Anthropic's Chief Executive, Dario Amodei, recently called for stronger safeguards on AI development speed, stirring reactions across Wall Street and Silicon Valley. Yet even as Amodei finds himself immersed in intense debates over AI safety, his company is pressing forward with plans for a landmark initial public offering that could rank among the largest in history.
According to four people familiar with the matter, Anthropic now anticipates annualized revenue exceeding $100 billion this year, a sharp rise from the $65 billion annualized pace recorded as of July. Investors seeking to join the IPO are citing these surging figures to support a remarkable valuation estimate of roughly $2 trillion for the company.
Two insiders noted that the AI lab could release detailed financial documents outlining its offering to the public within the coming weeks, potentially paving the way for trading to commence as early as November. These insiders cautioned that plans remain fluid, as IPO timing consistently hinges on factors such as shifting investor sentiment and market volatility. This is especially true for a company at the forefront of an unpredictable sector, where numerous executives, including its own leadership, increasingly warn of significant risks.
The continued advancement of this mega-IPO highlights a broader paradox surrounding AI: despite widespread concerns over the technology's disruptive power, the returns promised to investors are exceptionally enticing. Going public could provide Anthropic with rapid access to capital, helping to address its substantial funding needs, particularly the enormous computing costs associated with building and operating AI models.
Investors anticipate that the company will command around 5 gigawatts of computing capacity by the end of the year, with that figure roughly doubling by the end of next year, according to three sources familiar with the data. This would place it on par with rival OpenAI. On Thursday, Anthropic hosted a customer event at its San Francisco headquarters with roughly 100 investors from venture capital firms to discuss its technology offerings, as two sources indicated.
During the event, Anthropic co-founders Jared Kaplan and Ben Mann delivered presentations alongside Boris Cherny, who leads Claude Code, and Andrej Karpathy, an OpenAI co-founder who joined Anthropic as a researcher earlier this year. According to four people briefed on the discussions, Anthropic's bankers and executives, including Amodei and Chief Financial Officer Krishna Rao, have also met with prospective IPO investors in recent weeks.
The company has held sessions with those prepared to commit substantial capital or who are already existing Anthropic investors. Some smaller investors hoping to participate in the IPO have so far been unable to secure meetings with management, these sources noted. In certain meetings, the company emphasized the positive aspects of artificial intelligence, including its research in biology and the potential to cure diseases, according to three sources familiar with the pitches.
When queried about the potential threat from open-source AI models, Anthropic downplayed the competitive risk, with one insider relaying that the company told investors only a small fraction of businesses depend on such models. Several investors also conveyed confidence that Anthropic could expand its operations even if it slowed certain AI developments, given that the majority of its revenue comes from selling Claude to enterprise clients whose adoption of the technology remains in its early stages.
Amodei's safety warnings have raised questions about potential liabilities, such as scenarios where a rogue agent triggers litigation, risks that are typically not covered by insurance. Shortly before Amodei's call for slower development, a former Anthropic researcher warned that AI has a greater than 10 percent probability of causing human extinction within the next decade, fueling further macro-level concerns. Craig Coben, a former global head of equity capital markets at Bank of America, observed that such concerns resemble political risk, with investors simply learning to accommodate and manage them.
The Anthropic IPO also prompts the question of whether Amodei can ensure the safety of advanced AI models while enduring the pressures of running a public company beholden to investors' appetite for growth. Some Anthropic investors believe that as a listed entity, the company can achieve both objectives. Altimeter Capital's Brad Gerstner expressed support on social media on September 12, stating that an Anthropic IPO would be beneficial and crucial for bringing greater transparency, scrutiny, accountability, and participation to these great American companies.
As a public company, Anthropic would offer insights into its operations, particularly its financial data. It would also distribute the wealth it generates beyond a select group of private shareholders, potentially allowing more Americans to directly share in the benefits of AI success and safety. Jesse Fried, a Harvard Law professor who studies Anthropic's governance, noted that this transparency could be limited. Anthropic could still keep certain data confidential, such as activities within its laboratories. Fried remarked that if the concern is an agent escaping from Anthropic's testing environment and taking over the world, that approach offers little help.
OpenAI is taking a different path, with CEO Sam Altman announcing last week that the company would delay its IPO until 2027. The firm had previously weighed a 2027 listing while strengthening its financial position and pursuing a $1 trillion valuation, and is currently in talks for private financing at a $1.5 trillion valuation. Edward Best, a capital markets partner at Willkie Farr & Gallagher, suggested that OpenAI's postponement could make an Anthropic IPO more attractive now, before any regulatory changes reshape the industry. Best advised that if a company can push its valuation to market, one should strike while the iron is hot.