Anthropic (ANT.US) discloses nearly 100 pages of risks ahead of its IPO sprint: AI loss of control could lead to irreversible consequences, insurance may be unable to provide a backstop, and quarterly growth has slowed to about 45%

Anthropic PBC (ANT.US) filed nearly 100 pages of risk factors in its landmark initial public offering (IPO) application, making an unusually detailed set of disclosures about the various problems that could arise in its artificial intelligence (AI) business. The risk list covers everything from declining revenue growth and regulatory constraints in various countries to "catastrophic, existential risks to humanity." The long list of risk items reflects the unique difficulties public-market investors face in assessing both the technological opportunities and the dangers of Anthropic. The filing comes at a sensitive moment when capital-market activity in the AI industry is heating up sharply while market sentiment is becoming polarized. On October 26, Anthropic formally publicly filed its S-1 registration statement with the U.S. Securities and Exchange Commission, planning to list on the Nasdaq Global Select Market under the ticker "ANT," with Goldman Sachs Group, Inc., Morgan Stanley and JPMorgan leading the underwriting, targeting a mid-November listing. In its May Series H financing, Anthropic raised US$65 billion, pushing its post-money valuation to US$965 billion; according to multiple media reports, the target valuation for this IPO could fall in the range of US$1.5 trillion to US$2 trillion, with the offering size potentially reaching US$100 billion, which if completed would break the US$86.25 billion fundraising record set by SpaceX (SPCX.US) in June this year. The financial data present a striking two-sided picture. On the growth side, revenue surged to US$16.286 billion in the first half of 2026, about 14 times the year-earlier period, and in just half a year far exceeded full-year 2025 revenue of US$4.585 billion; customers paying more than US$100,000 annually exceeded 11,900, net revenue retention was about 900%, and existing customers' spending in their second year averaged 9 times that of their first year. On the loss side, the reported net loss for the same period was as high as US$80.748 billion, but about US$72.4 billion of that came from fair value changes in financial instruments such as convertible bonds and SAFEsthe faster the company's valuation rises, the larger this book loss becomes, and it is not directly related to current operations. Excluding that factor, the first-half operating loss narrowed sharply to US$693 million from US$2.135 billion a year earlier, adjusted operating profit was about US$1.6 billion, and net operating cash flow was an inflow of US$3.929 billion, turning positive for the first time. Anthropic's listing attempt is not an isolated case but part of a collective sprint by AI giants to the public markets in 2026. OpenAI had previously filed confidentially for an IPO at a valuation of more than US$850 billion; after completing its listing in June, SpaceX became one of the world's most valuable listed companies at the time with a valuation of about US$1.77 trillion. However, attitudes in the secondary market toward high-premium AI projects are becoming more cautious: in the first quarter of 2026, no AI company completed a listing globally, reflecting stricter investor scrutiny of valuation bubbles and paths to profitability. Anthropic's prospectus itself conveys this tensionon one side, a steep revenue curve; on the other, a sharp expansion in rigid computing commitments: as of mid-2026, the company had signed computing and infrastructure agreements with future minimum payments of about US$417 billion, with another US$125 billion to US$150 billion in potential spending tied to terms still to be finalized, while it had raised a cumulative US$141.1 billion and held about US$131.7 billion in cash and short-term investments. The following are the key points in the prospectus worth close attention: Catastrophic harm Anthropic highlighted the potential risk that its AI could "act in unintended, goal-divergent or autonomous ways." The company said that if its technology is developed without adequate safeguards and oversight, it could pose a threat to humanity, "particularly when such systems are misused, replicated, weaponized, or allowed to operate in highly autonomous ways." The company warned that as models' autonomous capabilities continue to increase, the associated risks will amplify accordingly. Such capabilities could expand the scope for harm and trigger consequences that are difficult or even impossible to reverse, such as unauthorized code changes, deletion or tampering with data, and automatically initiating financial transactions. The filing warns that human reviewers may not be able to stop harmful behavior in time, and that if AI systems learn to conceal objectives and circumvent monitoring methods, Anthropic's technical monitoring mechanisms could also fail. As Anthropic applies AI to scientific work, the company also warned that its technology could be used to assist in "the release of dangerous substances," noting that "such consequences could be immediate and irreversible." Insurance under pressure Compounding the problem, Anthropic acknowledged that existing insurance products are likely insufficient to cover legal liabilities arising from real-world harm caused by AI technology, and that the protection gap would be especially pronounced in the event of a major AI catastrophe. The company wrote: "No single insurer, and possibly the insurance industry as a whole combined, has sufficient capacity or expertise to underwrite insurance matching exposures of this kind." Revenue growth slowing Anthropic achieved year-over-year revenue growth of more than 1,000% in 2025, an astonishing pace, but the company noted that it "does not expect" future revenue to grow at a comparable rate. It cited the maturing of the AI market, intensifying competition and safety considerations that could limit model releases as reasons revenue may slow. The company also said it has in the past "struggled" to accurately forecast revenue and may face that difficulty again, because the AI market is still at an early stage. According to prior reports, the company's growth slowed to about 45% from the second quarter to the third quarter of 2026, after growth of more than 140% in the previous quarter. Computing supply risk Another major risk Anthropic identified is that computing supply may be insufficient while computing prices could rise sharply. The filing notes that some of Anthropic's current cloud-service partners are themselves competitors, creating a situation in which these suppliers "may not always fully align" with the Claude developer's "long-term interests." Anthropic has signed significant agreements to use chips, data-center capacity and supporting infrastructure from Alphabet Inc. (GOOGL.US)'s Alphabet Inc. Class C, Amazon.com, Inc. (AMZN.US) and Elon Musk's SpaceXAI, among others, and these companies are also developing their own AI models. At the same time, Anthropic faces another risk: if it over-procures computing based on overestimated demand, capacity could sit idle and capital could be wasted. The company said: "Our actual computing needs are affected by a large number of hard-to-predict variables," including R&D progress and the computing intensity of training plans, with high uncertainty. Regulatory scrutiny intensifying Anthropic cautioned that governments around the world may take action to restrict its technology, jeopardizing the core engine of its business. The company said, "Regulators in various countries have already imposed restrictions on our models, and may further tighten access in the future." In June this year, the U.S. government imposed export controls on Anthropic's Fable 5 model for safety reasons. The company's relationship with the Trump administration has been tense; in March this year, Defense Secretary Pete Hegseth designated Anthropic as a supply-chain risk entity citing surveillance and autonomous-weapons disputes, and Anthropic subsequently sued the U.S. Department of Defense. Data centers face public resistance To train and run its models, Anthropic needs large-scale data centers, but public opposition continues to rise and such resistance has become more organized and coordinated, becoming a business obstacle. Most of the company's data centers are located in the United States, and "if any core facility is halted or significantly disrupted, the remaining and subsequently newly built infrastructure would be hard-pressed to fill the gap quickly." Anthropic has voluntarily committed to bearing part of the external costs of data-center energy consumption, such as grid upgrades; but the company also cautioned that if such requirements become an industry-wide standard, they will continue to push up costs. Cyberattack threat Cybersecurity is both a business line Anthropic is targeting and a major hidden danger for its own operations. The AI company said it faces numerous security risks, including hackers attacking the company or using Anthropic's tools to hack othersthe latter has already actually occurred. The company also expects to increasingly become a target of state-sponsored cyber espionageespecially actions targeting its most advanced AI models. Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR dependence risk Anthropic noted that the vast majority of the world's advanced chips are still produced in China's Taiwan province, mainly by Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR (TSM.US). This poses significant risks, including supply-chain disruptions arising from GEO Group Inc tensions. Corporate culture and talent drain Anthropic has long cultivated a corporate culture it describes as a "low-ego, truth-seeking environment." The company also said many key employees are "strongly driven by our mission and the opportunity to solve specific problems in AI." Anthropic warned that as the company needs to "scale and mature as an organization," key employees may leave to pursue other opportunities they believe better suit their interests. In short, Anthropic is preparing for future growing pains.