OpenAI's S-1 Is Weeks Away, and the Whole Industry Is About to Get Graded
OpenAI's public S-1 filing is expected within weeks ahead of a targeted September IPO, and the numbers inside it will reset how every AI company is valued.
The most consequential document in AI this year will not be a model card. It will be a registration statement.
OpenAI is expected to file its public S-1 prospectus within the next few weeks, ahead of an initial public offering targeted for September 2026, according to reporting tracked by Build Fast with AI and IPO watcher StartupHub. The confidential version of that filing went to the SEC on June 8, per multiple outlets covering the process, and the public version is what everyone is actually waiting for. Confidential filings tell the regulator. Public filings tell the market.
The context matters. OpenAI's last private round, in March 2026, raised $122 billion at an $852 billion valuation, per StartupHub's IPO tracker. Reuters had reported back in October 2025 that the company was preparing for an IPO at a valuation of up to $1 trillion, weighing a capital raise of at least $60 billion. Annualized revenue has reportedly passed $25 billion, though that figure comes from press reports and analyst estimates, not audited statements. That is exactly the point. Everything we know about OpenAI's economics today is a leak, an estimate, or a number from a funding deck. The S-1 replaces all of it with signed, liability-carrying disclosure.
Goldman Sachs and Morgan Stanley are the lead banks, according to Gradually's IPO coverage. The same coverage flags the structural oddity that could slow things down: OpenAI is a Public Benefit Corporation formally controlled by a nonprofit foundation, an arrangement with no real precedent at this scale, and one that raises awkward questions about what new shareholders are actually buying.
Why practitioners should care
If you build or operate AI systems, this is not a markets story. It is your budget story.
The S-1 will show, for the first time, how OpenAI's revenue splits across consumer subscriptions, API usage, and enterprise contracts, and how that revenue compares to its compute bill. Those two line items are the gravity well everything else in this industry orbits. API pricing, enterprise discounting, the rate at which free tiers get more generous (OpenAI widened unlimited free ChatGPT access this very month, per Build Fast with AI, a classic pre-IPO engagement push): all of it is set by economics nobody outside the company has seen. Once the numbers are public, every procurement team and every CFO with an AI line item gets a benchmark that did not exist before.
There is a second-order effect worth noting. A successful OpenAI listing likely opens the door for other frontier labs to follow, and the resulting disclosure wave would force the entire sector to be valued on audited numbers rather than narratives. A weak listing does the opposite: it tightens private capital for everyone downstream, including the startups whose tools you probably use.
The case for
The bullish case, articulated across analyst coverage from Zacks to Value Add VC, is that this IPO is the AI boom's graduation ceremony. Revenue growth is real and steep: reportedly past $25 billion annualized, up from single-digit billions not long ago. Microsoft has separately disclosed $24.1 billion in AI revenue tied to its OpenAI relationship, per Build Fast with AI's roundup, which suggests the demand is not confined to one company's self-reported figures.
The structural argument runs deeper. Competing at the frontier now requires capital at a scale private markets struggle to supply. Anthropic alone has roughly $71 billion in compute commitments, including a $10 billion contract with infrastructure firm Volta, and the hyperscalers are spending near $700 billion on data centers in 2026, per the same reporting. Public equity is the only pool deep enough to fund that. On this view, the IPO is not an exit. It is refueling.
And the discipline cuts both ways. Public disclosure forces OpenAI to defend its unit economics quarterly, which bulls argue will push the whole industry toward sustainable pricing and away from subsidized land-grab behavior.
The case against
The skeptics have a simpler question: what if the numbers are bad?
Secondary market pricing already implies $900 billion to $1.1 trillion, Value Add VC's analysis notes, meaning pre-IPO buyers are paying IPO-level prices without IPO-level information. If the S-1 shows losses scaling with revenue, or unit economics that depend on compute costs nobody can control, the correction would not stay contained to one stock. It would hit the private valuations of every AI company benchmarked against OpenAI, which is most of them.
The governance critique is sharper still. Gradually's coverage points out that the nonprofit-controlled PBC structure may leave public shareholders with limited actual say, an unusual ask at a trillion-dollar valuation. Critics have also questioned the timing: filing into a market where Palantir trades up roughly 93 percent on AI enthusiasm, per Build Fast with AI, looks less like confidence and more like selling while the window is open. Fair question.
Finally, the capex skeptics. If OpenAI's disclosed compute commitments imply years of negative free cash flow, the IPO starts to resemble a pass-through from public investors to chipmakers and data center landlords. That is a harder story to sell on a roadshow.
What to watch
Three dates matter now. First, the public S-1 itself, expected within weeks: read the revenue mix and the compute cost lines before any headline. Second, the September target window, which Gradually notes could slip to Q4 given the corporate structure complexity. Third, the reaction of other labs. Anthropic's next move, and whether the rumored IPO wave actually forms, depends almost entirely on how this one lands.
Whatever the filing shows, the era of valuing frontier AI on vibes ends the day it drops.