Hero image for OpenAI's Confidential IPO S-1: $25B Revenue, But Losing $1.22 Per Dollar Earned

TL;DR

OpenAI filed a confidential S-1 on May 22, 2026, targeting a September public listing above $1 trillion even as the company projects net losses of $14–16 billion for this year — about $1.22 in operating loss for every dollar earned. The company’s $25 billion annualized revenue (roughly $2 billion per month) masks brutal unit economics, and at a public asking price above its $852 billion private valuation the revenue multiple surpasses even the priciest tech IPOs of the 2020s. Microsoft holds roughly 27% equity — edging out OpenAI Foundation’s 26% — raising governance questions about who ultimately steers the company away from nonprofit origins. Scrutiny deepens with Anthropic having just posted its first profitable quarter, giving public-market investors a direct profitability benchmark to hold against.

The S-1 Filing That Set the AI World Talking

OpenAI’s confidential S-1 registration statement, filed with the U.S. Securities and Exchange Commission on May 22, 2026, marked a watershed moment for the technology sector: the first serious attempt by a pure-play artificial intelligence company to go public at scale.[^1] The filing targets a September public listing with an asking price that would push past the $1 trillion mark — a threshold no AI-native business has ever approached.

Goldman Sachs and Morgan Stanley serve as lead underwriters for what could become the largest IPO of the decade. Under confidential filing rules, OpenAI was permitted to keep its preliminary pricing terms private, allowing the company to gauge institutional investor appetite before disclosing material financial data.[^1] That strategy held until detailed financials leaked in early June — figures that forced the market into an abrupt reckoning with the economics underneath the hype cycle.[^1]

There is no precedent for this. No pure-play AI company has successfully navigated a public offering at this magnitude, making OpenAI’s S-1 not just a corporate milestone but a sector-setting exercise in valuation methodology. Every remaining private AI startup will watch how the SEC scrutinizes its disclosures, and every public-market investor will study whether this listing establishes a new paradigm or confirms longstanding skepticism toward pre-profitability tech valuations.[^1]

$25 Billion a Year in Revenue — But Burning $1.22 for Every Dollar Earned

OpenAI’s top-line numbers are staggering by any standard: approximately $25 billion in annualized revenue, translating to roughly $2 billion per month. To put this in perspective, that puts OpenAI on pace to generate revenues approaching those of Fortune 500 incumbents — all within a compressed timeline that would have been unimaginable just five years ago. Revenue streams are diversified across API access tiers (GPT-4/GPT-5), ChatGPT subscription services, enterprise licensing agreements, and co-development deals with Microsoft through Azure.[^2]

But beneath that revenue ceiling lies a harrowing profitability picture. The S-1 disclosures reveal operating losses of $1.22 for every dollar earned — meaning the company is spending more to acquire each unit of revenue than it brings in. Projected 2026 net losses sit near $14 to 16 billion, one of the largest loss projections for any IPO candidate in modern history.[^2]

The gap between OpenAI’s revenue magnitude and its path to profitability is not closing — it is widening. This dynamic will be scrutinized far more aggressively by public-market investors than private ones ever would have. Where venture capital can forgive sustained losses in exchange for narrative dominance and optionality, public markets demand either a credible timeline to breakeven or an unassailable competitive moat. OpenAI has neither on clear display as it approaches the public listing.[^2]

The Valuation Disconnect: From $852 Billion to Public-Market Skepticism

OpenAI’s last private valuation of $852 billion arrived after a record-breaking $122 billion fundraising round in March 2026 — already an extraordinary figure for any company at this stage of development. At the above-$1 trillion public seeking price, OpenAI would command revenue multiples among the steepest of any IPO in history, surpassing even the most expensive technology exits of the past decade.[^3]

The tension between private-market enthusiasm and public-market discipline defines the central challenge of this offering. Private investors — many of whom are sovereign wealth funds and mega-funds with decades-long time horizons and deep relationships with Microsoft — rewarded OpenAI for narrative dominance, technological headroom, and option value in a market still being defined. Public investors, by contrast, trade daily on quarterly earnings calls and will demand far more concrete metrics: durable revenue retention, gross margin trajectories, customer acquisition cost trends, and clarity on where cash flows turn positive.[^3]

Historically, the biggest tech IPOs of the 2020s that commanded premium multiples — companies like Snowflake at its peak or Airbnb at listing — had demonstrable paths to profitability visible within their filings. OpenAI’s S-1 tells a different story: immense revenue potential married to equally immense capital intensity. The multiple gap between private euphoria and public realism is the defining tension of this offering.[^3]

Nonprofit-to-Foundation Restructuring: Who Really Controls OpenAI?

The corporate governance structure that will face the sharpest scrutiny in the S-1 centers on OpenAI’s transition from its original nonprofit mission to its current foundation-led form. The restructuring left the OpenAI Foundation with only 26% of equity — a figure that fundamentally alters the non-profit mission narrative that gave the company its early legitimacy and public goodwill.[^3][^2]

Microsoft, by contrast, holds approximately 27% of equity, making it the single largest stakeholder in the entity. This arrangement — a multinational tech corporation holding slightly more ownership than the charitable foundation after which the company is named — raises governance questions that regulators and institutional investors will pore over for months.[^3][^2]

Who can dictate OpenAI’s strategic direction if profit motivations begin overriding public-interest pledges? The complexity of this cap table will be a focal point for both institutional investors evaluating whether OpenAI’s dual-purpose structure produces stable governance, and regulators considering whether the company’s founding ethos has been hollowed out by capital-raising necessity. It is a detail that no financial metric can capture on its own but one that could prove consequential in how the market prices OpenAI’s long-term reliability as an investment.[^3][^2]

Anthropic’s Profitability Narrows the Narrative Gap

No discussion of OpenAI’s IPO is complete without noting that its closest competitor has already arrived at a milestone OpenAI cannot claim: first full quarter of profitability. Anthropic, founded by former OpenAI researchers Noam Shazeer and Dario Amodei after they left over disputes about the company’s commercial direction, has posted profits on revenue generated from API subscriptions, enterprise contracts, and government work — all while maintaining a leaner infrastructure footprint than OpenAI’s sprawling data center expansions.[^2]

This development narrows what had been OpenAI’s primary competitive moat — being the only major AI lab with commercially dominant model access. Anthropic’s leaner cost structure challenges the assumption that massive capital expenditure is necessary to build commercially viable AI companies, a point that will be amplified when public-market investors apply cross-company profitability benchmarks.[^2]

As the broader AI competitive landscape continues densifying — with Google DeepMind, Mistral, and xAI all advancing rapidly — OpenAI can no longer rely on first-mover narrative to justify its premium asking price. The profitability benchmark set by Anthropic essentially establishes a new baseline for public-market comparison that will inevitably drag down investor willingness to accept OpenAI’s current multiple expectations.[^2]

What This IPO Means for AI Practitioners and the Industry

The financial sustainability of OpenAI will have direct downstream effects on every stakeholder in the artificial intelligence ecosystem. Enterprise customers are actively monitoring whether subscription tier pricing, API cost structures, or usage caps shift materially following a public listing — history suggests that post-IPO companies facing margin pressure often pass costs along to business customers first.[^3]

Public-market discipline imposed on OpenAI’s burn rate will force competitors — both profitable and unprofitable — to accelerate their own paths toward unit economics sustainability. This could catalyze the next wave of consolidation in the AI sector, as undercapitalized mid-tier labs find valuation gaps with public markets widening rather than closing.[^3]

At the practitioner level, model-access stratification between free and premium tiers is likely to intensify before any potential market correction — a strategic imperative for platforms competing directly with GPT-class models when revenue sustainability becomes the overriding corporate concern.[^3]

Ultimately, whether this IPO succeeds or falters will set the valuation benchmark against which every remaining private AI company will be measured. The outcome signals not just for OpenAI but for how capital markets will price the entire next generation of AI infrastructure firms — a precedent that reverberates far beyond Wall Street.[^3]

The Line Between Hype and Historical Significance

OpenAI’s confidential S-1 paints a portrait of a company that has never had to reconcile ambition with accountability. From the $25 billion annualized revenue ceiling[^1] to its relentless per-dollar burn of $1.22[^2], the figures alone are enough to reframe how every remaining private AI lab will be measured. But beneath the headline numbers, three deeper tensions define this offering and, by extension, the entire AI sector.

The first is valuation itself: what took sovereign wealth funds and mega-funds twelve months of patient capital-raising to reach $852 billion[^3] could collapse on public-market discipline within days of listing. Revenue multiples beyond any precedent in the 2020s tech IPO landscape cannot be sustained by narrative alone when quarterly earnings calls arrive with ruthless clarity.[^3] The second tension is governance — a corporate structure that left Microsoft holding marginally more equity (27 percent) than the OpenAI Foundation itself (26 percent)[^3][^2] raises questions no SEC disclosure format can answer: who truly steers an entity named for its nonprofit origins after decades of capital-intensive growth?[^3][^2] Third, Anthropic’s successful first profitable quarter[^1][^2] strips away OpenAI’s last defensible moat — being the only major lab with dominant model access — and hands public-market investors a concrete profitability benchmark against which OpenAI’s spending will be judged.

The implications extend far beyond Wall Street. Enterprise customers should expect post-IPO margin pressure to land on subscription tiers and API pricing first.[^3] Undercapitalized mid-tier labs may find their valuation gaps deepening as the gap between public-market realism and private-market expectation widens.[^3] Model-access stratification between free and premium tiers will intensify before any correction comes. Whatever this IPO delivers — triumph or reckoning — it sets the precedent against which the entire next generation of AI infrastructure firms will be priced. The question is not whether OpenAI goes public, but whether the market that receives it will be willing to pay for what it actually is today rather than what its S-1 says it might become tomorrow.

Methodology

This article is based on three primary sources: Krishna Kareth’s LinkedIn analysis of the confidential S-1 filing (capturing key filing details and timeline), Sacra’s detailed financial breakdown of OpenAI’s revenue and loss figures, and Klover AI’s in-depth analysis of the IPO’s market implications and governance structure. All figures attributed to S-1 disclosures are sourced through these secondary analyses, as the confidential filing remains under SEC review and is not publicly accessible in its original form. Financial projections, valuation multiples, and competitive comparisons are drawn from these analyst interpretations and should be read as informed estimates rather than audited figures.

Editor’s Notes

This article was written before the formal S-1 pricing range was publicly disclosed. The $1 trillion+ valuation represents the asking price as understood through analyst sources rather than a confirmed IPO price range.
The Anthropic profitability comparison refers to the company’s first full quarter of positive net income as reported through public statements and market analysis; Anthropic’s financials are not publicly filed.
Revenue, loss, and equity-holding figures should be treated as analyst estimates drawn from leaked financial details rather than SEC-audited disclosures, which remain confidential until 15 days before the roadshow.

Source list

Krishna Kareth — LinkedIn analysis of OpenAI confidential S-1 filing (May 22, 2026): https://www.linkedin.com/posts/kkareth_openai-files-confidentially-for-ipo-following-activity-7469872564329164800-58ow
Sacra — OpenAI financial analysis and burn rate breakdown: https://sacra.com/c/openai/
Klover AI — In-depth OpenAI IPO viability analysis: https://klover.ai/openai_ipo_viable_business_indepth_analysis_2026/

Trust Stack

Last checked: 2026-06-29
Corrections: Contact us to report errors

Change log

2026-06-29: Article assembled and published. All sections (intro, body, conclusion) reviewed and signed off through the OKF pipeline.