SAN FRANCISCO (AP) — Anthropic, a major player in artificial intelligence, has submitted paperwork that may lead to the largest initial public offering (IPO) in history. This development comes amidst concerns from its CEO, Dario Amodei, about the potential existential risks posed by advanced AI.
The IPO is expected to position the San Francisco firm for a launch on Wall Street next month, possibly surpassing the record $86 billion raised by SpaceX in June. Should this IPO occur, it could also allow Anthropic to outpace OpenAI in their respective public offerings, marking a significant moment for both AI companies.
Just weeks before this filing, Amodei expressed in a written piece the need to slow the progression of highly advanced AI technologies. An IPO could furnish Anthropic with the resources to enhance its growth, though the exact funds it will accrue, along with its final valuation and share price, remains unclear. Nevertheless, even if the figures are lower than expected, this could still elevate the young startup into a position among the globe’s most valuable companies, potentially enriching Amodei and his co-founder sister, Daniela, by billions.
The extensive filing also reveals that Anthropic’s revenues are on the rise; however, like many tech firms going public, it has not yet recorded a profit. In the first half of 2026, the company, which created the chatbot Claude, reported a staggering net loss of $80.7 billion, an increase from a $13 billion loss in the same timeframe the previous year.
Despite these financial setbacks, private investors valued Anthropic at approximately $965 billion earlier this year, and forecasts suggest it could balloon to between $1.5 trillion and $2 trillion. If it reaches the $1.5 trillion mark, it would rank alongside Tesla as one of the most valuable public companies.
Reports indicate that the IPO could potentially raise up to $100 billion, smashing the current record set by SpaceX earlier this year.
In their regulatory filing, Anthropic announced plans to list its shares on the Nasdaq Global Select Market under the ticker symbol “ANT.”
Risks abound, yet substantial gains for founders
Anthropic indicated it may allocate some of the IPO funds for acquiring other companies and technology and bolstering its infrastructure to better compete against rivals.
Founded in 2021 by former leaders from OpenAI, Anthropic competes directly with that organization. The rivalry is intense, especially between Amodei and OpenAI CEO Sam Altman, who are seen as adversaries. But Anthropic faces significant competition from other giants like Google, newer startups in the U.S. and China, and Musk’s SpaceX, which has its own AI initiatives. Notably, these companies, including Anthropic, have been incurring substantial losses due to the high costs of AI development, leading to concerns about a potential economic bubble.
The anticipated IPO would not only help alleviate any shortfalls in cash flow but would also significantly enrich Anthropic’s seven founders, including Amodei and Daniela. Collectively, they hold nearly 500 million shares, but specifics on their ownership after the IPO are not detailed in the offering documentation.
Dario, 43, and Daniela, 39, who also serves as Anthropic’s president, have been estimated by Forbes to each have a net worth of $15.5 billion, ranking them as the 176th and 177th richest individuals globally.
The filing has also highlighted various risks facing Anthropic, such as retaining important staff, navigating regulatory challenges, and ensuring sufficient computing power for its AI systems.
Perhaps most concerning, the document emphasized that its AI models might “pose catastrophic or existential risks to humanity.”
Calls for a slowdown in AI development
This IPO comes during a time of heightened anxiety regarding AI technologies within and outside the industry. Reports persist about increasingly powerful AI systems, capable of solving complex problems but also possessing the potential to behave unpredictably and cause harm.
In September, Amodei advocated for a deceleration in AI development to allow safety protocols to catch up. He cautioned that, without this pause, AI systems could become capable of unleashing agents that might control the internet within as short a timeframe as six to twelve months.
Amodei proposed a strategy on his website aimed at increasing oversight within the industry, noting that Anthropic is already implementing some aspects independently. However, he expressed that broader coordination across the industry and with global governments, including authoritarian regimes, would be essential.
“I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong,” Amodei wrote.
Tension with the U.S. government
Anthropic has sought to distinguish itself from other AI firms, positioning itself as the “good guys” by prioritizing safety and responsible innovation over unchecked growth. Some of this attitude stems from its conflict with the U.S. government, which began in February when President Trump and Defense Secretary Pete Hegseth alleged that Anthropic posed a national security threat, branding it as a supply chain risk.
In response, Anthropic launched a lawsuit against the Pentagon in March, claiming the government was retaliating for its refusal to permit military use of its technology unrestrictedly. This legal battle has escalated the public debate surrounding the applications of AI in combat and surveillance.
OpenAI, creator of ChatGPT, arranged its own deal with the Pentagon just hours after the government took action against Anthropic over its stance. Anthropic has also initiated a separate legal challenge regarding another Pentagon rule aimed at designating the company as a supply chain risk.
A judge in California ruled in favor of Anthropic in the first case, while a Washington D.C. judge supported the government’s position in the second. Both parties are currently appealing those decisions.
