'IPO' Anthropic Expects $2 Trillion Valuation, Warns "AI Poses Risks to Humanity"

Aiming for a Nasdaq Listing This Fall
Dedication of 80 Pages in Prospectus to Risk Factors

Last year, revenues at Anthropic, a U.S. artificial intelligence (AI) startup that is preparing for an initial public offering (IPO) this fall, surged by 12 times compared to the previous year. The market is closely watching how quickly the company can improve profitability, given its continued heavy investment in AI infrastructure.

Dario Amodei, CEO of Anthropic. Photo by Reuters Yonhap News

According to international media outlets such as MarketWatch and the Financial Times (FT) on September 29, Anthropic recently disclosed in its IPO investment prospectus, which was distributed to some investment partners, that its revenue last year reached approximately $4.6 billion (6.2463 trillion won), a 12-fold increase year-on-year. With its growth accelerating this year, second-quarter revenue registered $11.5 billion. Both adjusted operating profits for the first and second quarters this year were reported to be in the black.

This remarkable growth rate has amplified expectations for a blockbuster IPO. There is also growing speculation in the market that Anthropic could achieve a valuation exceeding $2 trillion (about 2,711 trillion won), more than double the $965 billion valuation it received when it secured investment in May. Since June, some have even expressed hopes that Anthropic could surpass the achievements of SpaceX.

However, expenses have been rising as rapidly as external growth. Last year, Anthropic dramatically increased related expenditures to secure computing resources, leading to an operating loss exceeding $8 billion. As costs for computational resources needed to train and run AI models grew, total operating expenses increased to roughly $13 billion.

Future investment commitments are significant as well. Anthropic reported its plan to spend a total of $518 billion over the next few years on cloud, computing, and infrastructure to fuel rapid growth. The company has also signed contracts to secure computing resources from partners including Google and SpaceX.

Anthropic has warned that its technology could pose "existential risks to humanity," citing possibilities such as advanced AI models manipulating or threatening humans and exhibiting unpredictable behaviors. The company reportedly dedicated 80 pages to explaining risk factors—more than the 48 pages covering its business description. Anthropic also noted its high dependency on a limited number of customers as a risk factor; last year, about one-quarter of total revenue came from just two clients.

Amid a growing number of recent AI control failures, the industry is sounding alarms about the associated risks. Dario Amodei, CEO of Anthropic, has recently issued repeated warnings about AI dangers, arguing for slowing the pace of frontier AI development and strengthening safety measures. Sam Altman, CEO of OpenAI, and Elon Musk, CEO of Tesla, have also agreed on the need for collaboration to slow AI development and enhance safety.

This content was produced with the assistance of AI translation services.