OpenAI Files for a $1 Trillion IPO. It Loses Money on Every Dollar It Makes.

Correction (October 9, 2026): An earlier version of this article stated as fact that OpenAI filed its S-1 on May 22; that date appears in some reports but OpenAI, which announced the confidential filing on June 8, has not confirmed it. It also said ChatGPT had 500 million weekly users (OpenAI reported more than 900 million), listed JPMorgan as a lead bank, set a firm Q4 2026 target window, and cited a $25 billion GAAP loss estimate and a "$3 trillion AI IPO race" we could not verify. Those details have been corrected or removed, and Anthropic's filing date has been corrected to June 1.

OpenAI announced on June 8 that it had "recently submitted a confidential S-1" to the SEC, taking the first official step toward a public listing that, according to media reports, could value the company at up to $1 trillion. The number is staggering. The math behind it is even more so. On an adjusted basis, OpenAI lost an additional $1.22 for every dollar of revenue in Q1 2026 — a negative 122 percent non-GAAP operating margin reported by The Information, which on $5.7 billion of quarterly revenue works out to a loss of roughly $6.95 billion. The company that invented the modern AI era is now asking public markets to believe the losses are temporary and the valuation is not.

How We Got Here

OpenAI's last private funding round closed on March 31 — a $122 billion raise anchored by Amazon, Nvidia, and SoftBank at an $852 billion post-money valuation, the largest private technology financing in history. Goldman Sachs and Morgan Stanley have reportedly been lined up as lead underwriters, and media reports suggest a listing could come as soon as September. OpenAI itself says it has "not decided on timing yet; it may be a while." The confidential filing format keeps the full details private until at least 15 days before an investor roadshow begins, but the company has disclosed enough to construct a picture: OpenAI said in March it was generating $2 billion in revenue per month, roughly a $24 billion annual pace, and that enterprise makes up more than 40 percent of revenue. In February it reported more than 50 million ChatGPT subscribers and more than 9 million paying business users.

The Uncomfortable Numbers

The revenue trajectory is genuinely impressive. The loss picture is something else. Internal projections reported by The Information in 2024 put OpenAI's 2026 loss at up to $14 billion, a figure that excludes stock-based compensation, with no annual profit expected until 2029. If the Q1 margin held for the full year, the loss would run well above that. That makes OpenAI one of the most cash-intensive businesses ever to approach a public listing. The losses are structural: training frontier models, running inference at scale for more than 900 million weekly users, and maintaining the GPU infrastructure to stay ahead of Google, Anthropic, and Meta requires capital at a rate that current revenue cannot cover. The gap is not a bug in the business model — it is the business model, at least for now.

Why $1 Trillion Anyway

The trillion-dollar target is not irrational, even if it looks that way at first. The valuation is a bet on trajectory, not on the current income statement. If revenue continues growing at its current pace and the loss ratio compresses even modestly over the next two to three years — through improved inference efficiency, higher-margin enterprise contracts, and continued subscriber growth — the company reaches profitability at scale. The argument is the same investors made for Amazon in 2000 and Netflix in 2012: when the network effect is large enough and the strategic moat is wide enough, today's losses are the tuition bill for tomorrow's monopoly. OpenAI has more than 900 million weekly users, a deep distribution partnership with Microsoft that has embedded ChatGPT into enterprise workflows globally, and a brand that has become synonymous with the technology category it created. The strategic position is real. Whether it is worth $1 trillion right now is a different question.

What This Signals for the Entire Sector

What makes this IPO consequential beyond the dollar figure is what it does to the broader AI funding environment. Anthropic filed its own confidential S-1 on June 1, a week before OpenAI's announcement, at a $965 billion private valuation. Both filings together put the AI infrastructure era on a direct collision course with public market accountability — quarterly earnings calls, analyst coverage, shareholder pressure to show a path to profitability. The era of indefinite private burn, insulated from scrutiny by ever-rising private rounds, is ending. When OpenAI goes public, it will be required to disclose, in granular detail, exactly how much it costs to build and run the world's leading AI platform. That disclosure will reprice expectations across the entire sector. The same test is already reaching hardware: Agility Robotics is heading to Nasdaq at a $2.5 billion valuation on $1.8 million of 2025 sales.

The Question No One Can Answer Yet

The honest reality is that nobody knows where this lands. OpenAI is growing revenue faster than almost any company in history, but it is also spending at a rate that makes the absolute loss figures almost secondary — the trajectory matters more than any single quarter. The IPO will test whether public market investors are willing to extend the same faith that SoftBank, Amazon, and Nvidia already have: that the company building the infrastructure layer for the AI era deserves to be priced on potential rather than fundamentals. If they are, this will be the most watched market debut since Meta's 2012 listing. If they are not — if institutional investors balk at underwriting tens of billions in annual losses at a trillion-dollar price tag — every other private AI company's valuation gets re-examined overnight. The filing is out. The reckoning is coming.

Sources: MLQ.ai · CoinDesk · Where's Your Ed At, citing The Information · GIGAZINE · The Decoder, citing The Information