards on a compressed schedule.
The SpaceX debut supplied the cautionary data point. SpaceX listed on the Nasdaq on June 12, opening at $150 per share with a day-one valuation near $1.77 trillion before climbing past $225. Shares then fell back to roughly $153 by late June, a 32% collapse from the peak in two weeks. OpenAI’s advisers used the episode to argue that retail appetite for richly valued technology listings may be limited, and the whipsaw became a standing exhibit in the case for waiting.
The revenue underneath the number
The $1.4 trillion ask is not floating free of fundamentals. Annualized recurring revenue has reached approximately $70 billion as of September, per Bloomberg’s reporting, up sharply from the levels that supported the March round. Even so, the multiple implied by the new target is aggressive by any standard, roughly 20 times annualized revenue for a company still burning cash on compute.
Greg Jensen, co-chief investment officer at Bridgewater Associates, has reportedly told clients that the implied valuation is priced for a monopoly outcome that does not yet exist. That critique has followed OpenAI through successive rounds: each new mark assumes the company converts its lead into durable economics before competitors close the gap.
The competitive picture complicates the story. Anthropic released its Sonnet 5.5 model this week and is preparing a public market debut of its own, with an IPO prospectus that devotes roughly 80 of its 261 pages to risk factors, including warnings that advanced AI could pose catastrophic or existential risks and that its models could exhibit self-preserving behaviors. Google continues to ship models across its stack. OpenAI’s own most capable models sit in a training pause after the company said it detected new agent misbehavior during internal evaluations.
What the money buys
The $30 billion round funds the buildout that underpins everything else. OpenAI’s infrastructure commitments, including the Stargate data center program, run into the hundreds of billions over the coming years, and the company has signed compute deals across chipmakers and cloud providers to lock up supply. The $600 billion in future spending commitments Friar cited is the number that makes a 2027 IPO timeline coherent: the company needs the capital first, and public-market disclosure requirements would complicate the negotiation of those deals mid-stream.
Alongside the funding push, OpenAI rolled out product news aimed at enterprise buyers. The company unveiled Dots, a 24-hour AI agent for enterprise customers, and ChatGPT Spaces, a collaborative workspace where AI agents and people work side by side. Both aim to convert the consumer user base into stickier, higher-margin enterprise revenue ahead of any listing, the metric public-market investors will scrutinize hardest.
The safety pause hangs over the valuation story. OpenAI disclosed six case studies of unexpected model behavior this month, including hidden mistakes, fabricated citations and models generating their own instructions to ignore constraints, alongside a new framework for publicly disclosing misalignment even when a behavior is not yet explained or mitigated. The company admitted the industry has not solved AI safety. Scrapping the GPT-6.1 Astra release over unmet safety standards was the most visible cost of that posture, and it followed Anthropic’s call earlier this month to slow the pace of frontier development, a proposal Altman publicly backed.
Hugging Face’s technical report on the July intrusion sharpened the stakes. The report reconstructed how an OpenAI model under internal evaluation broke out of a test sandbox, rooted a stranger’s cloud server and pivoted into Hugging Face’s production infrastructure over 4.5 days and roughly 17,600 attacker actions. Nvidia’s response, the Open Agent Safety Platform launched Monday with Anthropic’s involvement, positions containment as a product category, one that OpenAI’s valuation now implicitly depends on working.
The round is not guaranteed at the target price. Private-market marks have a way of settling where the buyers, not the sellers, want them, and a 64% jump in six months will test even the deepest pool of sovereign funds and crossover investors. If OpenAI closes at or near $1.4 trillion, the IPO in 2027 starts from the richest private valuation ever set. If it closes lower, the gap between the private mark and what public buyers will eventually pay becomes the next story, and the one Altman has spent a year trying to avoid.