The first to lay bare the AI ledger to the public market, the first to undergo financial stress testing.

Written by: Long Yue

Source: Wall Street Insight

Anthropic has filed first, shifting the competition with OpenAI from models, revenue, and valuation, to the threshold of public market pricing.

On Monday, Anthropic announced that it has privately submitted its U.S. IPO application, ahead of competitor OpenAI in the listing process. Anthropic stated in a press release that the submission of the prospectus "gives us the right to choose to go public after the SEC's review is complete," emphasizing that "the proposed IPO will depend on market conditions and other factors."

This is not a formal public offering. According to Reuters, Anthropic's listing could happen as early as this fall, but it did not disclose the scale or terms of the offering. The significance of a private submission is that the company can advance its IPO preparations while temporarily not revealing sensitive financial details to competitors and the public.

This has made the IPO race for AI large model companies suddenly more concrete. Previously, the market was more focused on whose model was stronger and who had more users. Now, the question has become: who will first accept public market scrutiny, and who will set the pricing for "cutting-edge AI companies."

This filing will test whether investor enthusiasm for AI can withstand public market scrutiny and will determine which company will first establish a valuation template for the rapidly growing AI industry.

OpenAI has not yet followed suit with a filing. OpenAI CEO Sam Altman stated that he is "not focused on the timing of a potential IPO," and that the company "will go public at the right time."

However, the market was not originally betting this way. In the prediction market, most participants had previously anticipated that OpenAI would submit an IPO application before Anthropic.

The IPO window opens, but money is not unlimited

Anthropic is racing against time, backed by a very real market window.

Data from Dealogic shows that the IPO market has regained momentum in recent weeks, with global IPO financing reaching $87.5 billion as of May 26, the highest level for the same period since 2021.

Currently, the IPO window has clearly opened. AI chip company Cerebras saw a 68% rise on its first day of listing last month. According to FactSet data, among the companies that listed with valuations exceeding $10 billion in the past five years, only the digital design platform Figma had a higher first-day surge of 250% last year.

However, the opening of the window does not mean that funds are unlimited.

Notably, SpaceX is also pushing forward with a large IPO, aiming to raise $75 billion and achieve a valuation of $1.75 trillion, potentially trading in as little as two weeks. If SpaceX, Anthropic, and OpenAI all list in quick succession, the U.S. stock market would need to digest several large-scale tech assets simultaneously.

IPO research firm IPOX vice president Kat Liu told Reuters, "By filing shortly after SpaceX, Anthropic can take advantage of a favorable window and leverage strong investor interest in AI and growth stocks."

She further stated, "Compared to SpaceX, Anthropic's valuation appeal doesn’t appear as aggressive when viewed in isolation."

Patrick Healy, founder of Issuer Network, said, "There’s only so much oxygen in the room." He also mentioned, "SpaceX will consume a vast amount of capital, and the second one to enter will be in a better position than the third."

D.A. Davidson analyst Gil Luria remarked, "The combined capital requirements of SpaceX, OpenAI, and Anthropic will be substantial, likely disrupting the capital markets, so going public a bit earlier would be a huge advantage."

Filing first is seizing the narrative and absorbing risk

The benefits of being the first to go public are clear: setting the price first, raising funds first, allowing employees and early investors to achieve liquidity first.

However, going public first also has its costs: disclosing financial details first, accepting inquiries from institutional investors first, revealing the true cost structure of the AI company first.

PitchBook senior analyst Harrison Rolfes stated, "The conventional interpretation is that Anthropic has just gained a narrative advantage by filing first."

However, he also provided an alternative view: "The unconventional interpretation is that OpenAI actually comes out better: Anthropic voluntarily assumes all disclosure risks first, allowing OpenAI to observe for free how institutional investors respond to audited cutting-edge AI financial data before deciding on its own pricing."

This sentence highlights the crux of the AI company's IPO. The public market is concerned not only with the "AI story," but also with revenue quality, compute costs, cloud service share, cash burn, customer composition, and profit margins.

According to research cited by The Wall Street Journal, IPOs tend to appear in batches within an industry, and companies listed later in the cycle often perform worse than those listed early. The report explains that companies with deeper moats and higher quality generally go public earlier, followed by a batch of followers.

However, being first does not guarantee success.

In 2019, Lyft went public before Uber, but its stock performance after the IPO fell short of expectations, directly affecting Uber's listing two months later. Uber subsequently lowered its target valuation, but its stock price still fell post-IPO.

The report also noted that Facebook's stock price dropped more than half within three months after its IPO in 2012, amid market concerns about its ability to adapt to the shift towards mobile advertising. Facebook later validated its business model, but other companies that originally intended to go public, including Twitter, ultimately had to wait.

This means that Anthropic's head start could either secure it the pricing power for AI IPOs or make it the first large model company to have its ledger publicly scrutinized.

Why Anthropic: Changing narratives on revenue, valuation, and profits

Anthropic's willingness to move forward now is related to recent financial changes.

Wall Street Insight mentioned that Anthropic’s annualized revenue is approaching $45 billion. OpenAI's annualized revenue has just surpassed $30 billion, currently estimated at around $33 billion. By this measure, Anthropic's revenue scale is at least 35% higher than OpenAI.

This change has occurred rapidly. Reports indicate that by the end of 2025, Anthropic’s annualized revenue was only $9 billion, less than half of OpenAI’s. In the first five months of this year, Anthropic's revenue grew approximately fivefold; during the same period, OpenAI's revenue grew over 50%.

The revenue structures of the two companies are also different. OpenAI's revenue mainly comes from ChatGPT subscriptions; Anthropic relies more on selling API access for AI programming and other white-collar work scenarios to enterprises.

For the public market, both types of revenue will be compared. Subscription revenue looks at user scale and retention, while enterprise API revenue focuses on customer stickiness, usage frequency, and unit economics.

In terms of valuation, Anthropic has already surpassed OpenAI.

Reuters reported that Anthropic completed a $65 billion financing round in late May, resulting in a post-financing valuation of $965 billion, already exceeding OpenAI. OpenAI's latest valuation in March of this year was $852 billion.

The speed at which Anthropic's valuation has increased has also been rapid. In February of this year, Anthropic's valuation during a $30 billion financing round was $380 billion. By late May, the valuation had more than doubled. The latest round of investors includes Blackstone, Brookfield, D1 Capital Partners, GIC, General Catalyst, and Insight Partners.

Anthropic's rapid rise earlier this year had a significant impact on software and IT stocks, as investors worried that more autonomous AI tools would change traditional business models and accelerate industry disruption.

The profitability metric is of greater concern to the market.

According to a report by The Information, Anthropic is expected to achieve approximately $559 million in operating profit in the second quarter, with an operating profit margin of about 5%.

OpenAI, on the other hand, is still in significant losses. Reports indicate that OpenAI's operating loss for the first quarter reached as high as 122%, even after excluding significant items like equity incentives. According to the report's calculations, OpenAI's operating loss for that quarter was at least $7 billion.

Cost pressures primarily stem from compute power.

OpenAI predicted earlier this year that it would consume about $25 billion in cash for the year, with AI server rental costs reaching as high as $32 billion. Additionally, OpenAI must share 20% of total revenue with Microsoft, a pact that lasts until 2030. If it achieves the previously forecasted revenue of $30 billion this year, that would mean sharing about $6 billion with Microsoft.

Anthropic is not without cost pressures either. Anthropic also needs to share revenue with cloud partners. Its revenue accounting includes all the amounts sold through other cloud service providers, part of which will ultimately be returned to the cloud partners.

Anthropic’s current profitability status also carries risks. As revenue grows rapidly, the company needs to significantly increase server resources, which could push it back into losses.

This is also a point that the public market will question: how fast is revenue growth, do compute costs rise faster or slower; how much of total revenue ultimately has to be shared with partners; and whether enterprise customers are genuinely retained or merely buoyed by short-term AI enthusiasm.

This IPO race will ultimately become a pressure test for the public market

From a timeline perspective, Anthropic has stepped the pace forward.

From a financial narrative standpoint, it has also presented a combination that is more easily understood by the public market: higher annualized revenue, higher latest valuation, and better operating profit performance at least in the short term.

However, this does not mean that the IPO outcome is already determined. A private filing does not guarantee a successful public offering or a finalized valuation. The real test will begin once the prospectus is publicly disclosed.

The public market will compare Anthropic, OpenAI, and other AI companies on the same table: revenue growth rates, profit margins, cash burn, compute spending, cloud partnership shares, customer structures, model capabilities, and commercialization paths.

Who goes public first and how the market reacts may affect the future of both companies, as well as the next phase of the AI craze: either reinforcing market confidence in the transformative power of AI, or issuing warnings about overheating in AI.

For investors, this competition is no longer just about "whose model is smarter." Now it also depends on who can turn the AI story into financial statements that the public market is willing to buy.