Anthropic’s IPO filing is raising questions that go well beyond revenue, valuation and losses. In a review published late Oct. 27 Taipei time, CNBC said the company’s S-1 devotes more than 90 pages to risk disclosures, including several warnings that are unusual in a standard IPO document. They range from "catastrophic or existential risks to humanity" to data center opposition, electricity shortages, child safety concerns and the possibility that slower market growth could hurt the business.
Model safety is framed as a core business risk
The S-1 states the issue in blunt terms. If foundation models are not properly safeguarded, they could pose "catastrophic or existential risks to humanity," especially if they are misused, replicated, weaponized or operated with a high degree of autonomy.
Anthropic says safety has to be built into the product itself. If that effort falls short, the company says the impact could reach its brand, operations and financial results. The filing also says competitive pressure, market timing and stakeholder expectations make it harder to build and release frontier models safely. Anthropic adds that the technology is inherently complex, that it cannot guarantee its safeguards will always be sufficient, and that it may not be able to anticipate every scenario in advance.
Political scrutiny and regulatory friction appear in the filing
The introduction to the S-1 also points to the political environment. Anthropic says the broader climate, along with its public positions on AI regulation and safety, could bring heavier scrutiny, political backlash or asymmetric regulation. The filing also warns that certain government agencies could target the company in a disproportionate or selective way.
CNBC separately noted friction between Anthropic and the Trump administration, including disputes tied to Pentagon supply-chain risk labeling and export controls. Some litigation is still ongoing.
Data center expansion could face organized resistance
Claude’s training and operation depend heavily on expanding data center capacity. In the filing, Anthropic says opposition to data center development has become more organized and more coordinated across jurisdictions. It also says the company’s high profile could make related projects more likely to draw attention.
Citing the National Conference of State Legislatures, CNBC said about 16 U.S. states are considering pauses or restrictions on data center construction. The report added that the issue is becoming more political ahead of the midterm elections.
Power supply may become a direct limit on compute
Anthropic also identifies electricity as a major constraint. The S-1 says insufficient national and global access to adequate power and transmission infrastructure could limit the company’s ability to secure compute.
Even if some planned facilities use on-site natural gas generation, the filing says they would still face fuel price swings, pipeline capacity limits and transportation constraints. If demand tightens, electricity prices could also rise. In practical terms, the risk is not only whether data centers can be built, but whether they can be connected to affordable power on time. Sharp increases in power or compliance costs, regulatory curbs on electricity use, or project delays could all weigh on expansion.
Child safety is listed alongside platform access concerns
On child safety, Anthropic says its terms of service prohibit users under 18 from using Claude. At the same time, the company acknowledges in the S-1 that minors have accessed the platform in the past and may do so again.
The filing says child safety risks could become more acute as adoption grows, including among minors and other vulnerable users. The report also notes that peers such as OpenAI and Meta have faced pressure from allegations and lawsuits tied to child safety. Meta recently agreed to a settlement of about $17 billion with multiple state attorneys general over issues including youth mental health and addiction, according to the report.
Moving too slowly is also described as a risk
Anthropic’s filing does not only warn about moving too fast. It also says slower-than-expected market development could hurt growth. If businesses cut AI spending, or if the market for models and applications grows more slowly than expected or contracts, the company says its growth could slow or stall.
The filing adds that customers may decide the incremental gains from frontier models do not justify a premium price and may switch to cheaper or open-source alternatives.
Offering size and price are still blank
For now, Anthropic has disclosed that its S-1 is on file before effectiveness and that its ticker will be ANT. The size and pricing of the offering, however, have not been filled in. Investors are now likely to watch how these risk disclosures are handled during the roadshow, and whether disputes around data centers, electricity and politics intensify before the listing.
