Artificial intelligence (AI) company Anthropic has just published the documents required for its highly anticipated debut on the US stock market.
The 335-page document contains new details about how the owner of the chatbot Claude makes money, what this money is used for and who controls the company.
Anthropic also outlines some of the risks that the company believes AI technology poses – including the possibility of human death. This information is presented alongside the many benefits that the technology claims to have.
Here are five key points from the documentation that are worth highlighting.
1. The end of the world – an occupational hazard
All listed companies must inform potential investors about what could go wrong.
Anthropic’s ”risk factors” are unique; the company mentions, among other things, that the AI technology it is developing could lead to the end of humanity.
“Sufficiently advanced AI systems can – if not developed, implemented and operated with appropriate safeguards, oversight and security procedures – pose catastrophic or existential risks to humanity,” writes Anthropic.
The company notes that less than a year ago, a hacking group – which Anthropic suspects is backed by China – “manipulated” Claude to successfully infiltrate a variety of organizations, including large technology companies, financial institutions, chemical companies and government agencies.
At the same time, Anthropic emphasizes that the company’s success depends on the ability to constantly ”develop, improve, and implement the most advanced, capable, and useful AI models.”
One of the biggest risks to this capability is access to the necessary infrastructure – such as computer chips and data centers – required to build, train, and run the models that form the foundation of the technology.
2. The company is bleeding money
Anthropic has not yet set a target price for the share ahead of the IPO.
There is widespread expectation that the company will be valued at $2 trillion or more upon its debut, making it the largest IPO in history. However, the final valuation will be heavily influenced by the company’s financial performance.
In the first half of this year, Anthropic made a loss of $80.7 billion (£60 billion) – a figure that is double the total loss for the whole of 2025. The company’s total revenue was $16.2 billion in the six-month period. Revenue is growing rapidly, but the company has a long way to go before breaking even.
Troy Hooper, co-head of equity capital markets at Mergermarket, describes Anthropic’s revenue growth as “exceptional.” However, he adds that investors today are “less forgiving” of tech companies promising huge future returns than they were just a few months ago.
Anthropic’s debt situation is also unclear, as the company has extensive accounting costs linked to certain types of investments.
“Therefore, investors do not yet have a clear picture of Anthropic’s underlying profitability, or whether the company’s cash flow generation is sufficient to support the business as it scales up,” Hooper continues.
3. But luckily there is plenty of cash.
Despite the company being at a loss, Anthropic has large amounts of cash available.
Of the $131.6 billion in cash on hand, almost all of it is classified as “working capital,” meaning the money can be used to finance day-to-day operations.
Currently, Anthropic has more cash available than Meta and Google combined.
Anthropic’s costs may rise as the company faces stiff competition from other AI giants – both in terms of recruiting top talent and securing computing power.
Data center rent alone is expected to cost the company $321.4 billion, based on contracts signed so far this year.
According to Hooper, these large costs are expected to increase further. Even with such large cash reserves, it will not be enough to maintain the growth rate.
Investors in Anthropic “are paying for future dominance, not for current economic reality,” he says.
4. Anthropic is dependent on a few large customers.
Anthropic sees unlimited opportunities to sell its technology to customers.
A large portion of the current business comes from just 250 large corporate customers, who the company says have paid at least $10 million each in the past year.
An even smaller group – just five corporate clients – accounts for a quarter of the current business. Michael Monaghan, portfolio manager for Founder ETF, pointed out that “many of Anthropic’s largest clients do not have long-term contracts.”
The high concentration of its customer base means that Anthropic is dependent on a small percentage of its paying customers to drive growth – with no guarantee that this will continue.
Anthropic also has business relationships with players such as Google and Amazon. Both of these are investors in the company – but also competitors – which could create potential business conflicts in the future, Monaghan noted.
According to Anthropic, the total base of corporate customers has grown rapidly: from about 1,500 customers last year to almost 12,000 this year.
Part of the growth strategy is to offer tools based on Claude for free, then “convert these users into paying customers over time.”
This strategy has led to Claude having “the fastest spread among companies of any technology in history,” the company claims.
5. AI promises ”extraordinary” benefits for the world
Despite all the risks Anthropic has said AI poses, they also reiterated their claim that AI will produce an ”extraordinary” amount of good things.
”Achieving our mission means the world reaping the enormous benefits of frontline AI while mitigating the most significant risks,” the company said.
That mission is defined as: ”Responsible development and maintenance of advanced AI for the long-term benefit of humanity.”
”We do not pursue projects based solely on commercial opportunities or research news; our efforts must advance our mission.”
Such hypothetical benefits include advances in healthcare, such as the development of tailored treatments for diseases and improved disease detection, tailored education and legal advice, and the efficiency of electricity grids and batteries.
Anthropic said AI could one day provide individually tailored training and expert advice in financial and legal advice, develop power grids and battery designs for greater efficiency and reliability, and help ”developing countries” build more technological capacity.
”Ultimately, we believe that frontline AI will help drive global economic growth,” the company said.