Anthropic Publishes IPO Prospectus, Disclosing Five Key Pieces of Information

Anthropic's blockbuster listing — five startling things we know so far

Image source, EPA

Image caption, Anthropic CEO Dario Amodei has become one of the key figures of the current AI era. Article Information

Artificial intelligence (AI) company Anthropic has published the documents required for its US stock market listing, a flotation the market has long awaited.

The 335-page document reveals new details about how the owner of the Claude chatbot makes money, where it spends it, and who controls it.

Anthropic also set out in detail some of the risks it believes AI technology could pose, including the possibility of taking human lives. This information is presented alongside the many benefits the technology is claimed to offer.

Here are five points worth noting in the document.

1. 'Doomsday' is an occupational hazard

Every listed company must tell potential investors how things could go wrong.

But Anthropic's "risk factors" are quite unique: it also discloses how the AI technology it is developing could lead to the extinction of all humanity.

Anthropic said: "Without appropriate safeguards, oversight and security measures to develop, deploy and operate them, sufficiently advanced AI systems could pose catastrophic or existential risks to humanity."

Mbappé caught up in 'Chinese dictator' meme storm: brand terminations, CCTV suspension

Australian prisoner loses bid to eat Vegemite, why is the 'national food' banned in prison?

Serbia turns: Vučić's 14-year rule ends, where does the China-Serbia 'ironclad friendship' go?

How does birdwatching reshape the brain and delay cognitive decline?

English text is being prepared.

But, Anthropic said, its success depends on its continued ability to "develop, improve and deploy the most advanced, most powerful and most useful AI models."

Moreover, one of the most significant risks to this ability is whether Anthropic can obtain the infrastructure it needs, including computer chips and data centers, to build, train and operate the models underpinning its technology.

Former employee of AI giant: Industry has 'genuine fear' of human extinction, the West must coordinate with China September 13, 2026

Why do AI companies want you to be afraid? May 1, 2026

China and the US lead in different AI tracks, but either could ultimately pull ahead April 9, 2026

The market widely expects the company's market value at listing to reach $2 trillion (13.34 trillion yuan; NT$64.8 trillion) or more, making it the largest stock listing in history.

But that figure will be heavily influenced by its financial performance.

Anthropic lost $80.7 billion (538.1 billion yuan; NT$2.62 trillion) in the first half of this year, twice its full-year loss in 2025. The company's overall revenue for those six months was $16.2 billion.

Although revenue is rising rapidly, it still has a long way to go to break even.

Troy Hooper, co-head of equity capital markets at Mergermarket, stressed that Anthropic's revenue growth is "extraordinary," but added that investors are less forgiving of tech companies promising huge future returns than they were a few months ago.

In addition, Anthropic's debt position is unclear, because it has substantial accounting charges related to certain forms of investment.

Hooper added: "As a result, investors do not yet have a clear picture of Anthropic's potential profitability, nor of whether its cash generation rate is sufficient to support business expansion." Skip YouTube post, 1 Allow Google YouTube content

This article contains content provided by Google YouTube. Because these contents use technologies such as cookies, we seek your consent before loading any content. You may wish to read Google YouTube's cookie policy and privacy policy before giving permission. To read the above content, please click "Accept and continue." Note: BBC is not responsible for the content of external websites. YouTube content may contain ads

At $131.6 billion, almost all of its cash is defined as "working capital," meaning it can be used to pay for day-to-day operations.

At present, Anthropic has more cash on hand than Meta and Google combined.

However, as it fiercely competes with other AI giants for top talent and computing power, Anthropic's costs may rise. Renting data centers alone has cost the company $321.4 billion so far this year.

Hooper analyzed that such massive costs are expected to increase further. Even with this amount of cash, it is not enough to sustain its business growth.

He said anyone investing in Anthropic "will be paying for future dominance, not current economics."

4. Anthropic depends on a small number of large customers

Anthropic believes there are countless ways to sell its technology to customers. A large part of the company's current business comes from just 250 major enterprise customers, which have paid at least $10 million over the past year.

An even smaller number — just five enterprise customers — accounts for a quarter of its current business.

Michael Monaghan, portfolio manager at Founder ETF, noted that "many of Anthropic's top customers do not have long-term contracts."

This business concentration leaves Anthropic dependent on a small group of paying customers to support growth, with no guarantee this will continue. Anthropic also has corporate partnerships with companies such as Google and Amazon, both of which are simultaneously its investors "and competitors."

On this, Monaghan believes this could create potential business conflicts in the future.

Anthropic says its overall enterprise customer base is growing rapidly, from about 1,500 last year to nearly 12,000 this year. Part of the company's growth strategy is to offer Claude tools for free and then work to "convert these users into paying customers over time."

The company also says this strategy has made Claude "the fastest-adopted technology in enterprise history." Skip YouTube post, 2 Allow Google YouTube content

This article contains content provided by Google YouTube. Because these contents use technologies such as cookies, we seek your consent before loading any content. You may wish to read Google YouTube's cookie policy and privacy policy before giving permission. To read the above content, please click "Accept and continue." Note: BBC is not responsible for the content of external websites. YouTube content may contain ads

End of YouTube post, 2

5. AI is expected to bring 'extraordinary' benefits to the world

"We will not pursue projects solely based on commercial opportunity or research novelty; our efforts must advance our mission."

These assumed benefits include advances in healthcare, such as developing customized treatments for diseases and improving disease detection, tailored education and legal advice, and more efficient power grids and batteries.

Furthermore, decades of scientific progress could be compressed into a few years, accelerating drug discovery, disease detection and the delivery of personalized therapies.

Anthropic says AI can provide personalized education as well as expert-level financial and legal advice, advance power grid and battery design to improve efficiency and reliability, and help the "developing world" build more technological capacity.

This article was originally written in English. We used artificial intelligence to assist with the translation, and a BBC journalist reviewed it before publication. Learn more about how we use artificial intelligence.