Anthropic (ANT.US) revealed nearly 100 pages of risk before sprinting to IPO: AI lost control or caused irreversible consequences, difficult insurance coverage, and quarterly growth rate dropped to about 45%

Zhitong Finance App learned that Anthropic PBC (ANT.US) submitted nearly 100 pages of risk factors in its landmark initial public offering (IPO) application and made an unusually detailed set of disclosures on various issues that may arise in its artificial intelligence (AI) business. This risk list covers all aspects: declining revenue growth, regulatory restrictions in various countries, and even “posing disastrous, survival-level risks to humans.” A long list of risks reflects the unique challenges that open market investors face when evaluating the opportunities and risks of Anthropic technology.

The submission of listing documents comes at a sensitive window where the capitalization process in the AI industry is rapidly heating up and market sentiment tends to diverge. On October 26, Anthropic officially disclosed the S-1 registration documents to the US Securities and Exchange Commission. It plans to use “ANT” as the code to enter the NASDAQ global selection market. Goldman Sachs, Morgan Stanley, and J.P. Morgan Chase will take the lead in underwriting, with the goal of completing the listing in mid-November.

In the H round of financing in May, Anthropic used 65 billion US dollars to push the post-investment valuation to 965 billion US dollars; according to various media reports, the target valuation of this IPO may fall in the range of 1.5 trillion US dollars to 2 trillion US dollars, and the fund-raising scale may reach 100 billion US dollars. If successful, it will set a new $86.25 billion fundraising record set by SpaceX (SPCX.US) in June of this year.

The financial data showed two distinct aspects. On the growth side, revenue in the first half of 2026 soared to US$16.286 billion, about 14 times that of the same period last year, far exceeding US$4.585 billion in just half a year; more than 11,900 customers paid more than US$100,000 a year, with a net revenue retention rate of about 900%, and the average second-year expenditure of old customers was 9 times that of the first year.

On the loss side, net book loss reached US$80.748 billion during the same period, but about US$72.4 billion of this amount came from changes in the fair value of financial instruments such as convertible bonds and SAFE — the more rapidly the company's valuation rose, the greater this portion of the book loss, and was not directly related to current operations.

After excluding this factor, the operating loss for the first half of the year narrowed sharply to US$693 million from US$2,135 million in the same period last year. The adjusted operating profit was about US$1.6 billion, and the net operating cash flow inflow was US$3,929 million, which was corrected for the first time.

Anthropic's launch attempt is not an exception; it is part of the 2026 AI giants' collective sprint to the open market. OpenAI had previously secretly submitted an IPO application, with a valuation of more than 850 billion US dollars; after completing the listing in June, SpaceX became one of the listed companies with the highest market capitalization in the world at the time with a valuation of about 1.77 trillion US dollars.

However, the secondary market's attitude towards high-premium AI projects is becoming more cautious: in the first quarter of 2026, non-AI companies in the world completed listing, reflecting investors' stricter scrutiny of valuation bubbles and profit paths.

Anthropic's prospectus itself conveys this tension — a steep curve in revenue and a sharp expansion in rigid computing power promises: as of mid-2026, the company has signed a computing power and infrastructure agreement with a minimum future payment of approximately $417 billion, and potential expenses of 125 billion to 150 billion US dollars to be finalized, with cumulative financing of $141.1 billion, cash on account, and short-term investments of about $131.7 billion.

Here are the key points to focus on in the prospectus:

Catastrophic hazards

Anthropic has highlighted the potential risk that its AI may “act in ways that are not anticipated, off target, or autonomously.” The company said its technology could pose a threat to humans if developed without adequate safeguards and supervision, “particularly when such systems are misused, copied, weaponized, or allowed to operate in a highly autonomous manner.”

The company warned that as the model's autonomy continues to improve, the associated risks will increase simultaneously. Such capabilities may expand the harmful space and cause consequences that are difficult to reverse or even irreversible, such as unauthorized submission of code changes, deletion of tampered data, and automatic initiation of financial transactions.

The application documents warn that manual reviewers may not be able to stop harmful behavior in a timely manner, and if the AI system learns to hide targets and bypass surveillance methods, Anthropic's technical monitoring mechanism will also fail.

As Anthropic applies AI to scientific work, the company also warned that its technology could be used to aid the “release of hazardous substances,” and stated that “such consequences could be immediate and irreversible.”

Insurance pressure problem

To make matters worse, Anthropic admits that existing insurance products are unlikely to be sufficient to cover the legal liabilities derived from actual damage caused by AI technology. Once a major AI disaster occurs, the coverage gap will be particularly prominent.

“No insurer, or perhaps the entire insurance industry combined, has sufficient capacity or expertise to cover insurance that matches this risk exposure,” the company wrote.

Revenue growth is slowing

Anthropic achieved over 1000% year-on-year revenue growth in 2025, an impressive increase, but the company notes it “doesn't expect” future revenue to grow at a comparable rate. The company cites the maturity of the AI market, increased competition, and security considerations that may limit the release of its models as reasons for the possible slowdown in revenue. The company also said that it was “difficult” to accurately forecast revenue in the past and may face this difficulty again in the future because the AI market is still in its early stages.

According to previous reports, the company's growth slowed to about 45% from the second quarter to the third quarter of 2026, compared to more than 140% in the previous quarter.

Computing power supply risk

Another major risk Anthropic points out is that the supply of computing power may be insufficient, and at the same time, the price of computing power may rise sharply. The application documents state that some of Anthropic's current cloud service provider partners are themselves competitors, creating a situation where these vendors “may not always fully meet” the long-term interests of “this Claude developer.” Anthropic has signed important agreements to use chips, computer room resources, and supporting infrastructure from companies such as Google, Amazon (AMZN.US), and Elon Musk's SpaceXAI under Alphabet Inc. (GOOGL.US), and these companies are also developing their own AI models.

At the same time, Anthropic also faces another risk: if computing power is excessively purchased based on overestimated demand, computing power will be idle and funds will be wasted. The company said, “Our actual computing power demand is affected by a large number of variables that are difficult to predict,” including the progress of research and development, the computing power consumption intensity of training programs, etc., and the uncertainty is very high.

Increased regulatory scrutiny

Anthropic suggests that governments around the world may act to limit their technology, thereby jeopardizing the core engine of their business. The company said, “Regulators in various countries have already imposed restrictions on our model, and access may be tightened further in the future.”

In June of this year, due to safety considerations, the US government imposed export controls on Anthropic's Fable 5 model. Relations between the company and the Trump administration have always been strained; in March of this year, Secretary of Defense Pete Hugseth listed Anthropic as a supply chain risk subject on the grounds of surveillance and autonomous weapons disputes, and then Anthropic sued the US Department of Defense.

Data centers face public boycotts

Anthropic needs large-scale data centers to train and operate the model, but public opposition continues to rise, and this boycott has become more organized and collaborative, and has become a hindrance to business. Most of the company's computer rooms are located in the US. “Any core facility is shut down or significantly disrupted, and it is difficult to quickly make up for the remaining and subsequent new infrastructure.”

Anthropic has voluntarily promised to cover some of the external costs associated with data center energy consumption, such as power grid upgrades; however, the company also warned that if such requirements become common industry standards, costs will continue to rise.

The threat of cyber attacks

Cybersecurity is not only Anthropic's target business circuit, but also a major threat to its own operations.

The AI company says it faces a number of security risks, including hacking the company or using Anthropic's tools to hack others — the latter has actually happened. The company also anticipates increasingly being targeted by state-sponsored cyber espionage — particularly operations targeting its most advanced AI models.

TSMC dependency risk

Anthropic notes that the vast majority of the world's advanced chips are still being produced in Taiwan Province of China, mainly manufactured by TSM.US. This poses significant risks, including supply chain disruptions due to geographical tension.

Corporate culture and talent loss

Over time, Anthropic has created a corporate culture that describes itself as a “low self, realistic environment.” The company also said that many of its key employees are “strongly driven by our mission and the opportunity to solve specific problems in the AI field.”

Anthropic warns that as the company needs to “scale up and mature as an organization,” key employees may leave their jobs to find other opportunities they think are better suited to their own interests. Simply put, Anthropic is preparing for the pains of growing up ahead.