Anthropic is going public. The numbers – and the risks – are staggering
Updated . First published at
One of the world’s hottest artificial intelligence companies is trying to raise tens of billions of dollars in a potentially record stock market debut. It’s also warning that it might wipe out humanity and posting losses in the tens of billions of dollars.
The phrase “catastrophic or existential risks to humanity” stands out in a 335-page financial document laying out concepts such as dividend policy, governance structure and executive compensation.
Anthropic says its AI models could pose a “catastrophic or existential risk to humanity” and can “resist shutdown,” according to an initial public offering prospectus. Patrick Sison
When listing on the stock market, companies have to disclose risks but they’re commonly focused on potential problems such as competitors, government regulations or economic downturns.
Rarely do they go as far as Anthropic, which dedicated 93 pages to risk factors in its Monday morning (just after midnight on Tuesday AEDT) filing.
Those risks included some pretty standard stuff: retaining key staff, navigating regulatory roadblocks and securing enough compute power to run its AI systems.
But the company behind the Claude chatbot also went deep on some of the key concerns that have generated opposition to the breakneck AI race in recent years.
The initial public offering (IPO) filing details major categories of risk from artificial intelligence: misuse, such as being used to design and implement bioweapons or cyberattacks; loss of human control; authoritarian abuse, such as its potential use to create surveillance states; and economic disruption.
“Concerns have been raised both within our organisation and by other researchers, regulators, policymakers, and industry participants that sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity,” the company warned.
Anthropic says AI “could rapidly accelerate scientific innovation, greatly increase economic growth rates, bring cures for major diseases”. Anthropic
That would particularly be the case if the systems “were misused, replicated, weaponised, or allowed to operate with a high degree of autonomy”.
As AI was increasingly used to build new AI, those risks would increase, Anthropic said.
The company warned its model safety testing had already produced “incomplete or unreliable results” with models “behaving differently when they detect or infer that they are being evaluated”.
That all refers to existing technology. More advanced models could pose risks relating to “autonomy, manipulation, and the simulation of aspects of human psychology”, the filing stated.
That could include "self-preserving behaviours" such as refusing to be shut down, concealing or manipulative information and ‘behaviour resembling blackmail".
"Our increasing use of AI systems to accelerate our own research and development may cause the capabilities of our models to advance more rapidly than our ability to develop, validate, and implement corresponding safety, alignment, and security measures," it said.
As AI is increasingly used to build new AI, the risks would increase, Anthropic said. Anthropic
Call for AI development slowdown
The warnings are not out of character for Anthropic. Chief executive Dario Amodei last month called for a global slowdown in AI development amid warnings the technology could be capable of letting loose a swarm of agents that could take over the internet within six to 12 months.
“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,” he wrote.
Amodei proposed a range of measures for the industry, which he has committed to himself, which he acknowledged would not be easy.
“But I believe we owe it to humanity to try,” he said.
Anthropic CEO Dario Amodei is moving ahead with its initial public offering. Getty Images
On the upside, the filing argues “AI is one of the most consequential technologies in human history”.
“If managed well, it could rapidly accelerate scientific innovation, greatly increase economic growth rates, bring cures for major diseases within reach, and solve problems that have resisted generations of human effort,” Amodei and other company leaders wrote in an opening letter to prospective shareholders.
Show me the money
The filing revealed staggering details about the finances of the San Francisco-based company.
It reported a net loss of $US80.7 billion ($115 billion) in the first half of 2026. That compares to a loss of $US13 billion ($18.5 billion) in the same period a year earlier.
Yet its value has skyrocketed, from $US965 billion by private investors earlier this year to current reported figures of $US1.5 trillion ($2.1 trillion) or even $US2 trillion ($2.85 trillion).
But there were positive signs too. Revenue grew 1300 per cent year-on-year in the first six months of 2026 to $US16.3 billion ($23.2 billion).
Read more
SpaceX and Tesla could soon merge in a $5 trillion move. What would that mean for Australia?
And only a fraction of the loss was actual spending – such as on AI infrastructure and computing – during the first half of the year, with $US72 billion attributed to “change in fair value of convertible notes, derivative liabilities, and SAFEs.”
Anthropic’s IPO will test Wall Street’s appetite for AI. The general public, policymakers and key industry figures are raising questions about direct risks as the rapidly advancing research and data centre buildout finds increasing opposition in the western world.
"Anthropic is entering an IPO market that is far more challenging than the summer. Investors were not questioning valuations the same way a few months ago," said Matt Kennedy, senior strategist at Renaissance Capital, a provider of IPO-focused research and ETFs.
In the filing, Anthropic notes it has already signed leases for data centre capacity in Australia and Canada, beyond its US operations.
A protest outside Utah’s State Capitol to oppose the construction of a data centre in Salt Lake City. Getty Images
Anthropic is listing on the Nasdaq Global Select Market under the ticker symbol "ANT”, and could go public as early as November 16.
The offering could bring in as much as $US100 billion, according to reports, which would eclipse the amount raised by SpaceX in June.
It would also put Anthropic ahead of OpenAI in going public in what is expected to be a pair of blockbuster stock debuts for the rival AI labs.
Amodei, who was once a leader at OpenAI, is widely seen as a bitter rival to the ChatGPT maker’s CEO, Sam Altman. The companies are locked in a furious battle for the future of AI along with companies such as Google and SpaceX’s AI division.
"Anthropic's IPO will establish a benchmark for valuing leading AI labs, influencing both upcoming listings and confidence in the companies funding the AI infrastructure build-out," IPOX research associate Lukas Muehlbauer said.
"Its success would not necessarily benefit every AI business equally.”
Dario Amodei, 43, and 39-year-old sister Daniela, who is Anthropic's president, are worth $US15.5 billion each, according to Forbes, which ranks them as the 176th and 177th richest in the world.
The listing would potentially boost their wealth by billions of dollars.
- Reported with wires
Share a tip-off, video or photo with us
