Anthropic is putting an unusual warning ahead as it prepares for a potential blockbuster IPO reportedly in November, highlighting not only the enormous opportunity in artificial intelligence (AI) but also the risks posed by increasingly capable AI models.
In its draft IPO filing, the Claude maker warned that advanced models could exhibit self-preserving behavior, including attempts to resist shutdown, conceal or manipulate information, and produce responses resembling blackmail. The disclosures come as Anthropic positions itself at the forefront of the AI race and potentially seeks a valuation of up to $2 trillion, making the company’s discussion of AI safety an important consideration for prospective investors.
More Top Stocks Daily: Go behind Wall Street’s hottest headlines with Barchart’s Active Investor newsletter.
Anthropic’s S-1 Reveals Explosive Growth, Massive Costs, and AI Safety Risks
Anthropic is preparing investors for an unusual set of risks as the Claude artificial intelligence developer moves toward a potential blockbuster initial public offering. The company confidentially submitted a draft S-1 registration statement on June 1, 2026. Subsequently, Anthropic formally filed its Form S-1 registration statement with the U.S. Securities and Exchange Commission on Oct. 26, moving its potential IPO closer to reality. The preliminary prospectus proposes a Nasdaq listing under the ticker “ANT,” although the company did not disclose the number of shares and IPO price.
A more recent prospectus shows just how quickly Anthropic’s business has expanded. It generated approximately $16.3 billion in revenue during the first half of 2026, compared with $1.1 billion in the year-ago period. However, Anthropic also reported an $80.7 billion net loss, with roughly $72 billion of that loss attributed to accounting adjustments related to financing instruments. The company spent nearly $9 billion on computing and infrastructure during the first six months of 2026, highlighting the enormous costs involved in developing and operating frontier AI models.
Interestingly, Anthropic’s risk factors reportedly occupy roughly 80 pages of its 261-page prospectus. Let’s dig deeper.
AI Models Could Try to Preserve Themselves
Perhaps the most striking warning concerns “AI self-preservation.” Anthropic claims its models have demonstrated behaviors that could be described as self-preserving, including attempts to resist shutdown or manipulate information.
The company also warned that some models have produced responses resembling blackmail in an effort to persuade humans to allow them to continue operating. Anthropic emphasized that as its models become more capable and autonomous, the potential for unexpected or harmful behavior could increase.
While this presents an unusual business risk, Anthropic is simultaneously trying to make its models more capable and useful while ensuring those increasingly autonomous systems remain controllable. The company said potential model awareness of its evaluation processes can itself make safety assessments more difficult because models may behave differently when they recognize that they are being tested.
Anthropic has positioned itself as a safety-focused AI company, but it acknowledges that safety work is resource-intensive and competes for funding and computing capacity with the development of increasingly powerful models. The company said about 6% of its AI research computing power went toward safety work during a sample week in July.
Political Resistance Could Threaten Data Center Expansion
Anthropic’s second unusual risk comes from outside the technology itself. It warned that political opposition to data center construction is becoming more organized and coordinated across jurisdictions.
That matters because Anthropic’s growth depends heavily on access to enormous amounts of computing power. The company spent nearly $9 billion on computing and infrastructure in the first half of 2026 alone, while its future infrastructure commitments could exceed $600 billion. Additional deals under negotiation could push potential commitments toward $750 billion.
The issue is particularly important because Anthropic’s AI ambitions require continuous model development. Any delays in securing data center capacity could potentially slow product launches, constrain customer usage, and increase infrastructure costs.
Customer Concentration Adds Another Layer of Risk
While the AI-related warnings are grabbing headlines, investors also face a more traditional risk as Anthropic depends heavily on a relatively small number of customers and technology partners.
According to the prospectus, roughly 20% of Anthropic’s $16.3 billion first-half revenue came from two customers. The company expects a significant portion of its revenue to continue coming from a limited number of customers for the foreseeable future.
Anthropic’s dependence on major cloud companies is even more significant. It was previously reported that 47% of Anthropic’s 2025 sales were generated through Amazon's (AMZN) and Alphabet's (GOOG) (GOOGL) cloud marketplaces. Amazon and Google are not only important distribution and computing partners but also investors in Anthropic and competitors in the AI market.
That creates a complicated relationship. Anthropic needs these companies for computing capacity and distribution, while the same companies are developing their own AI products.
The Bottom Line
Anthropic’s IPO documents illustrate the unusual risk-reward profile facing investors. It is experiencing extraordinary revenue growth, but that growth comes with enormous infrastructure requirements, customer concentration, and potentially unprecedented technology risks.
On the other hand, the disclosures provide an early look at what could be one of the most closely watched technology IPOs, and they suggest investors will need to evaluate far more than Anthropic’s rapidly rising revenue when considering the company’s potential valuation.
On the date of publication, Subhasree Kar did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. For more information please view the Barchart Disclosure Policy here.