Anthropic’s warnings in its S-1 filing about artificial-intelligence systems potentially threatening humanity have understandably garnered a lot of attention in recent weeks. However, there are also far more quotidian financial and competitive risks the company faces to which investors should pay attention.
The maker of frontier AI model Claude filed it’s long-anticipated prospectus ahead of its initial public offering which is expected next month. There’s a lot to parse in the 318-page document, but one part that has received enormous attention is the lengthy risk factors section.
A risk factor section is a disclosure in a public company’s securities filings outlining issues that could negatively impact the business. Anthropic’s risk factor section is a whopping 94 pages, and includes unconventional risks, including the possibility that its AI systems may pose “catastrophic or existential risks to humanity.”
“Anthropic highlights risks in its S-1 and while this is standard for any public filing, its risks have certainly garnered more attention than most IPOs,” says Ryan Lee, senior vice president of product and strategy at Direxion. “While it’s impossible to know which risks will be most relevant until trading begins, each can be viewed as a trading catalyst.”
In its filing, Anthropic says it faces “intense competition, and if we are unable to compete effectively, our business, financial condition, and results of operations could be adversely affected.” Much of that competition comes from OpenAI, Microsoft, Alphabet, and Meta Platforms, each of which are building their own competing AI models.
With so many players competing for dominance, there’s also the risk of AI models becoming commoditized, meaning they could become similar enough that companies like Anthropic lose competitive advantages and pricing power.
“Newer models are meaningfully expanding what customers can accomplish, giving them a reason to move quickly. That is encouraging for future demand and the pace of technological advancement, but it also shows the importance of consistently delivering leading [large language model] capabilities,” says Willy Lee, principal at Neostellar Capital.
Another risk is that Anthropic’s five largest customers represented approximately 25% of its revenue for the nine months ended Sept. 30. If any of those customers curtail spending Anthropic says it could “adversely affect our prospects, business, and financial condition.”
Then there is the persistent and well-established risk of capacity constraints, which are affecting every AI business. Despite the enormous growth in compute capacity, there still isn’t enough to match demand. Anthropic warns it “may be unable to train models that remain competitive with those of our competitors or adequately serve our customers.”
In an effort to gain access to more compute, Anthropic has commitments to spend roughly $630 billion on data-center capacity and chip leases over the next 10 years.
Anthropic also says it faces potential government intervention and regulation, which could increase costs and restrict operations. The conversation around how AI companies should be regulated has intensified as concerns about the capabilities of their models—and those models’ tendency to perform in unexpected ways—have grown.
It’s likely many of the potential risks Anthropic lists won’t come to pass, but investors would do well to keep them in mind when deciding whether to invest in a business that will be the first of its kind to be publicly traded.
Comments