Anthropic’s long-awaited IPO filing is doing two things at once: it is trying to sell investors on explosive growth, and it is warning them that the technology behind that growth could pose “existential risks to humanity.” The S-1 from the Claude maker lands as the startup says revenue surged to nearly $4.6 billion in 2025, while operating loss still topped $8 billion. The filing frames Anthropic as a fast-scaling AI leader, but also as a business whose own prospectus spends more space on danger than on operations.
The balance inside the document is striking. Reuters, citing the filing via NDTV Profit, said roughly 80 of the prospectus’s 261 pages are devoted to risk factors, compared with 48 pages on the business. That tilt tells investors exactly how Anthropic wants the market to read this deal: high growth, but with an unusually large warning label attached. The company’s prospectus language, as reported by the Financial Times and Reuters, says the technology could create “existential risks to humanity” and “catastrophic or existential risks to humanity.”
That is not standard IPO boilerplate. It is a signal that Anthropic is not trying to hide the policy and safety debate surrounding frontier AI. Instead, it is putting those concerns in the center of the story. For public-market buyers, that matters because the valuation case now has to survive not just revenue growth, but a much bigger set of operational, regulatory and reputational questions than most software listings face.
The topline numbers are the headline attraction. Reuters reported Anthropic’s 2025 revenue at nearly $4.6 billion, roughly 12 times higher than a year earlier. That kind of growth is the kind public investors typically chase in AI, especially when the company is positioned as a key rival in a market that has become one of the most closely watched in tech. But the filing also shows how expensive that growth remains.
Operating loss exceeded $8 billion in 2025, reaching $8.06 billion, up from $2.98 billion in 2024, Reuters reported. Net loss was about $42 billion, though roughly $34 billion of that was a non-cash accounting charge tied to convertible financing rather than operating spend. Even with that adjustment, the message is clear: Anthropic is still burning a huge amount of capital to scale its models, infrastructure and customer base. Investors considering the IPO will have to decide how much of that burn is temporary and how much is structural.
One reason the losses are so large is the scale of the infrastructure commitment. Reuters reported the company plans $518 billion in cloud, computing and infrastructure obligations in coming years. That number underscores how capital intensive the AI race has become. Anthropic, like its peers, is not just building software. It is locking in access to massive computing power, and those commitments can shape everything from margin assumptions to financing needs.
At the same time, Anthropic’s balance sheet is not empty. Reuters said the company held $20.28 billion in cash and short-term investments as of Dec. 31, 2025. That gives it room to keep investing, but it does not erase the scale of the obligations ahead. For public investors, the question is not whether Anthropic has enough cash today to get through the next phase. It is whether the current pace of spending, once exposed to the discipline of public markets, can translate into durable profits before the cost curve overwhelms the growth curve.
The filing also reveals concentration risk. The Financial Times, via Yahoo Finance, reported that nearly a quarter of 2025 revenue came from just two customers, and many contracts were not long term. That makes the growth story look less diversified than the headline revenue number suggests. For a company that is pitching itself as a central platform for enterprise AI adoption, customer concentration can be a concern because it gives a small number of buyers outsized influence over future revenue.
That does not mean the business is fragile. It does mean the revenue base may not yet have the kind of breadth public-market investors usually want in a premium listing. In a market where AI valuations already assume strong forward execution, any sign that revenue depends on a narrow set of large customers can make investors more cautious about the sustainability of the growth rate.
The market is not waiting for the listing to begin debating the price. Backers expect Anthropic to debut at a valuation above $2 trillion, according to the Financial Times, more than double its $965 billion valuation in a May funding round. That gap shows how much optimism is already embedded in the company’s path to market. It also suggests investors are being asked to underwrite not just current revenue, but years of expansion, margin improvement and platform dominance.
That is a high bar even for a company posting a 12-fold revenue jump. And because Anthropic is not yet publicly traded, there is no live stock price to anchor the debate. The valuation discussion is still hypothetical, but it is already central to the story. Any IPO priced near the levels now being discussed would put the company in rarefied territory before it has shown sustained profitability.
The filing points to an IPO on Nasdaq this autumn, though Reuters said the deal is likely to be pushed until after the November 2026 US midterm elections. That timing matters because public-market appetite for a deal this large can shift quickly with macro conditions, regulatory attention and political noise around AI. The later the listing, the more time Anthropic has to extend growth and perhaps reduce the loss profile investors will scrutinize.
There is also evidence the business momentum is still improving. The Financial Times reported second-quarter 2026 revenue reached $11.5 billion, putting the company on track for a second straight adjusted operating profit. That is important because it suggests the company is not simply scaling revenue at any cost; it is beginning to show some operating leverage on an adjusted basis. But that does not erase the earlier losses, the infrastructure commitments or the safety warnings. It simply means the story is becoming more complex.
Anthropic’s filing is unusual because it asks investors to believe two opposite things at once. On one hand, the company says it is growing at exceptional speed and moving toward better operating performance. On the other, it says the technology itself may pose “existential risks to humanity,” and it has built a prospectus where risk dominates the page count. That combination may actually suit the current AI market, where fear and enthusiasm have become intertwined.
But the IPO will force a harder question: can a company warning about existential danger still command a premium valuation above $2 trillion? The answer will depend on whether investors focus more on the scale of the revenue opportunity or on the scale of the liabilities, losses and uncertainties wrapped around it. For now, Anthropic has done the one thing a listing prospectus is supposed to do. It has made the market look.
