The Filing That Changes Everything About AI Investment
Within twenty-four hours of SpaceX publishing its S-1 prospectus, CNBC and the Wall Street Journal reported that OpenAI is preparing to confidentially file a draft IPO prospectus, with Goldman Sachs and Morgan Stanley advising on the process. The company is targeting a public listing as early as September 2026 at a valuation above $1 trillion, pending regulatory review and market conditions. OpenAI's March 2026 funding round closed at a post-money valuation of $852 billion, with major participation from Amazon, Nvidia, SoftBank, and other investors. Chief Financial Officer Sarah Friar confirmed in a January 2026 blog post that the company's annualised revenue rate exited 2025 above $20 billion. The path from a $20 billion annualised revenue rate to a $1 trillion valuation implies a revenue multiple of approximately 50 times — a premium that reflects not the current financial performance but the market's assessment of what OpenAI controls: the most widely used AI interface in the world, the most capable frontier model available to consumers, and the infrastructure layer through which an increasing share of global knowledge work is being performed.
The OpenAI IPO would, if completed at target, rank as the second-largest IPO in history behind only SpaceX — with the two listings potentially occurring within three months of each other. The combined equity supply from SpaceX and OpenAI, estimated at $135 billion to $155 billion, represents the most significant test of global institutional capital market absorption capacity since the dot-com era — but in a market context that is fundamentally different from 2000. The companies going public in 2026 have actual revenue, actual users, and actual market positions that justify serious valuation analysis. The absorption challenge is not whether the underlying businesses deserve large valuations — they may well — but whether the institutional capital available to fund both listings simultaneously is sufficient to meet the supply without requiring price concessions that undermine the IPO economics for both issuers and their existing investors.
The Corporate Structure Problem: How Do You Take a Nonprofit's AI Lab Public?
OpenAI's path to a public listing is complicated by a corporate structure that has no precise precedent in public market history. The company began as a non-profit research organisation, created a for-profit subsidiary to raise commercial capital, and has been in the process of converting to a fully for-profit entity — a conversion that has generated significant legal challenges, most prominently from co-founder Elon Musk, whose lawsuit seeking to block the conversion was resolved in a manner favourable to OpenAI's restructuring plans. Microsoft holds a 49% economic interest in OpenAI's for-profit subsidiary — a stake that was negotiated in exchange for the $13 billion in cloud computing infrastructure investment that Microsoft provided between 2019 and 2025. The Microsoft relationship creates both a strategic asset and a governance complexity: Microsoft's Azure is OpenAI's primary compute provider, meaning the two companies are simultaneously investor and portfolio company, strategic partner and customer, and potential future competitors as Microsoft develops its own AI capabilities alongside its OpenAI investment.
The valuation implications of the Microsoft stake are material for IPO investors attempting to size their allocation. If Microsoft's 49% economic interest is valued at the implied IPO valuation — approximately $490 billion at a $1 trillion total valuation — it represents an unrealised gain on Microsoft's $13 billion investment of approximately $477 billion, a return profile that would rank among the most successful corporate venture investments in history. Whether Microsoft will retain, monetise, or strategically restructure its OpenAI stake in the context of an IPO is a question that the confidential S-1 filing will address but that has not been publicly disclosed. A Microsoft decision to sell even a fraction of its stake in the public offering would significantly increase the supply of OpenAI equity available to public market investors, potentially requiring the offering to be priced more conservatively to clear the market.
The Revenue Model: Why $20 Billion Annualised Revenue at 50x Is a Thesis, Not a Fact
OpenAI's revenue model is more concentrated and more nascent than the $20 billion annualised figure alone suggests. The overwhelming majority of current revenue is derived from ChatGPT subscriptions — the $20 per month consumer plan and the $30 per month Plus plan — and from API access sold to enterprise customers and developers who integrate GPT-4 and o-series models into their own applications. The business model is straightforward and proven at scale: sell access to large language model inference, priced by the token or by the subscription, to users who find the capability valuable enough to pay for it. What the revenue figure does not yet demonstrate is the durability of that model against free alternatives from Google (Gemini), Meta (Llama), and Anthropic (Claude), or against the potential for OpenAI's enterprise customers to migrate to open-source models that they can host themselves without ongoing API costs.
The bull case for OpenAI's $1 trillion valuation rests on three premises that are plausible but not certain. First, that the ChatGPT brand and user base represent a durable consumer franchise that survives competitive pressure — the way Google's search brand has survived decades of would-be challengers. Second, that the transition from subscription and API revenue to agentic AI platform revenue — where OpenAI's agents perform work on behalf of users and enterprises, generating per-task revenue rather than per-token charges — multiplies the revenue per user by orders of magnitude as AI takes on more consequential work. Third, that OpenAI's position at the frontier of AI capability is sustainable through its continued investment in model training, and that this frontier position commands a premium that justifies platform-scale valuation multiples rather than software-company multiples. Each of these premises is contestable by serious analysts. The IPO process, and the public market's verdict on the offering price, will be the most consequential stress test of all three simultaneously — and it will establish the valuation framework within which the entire AI sector is assessed for years to come.
What OpenAI Going Public Means for the AI Market Structure
The OpenAI IPO matters for the AI market not just as a financial event but as a competitive signal. A successfully priced public listing at or above $1 trillion would validate the commercial scale of the AI application market in a way that no amount of private funding rounds can achieve — because public market pricing reflects the judgment of thousands of professional investors with fiduciary obligations, not the judgment of a small number of late-stage venture funds seeking deployment opportunities. It would also create a liquid, publicly traded equity benchmark against which every other AI company — Anthropic, Mistral, Cohere, Perplexity, and the AI divisions of every major technology company — is implicitly valued. The OpenAI IPO is therefore not merely the listing of one company. It is the establishment of the reference price for an industry. The capital that flows into OpenAI's public offering will flow out of something else — other AI equities, other technology companies, or other asset classes — and the reallocation effects of both the SpaceX and OpenAI listings will reshape portfolio positioning across the technology sector for the remainder of 2026 and well into 2027.