
OpenAI files confidential IPO with SEC one week after Anthropic. The $150B valuation faces a 40x revenue multiple test as two AI prospectuses arrive in quick succession.
OpenAI has confidentially filed a draft registration statement for its initial public offering with the U.S. Securities and Exchange Commission. The filing sets up what could be the largest technology IPO in years, coming roughly one week after rival Anthropic submitted its own confidential IPO paperwork. The concentrated wave of AI company listings will test investor appetite for high-revenue, high-burn business models.
The confidential filing means OpenAI's financials, risk factors, and use-of-proceeds plan remain sealed until the company chooses to make them public, typically 15 to 21 days before the roadshow begins. The SEC's confidential filing process, available under the JOBS Act for companies with under $1.07 billion in annual revenue, allows issuers to refine their prospectus without public scrutiny. OpenAI's revenue has been reported by multiple outlets to exceed that threshold, so the confidential filing may indicate the company is still finalizing its revenue recognition methodology or addressing SEC comments on its business model.
Anthropic's confidential filing roughly one week before OpenAI's creates a sequencing problem for institutional investors. Most IPO allocations are decided by a small group of anchor investors who can absorb large blocks of shares. Those same investors now face two AI IPO prospectuses in rapid succession, each with different revenue profiles, burn rates, and valuation expectations.
OpenAI reported annualized revenue of roughly $3.7 billion as of late 2024, according to published reports, with the majority coming from ChatGPT subscriptions and API access. Anthropic's revenue is estimated at roughly $1 billion on an annualized basis, primarily from API usage by enterprise clients. The valuation gap is wide: OpenAI's last private secondary transactions valued the company at roughly $150 billion, while Anthropic's last funding round valued it at roughly $18 billion.
The risk for both companies is that investors treat them as substitutes rather than complements. If anchor investors allocate capital to the first filer, the second filer may face a thinner book and a lower pricing range. The one-week gap between filings compresses the window for investors to perform due diligence on both companies simultaneously.
OpenAI's revenue mix carries different quality signals than Anthropic's. Subscription revenue from ChatGPT Plus and Team plans is recurring and predictable, with monthly billing cycles and relatively low churn. API revenue, by contrast, is usage-based and tied to developer activity, which can fluctuate with product releases, pricing changes, and competition from open-source models.
Anthropic's revenue is more heavily weighted toward API consumption by enterprise clients, which typically involves longer sales cycles but higher contract values. The company's Claude model family has gained traction in regulated industries such as healthcare and legal services, where data privacy requirements favor a closed-source API model over consumer-facing chatbots.
For IPO investors, the key metric will be gross margin on each revenue stream. Subscription platforms typically run gross margins of 70% to 80% once infrastructure costs are amortized. API businesses face higher variable costs from compute usage, which can compress gross margins to 50% or below during periods of rapid model training and deployment.
The $150 billion valuation attached to OpenAI in secondary markets implies a multiple of roughly 40x trailing revenue. Anthropic's $18 billion valuation implies roughly 18x trailing revenue. Both multiples are high by historical standards for technology IPOs, the comparison matters less than the absolute pricing.
A 40x revenue multiple requires either sustained revenue growth above 50% annually for several years or a path to operating margins that justify the premium. OpenAI's cost structure includes massive compute spending, talent retention costs, and legal expenses related to intellectual property disputes. The company has not disclosed operating income. Published estimates suggest it is burning cash at a rate of several billion dollars per year.
The SEC's review of OpenAI's draft registration will focus on several areas that are unique to AI companies:
Each of these factors could appear as a risk factor in the final prospectus. The SEC may require additional disclosure before declaring the registration effective.
The confidential filing does not set a specific IPO date. The typical timeline from confidential filing to public filing is 60 to 90 days, followed by a two-week roadshow. That would place a potential OpenAI IPO in the second quarter of 2025, assuming no SEC delays or market disruptions.
The broader IPO market has been receptive to large technology listings in 2024 and early 2025, with several multi-billion-dollar offerings pricing at or above their filing ranges. The AI sector carries additional volatility risk. Any major product failure, regulatory action, or competitive shift in the large language model market between now and the IPO date could reset investor expectations.
For traders tracking the AI IPO wave, the next concrete marker is the public filing of OpenAI's S-1 registration statement. That document will reveal the company's financials, risk factors, and proposed ticker symbol. It will determine whether the $150 billion valuation holds or adjusts to market conditions. The one-week gap between the Anthropic and OpenAI filings means the two prospectuses will arrive in quick succession, forcing investors to make allocation decisions with incomplete information on both companies.
OpenAI's confidential IPO filing is the starting gun for what may be the most consequential technology IPO cycle since the 2019 wave of high-growth software listings. The outcome will set the valuation benchmark for every AI company that follows, from Anthropic to a pipeline of smaller model developers and infrastructure providers.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.