Lauren Hirsch, Sri Muppidi, Erin Griffith & Mike Isaac

Anthropic CEO Dario Amodei roiled Wall Street and Silicon Valley recently when he called for more guardrails on the pace of artificial intelligence development.

But even as Amodei became embroiled in a contentious AI safety debate, his company — which makes the Claude chatbot — continued to prepare for a blockbuster initial public offering that may shape up to be the biggest listing ever.

The company is expected to reach more than $100 billion in annualised revenue by the end of this year, according to four people familiar with the matter. That’s up from $65 billion in annualised revenue as of July. Investors seeking to get a piece of the IPO are using those soaring numbers to justify Anthropic’s staggering potential valuation of $2 trillion.

The AI lab could publicly release financial documents detailing its offering as soon as the coming weeks, which could put it on track for its shares to start trading as soon as November, said two people with knowledge of the situation who, like others quoted in this article, spoke on condition of anonymity to discuss the confidential plans.

The people stressed that the plans could still change. The timing of IPOs is always subject to change because of factors including shifting investor sentiment and market volatility.

That is particularly the case for a company that is a leading force in an unpredictable industry in which many executives, including its own leader, are increasingly warning about sizable risks.

That the giant IPO is ploughing ahead highlights a broader tension surrounding AI: Despite worldly concerns about the technology’s disruptive powers, the returns these companies promise to investors are enormously tantalising.

Going public could give Anthropic quick access to capital to help fund its large financial needs, in particular the substantial computing costs necessary to build and run its AI models.

Investors expect the company to have about 5 gigawatts’ worth of compute — the industry term for computing power — available to use by the end of the year, and would roughly double that capacity by the end of next year, according to three people familiar with the figures. That would put it on par with its rival OpenAI.

On Thursday, Anthropic hosted about 100 investors from venture capital firms at its San Francisco headquarters to discuss its technology offerings in a customer event, according to two people familiar with the event.

Jared Kaplan and Ben Mann, two of Anthropic’s co-founders, and Boris Cherny, the head of Claude Code, made presentations alongside Andrej Karpathy, a researcher and co-founder of OpenAI who joined Anthropic this year, the people said.

Anthropic’s bankers and its executives, including Amodei and its chief financial officer, Krishna Rao, have also met with potential investors in the IPO in recent weeks. The company has taken meetings with those who are willing to write large checks or are preexisting investors in Anthropic, four people familiar with the discussions said.

Some smaller investors are hoping to invest at the IPO but haven’t yet managed to obtain an audience with management, these people say.

In some of these meetings, the company has pointed to the positive aspects of artificial intelligence, including its research into biology and potential to cure diseases, according to three people familiar with the pitch.

When asked about the potential threat from China offering cheaper open-weight AI models — which publicly provide the underlying calculations of their systems — Anthropic has downplayed the competition. The company has told investors that only a small portion of businesses rely on such models, one person familiar said.

Several investors also said they believe Anthropic can expand its business even if it slows certain AI development, as Amodei recently called for. That’s because the company generates a majority of its revenue selling Claude to business customers, who are still early in their adoption of the technology.

Still, bankers and lawyers say Anthropic is wading through largely uncharted territory with its IPO.

Amodei’s safety warnings raise questions about potential liabilities for the company if, for example, a rogue agent leads to litigation. (These risks are not generally insured.) Amodei called for the development slowdown shortly after a former Anthropic researcher warned the chance that AI could make humans extinct within a decade was greater than 10%, creating more cosmic concerns.

“My experience is, it’s a bit like political risk; investors just kind of get their arms around it and just deal with it,” said Craig Coben, a former global head of equity capital markets at Bank of America.

Anthropic’s IPO also raises questions about whether Amodei can ensure the safety of advanced AI models and withstand the pressures of running a publicly traded company that is subject to the whims of public investors clamouring for growth.

Some Anthropic investors argue that as a public company, it can achieve those goals simultaneously.

“Anthropic will IPO,” wrote Brad Gerstner of Altimeter Capital on social media on Sept. 12. He added, “And it's beneficial/critical that we bring even more transparency, scrutiny, accountability, & participation to these grt American companies!”

OpenAI is taking a different approach. Last week, its CEO, Sam Altman, said the company was putting off its public listing until 2027. The company had been weighing a 2027 listing as it shores up its finances and tries for a $1 trillion valuation. (It is in talks to privately raise funds at a $1.5 trillion valuation.)

(The New York Times has sued OpenAI and Microsoft, claiming copyright infringement of news content related to AI systems. The two companies have denied those claims.)