Data Labeling

Altman Tells OpenAI Staff to Expect IPO Within the Next Year

Altman says filing now buys flexibility, but faster AI progress could push the listing back.

Liza Chan
Liza ChanAI & Emerging Tech Correspondent
June 11, 20263 min read
Share:
Abstract representation of a tech company weighing public markets against rapid AI advancement, financial charts and server infrastructure

Sam Altman told OpenAI employees in a Slack message Monday that he expects the company to go public within the next year, according to The Information. The message landed around the same time OpenAI said on its company blog that it had confidentially filed a draft S-1 with the SEC.

The caveat is the story

Most of the coverage led with the timeline. The more interesting part is the condition Altman attached to it. He told staff that if OpenAI's technology advances to the point where its AI can build new AI on its own, what the field calls recursive self-improvement, a quick listing becomes less attractive. The faster that takeoff looks, he reportedly wrote, the more it could pay to stay private, because the technology and the world could shift in unexpected ways.

Read plainly, that is a CEO saying the more the core bet pays off, the less he wants public shareholders in the room. Filing now, in his framing, is just keeping the door open. PYMNTS reported the cautious tone clashes with a louder investor narrative that OpenAI and Anthropic are sprinting to list as fast as possible.

So why file at all?

Money, mostly. OpenAI is in advanced talks to lease a 10-gigawatt data center campus on federal land in southern Ohio, a project Reuters reported could cost at least $500 billion to build out across chips, power, and labor. Nvidia may guarantee the lease and the developer's financing. The first phase isn't expected online until 2028.

That is the tension. Recursive self-improvement argues for staying private. A half-trillion-dollar compute habit argues for tapping public capital. Altman is trying to hold both, and "filing now gives us optionality" is the corporate way of saying he hasn't decided.

"We recently submitted a confidential S-1. We expect it to leak so we're just announcing it." That was OpenAI's own blog, which is a more honest line than most pre-IPO communications manage.

A new model, conveniently timed

Separately, chief scientist Jakub Pachocki told staff that a new model, codenamed 5.6, is a meaningful improvement over the current flagship, GPT-5.5, and is expected this month. "Meaningful improvement" is doing a lot of work in that sentence, and OpenAI ships point releases often enough that the bar for the word has drifted. We'll know more when benchmarks land.

One detail with a clearer dollar value: OpenAI is preparing a tender offer letting employees sell shares at the current price of $687.69. For staff watching the IPO clock, that number is the one that matters right now.

The model is due this month. The Ohio lease, if it closes, starts payments only once the site runs in 2028. The IPO window stays open, undated, and dependent on how fast the technology moves.

Tags:OpenAIIPOSam Altmanartificial intelligencerecursive self-improvementdata centersNvidiaGPT-5.5SEC filing
Liza Chan

Liza Chan

AI & Emerging Tech Correspondent

Liza covers the rapidly evolving world of artificial intelligence, from breakthroughs in research labs to real-world applications reshaping industries. With a background in computer science and journalism, she translates complex technical developments into accessible insights for curious readers.

Related Articles

Stay Ahead of the AI Curve

Get the latest AI news, reviews, and deals delivered straight to your inbox. Join 100,000+ AI enthusiasts.

By subscribing, you agree to our Privacy Policy. Unsubscribe anytime.