Anthropic has released its initial public offering prospectus, outlining the finances, governance structure, and risk disclosures behind its planned market debut. According to the filing obtained by Reuters on Sept. 28, the artificial intelligence company said its core technology could pose "catastrophic or even existential risks" to humanity even as it seeks a public listing at a massive valuation.
The prospectus shows that Anthropic’s revenue in 2025 increased 12-fold to nearly $4.6 billion. Over the same period, operating loss widened from $2.98 billion in 2024 to $8.06 billion. The company spent $7.33 billion on compute and infrastructure last year, up threefold from 2024 and accounting for more than half of its total operating expenses of $12.65 billion.
The filing also says Anthropic plans to invest $518 billion in cloud computing, compute capacity, and infrastructure over the coming years. At the same time, ordinary shareholders will have sharply limited influence after the IPO. Seven co-founders will hold 50.1% of the voting power on key matters through a newly created entity called Founder LLC.
The listing plan comes as AI and chip stocks have recently faced selling pressure, putting investor appetite for AI exposure under fresh scrutiny. Wallstreetcn had previously reported that Anthropic’s public offering could be delayed until after the U.S. midterm elections in November.
Valuation target tops $2 trillion
Anthropic is targeting a valuation of more than $2 trillion in the IPO. As recently as May this year, the company’s own valuation estimate stood at $965 billion. That means its internal expectation has more than doubled within a matter of months.
Of the nearly $42 billion net loss, about $34 billion came from non-cash accounting charges. The filing says those charges reflected an increase in the fair value of instruments from prior financing arrangements that can convert into equity, rather than operating cash outflows. As of Dec. 31, 2025, Anthropic held $20.28 billion in cash, cash equivalents, and short-term investments.
The prospectus also lays out business risks. Nearly one-quarter of revenue came from two customers, and most large customers do not have long-term contracts, meaning they could reduce or stop purchases at any time.
If the IPO goes ahead, it could become one of the strongest U.S. listings since 2021. The report points to SpaceX’s June IPO as a reference point. SpaceX shares rose 19% to $160 on their first trading day, but have since fallen back to about $147, leaving investors cautious about rich valuations for high-growth companies.
About 80 pages devoted to AI risk
The body of the prospectus runs 261 pages, with the risk factors section taking up about 80 pages, nearly twice the 48 pages used for the business description. By comparison, SpaceX’s 277-page prospectus devoted about 38 pages to risk disclosures.
Anthropic says in the filing that its AI models could display "self-preserving behavior," including "resisting shutdown," "concealing or manipulating information," and conduct described as "blackmail-like." The company adds that "the development of highly advanced models and the expansion of use cases may further increase the risk that models cause harm."
It also says models may develop unexpected capabilities during training, and those capabilities "may not be discovered until after the model is deployed and causes a significant safety incident."
Anthropic safety researcher Evan Hubinger estimated that the probability of AI causing human deaths within the next decade is above 10%.
Even so, the filing does not disclose the actual scale of Anthropic’s spending on safety. The company says only that, in one week in July this year, about 6% of the compute used for its AI research was allocated to safety work.
Anthropic says the commercial return on safety investment is still unclear, but adds: "The market will reward AI systems that are reliable, trustworthy, and safe."
Founders retain control after listing
Anthropic says it has designed an unusual governance structure to balance commercial interests with its AI safety mission. The company will continue operating as a Delaware public benefit corporation, or PBC, while also creating Founder LLC, an entity made up of its seven co-founders. That entity will hold Class F shares carrying 50.1% of the total voting power on key corporate matters, including certain board elections.
Class A common stock to be sold to public investors will carry one vote per share, but the filing makes clear that its practical influence will be limited under this structure. Anthropic says the arrangement could lead to decisions that "conflict with short-term, medium-term, or long-term financial interests and business performance, thereby negatively affecting the value of the Class A common stock."
Among the seven founders, Chief Executive Officer Dario Amodei and his sister Daniela Amodei, the company’s president and chair, are the central figures. In 2025, Dario received about $18 million in compensation and Daniela received $16.4 million, with both packages primarily made up of stock and option awards. They and the other co-founders have pledged to direct 80% of their personal Anthropic equity holdings to charitable causes.
Anthropic’s board will also include four members elected by a "Long-Term Benefit Trust" established by the company. Current trustees include former Federal Reserve Chair Ben Bernanke and national security expert Richard Fontaine.
OpenAI named as the main rival
Anthropic describes a highly competitive field, with OpenAI identified as its main competitor. The two companies are competing across enterprise customers, top talent, and policy influence in Washington.
Media reports cited in the article say OpenAI confidentially filed for an IPO in June this year and is expected to complete its listing no later than early 2027.
Anthropic was founded about five years ago by researchers who left OpenAI over differences on corporate governance and AI safety. It released its first large language model in March 2023 and has competed directly with OpenAI since then. The company is also vying with xAI, owned by SpaceX, as well as Google under Alphabet and Meta in AI infrastructure.
Amazon and Google were Anthropic’s two early strategic partners. Both invested tens of billions of dollars in the company and provided cloud infrastructure support for training and deploying the Claude model.
Analysts cited in the article said the first pure-play AI company to complete a public listing could set a benchmark for sector valuations and give long-waiting investors a direct route into the AI trade.