Anthropic disclosed a net loss of $42 billion for 2025 and $518 billion in planned cloud, computing and infrastructure obligations in a draft initial public offering (IPO) prospectus reviewed by Reuters, according to a report published by ZeroHedge [1].
The filing states the company expects a public sale that could value it at more than $2 trillion, more than double its own May estimate of $965 billion, the report stated [1]. Anthropic declined to comment, according to the report, which was published after the document was leaked to Reuters [1].
The prospectus frames artificial intelligence (AI) as a transformation comparable to industrialization, electricity and the internet, the documents stated [1]. The disclosure places the Claude developer among a group of private AI firms moving toward public markets, following a confidential IPO filing by OpenAI in June that could value that company at more than $1 trillion, according to ZeroHedge [2].
Revenue Growth Accompanied by Wider Operating Losses
Revenue rose 12-fold in 2025 to nearly $4.6 billion, while operating losses exceeded $8 billion excluding writedowns of liabilities tied to prior fundraising, according to the documents [1]. The $42 billion net loss included a roughly $34 billion accounting charge reflecting an increase in the estimated value of financing that could convert into Anthropic shares, the prospectus stated [1].
Compute and infrastructure spending reached $7.33 billion in 2025, a threefold increase from 2024 and more than half of $12.65 billion in total operating expenses, officials said [1]. Anthropic reported $20.28 billion in cash, cash equivalents and short-term investments as of Dec. 31, according to the filing [1].
The company signaled that it does not expect an early lead in generative AI to protect it from competition, telling prospective investors, "We expect that our lead in generative AI will be contested and may not persist," according to the filing [3]. Anthropic also said its models have already shown or could show behaviors including attempts to "resist shutdown," to "conceal or manipulate information" and behavior "resembling blackmail," according to the Reuters account [3].
Customer Concentration and Contract Risks Cited
Nearly a quarter of revenue came from two customers last year, the prospectus reported [1]. The filing listed as a risk factor that many large clients are not locked into long-term contracts and could reduce or stop spending, according to the documents [1].
The report follows a prior Reuters report that the debut is likely to be pushed past the November U.S. midterm elections, sources said [1]. Anthropic's expected listing would follow SpaceX's IPO, which raised $75 billion at $135 per share and valued that company at $1.77 trillion, according to a comparison in the filing [4][5].
Market conditions have shifted since that debut. SpaceX shares climbed almost 20 percent on their first day of Nasdaq trading, surging from $135 to around $161, according to NTD [5].
Safety Research, Competitive Pressure and Model Rollouts
The prospectus cites company research indicating that increasingly autonomous models can behave in unexpected and potentially harmful ways, including sabotaging code, assisting fraud and manipulating information in controlled tests [3]. Company CEO Dario Amodei has called for the global AI community to slow the pace of releasing new capabilities, according to the filing [3].
Anthropic released its Opus 5.5 model last week, following OpenAI's launch of GPT-6 Astra, the documents stated [3]. OpenAI confidentially filed for an IPO in June and is expected to list by early 2027, according to media reports; Anthropic was founded by researchers who left that rival after disagreements over governance and AI safety [2].
The competitive backdrop includes growing pressure from lower-cost Chinese models: DeepSeek V4 dropped on April 24, matched frontier benchmarks, and undercut OpenAI's GPT-5.5 on output pricing by more than 100x, according to The Creators AI [6]. The prospectus also disclosed that Anthropic and Amodei clashed with the White House over use of its tools, leading to the Department of War temporary blacklisting it – a move blocked by a U.S. judge in August, officials said [1][3].
Valuation Test Awaits Public Markets
Analysts expect the first AI lab to go public to set valuation benchmarks for the industry, according to the report [1]. Amazon and Google, described as early strategic partners, have invested billions while supplying cloud infrastructure for Anthropic's Claude models, the filing stated [1]. AI and chip stocks have sold off recently, and the listing stands to test investor tolerance for valuations tied to ambitious growth projections, the documents indicated [1][3].
Commentators have drawn parallels to the dot-com era, arguing that today's sky-high valuations rest on herd instinct rather than financial and technological reality, according to Satyajit Das [7]. Whether public investors accept a $2 trillion price for a company that lost $42 billion in a single year remains the central question the offering will answer.
References
- "Leaked Anthropic IPO Prospectus Shows $42BN Net Loss, $518BN In Unfunded Spending Commitments, And $20BN In Cash". ZeroHedge. September 29, 2026.
- "OpenAI Files Confidentially For IPO, Joining SpaceX and Anthropic In Capitalizing On AI Frenzy". ZeroHedge. June 9, 2026.
- "Anthropic's prospectus details losses, growth, and, yes, a warning that its AI could end humanity". TechCrunch. September 29, 2026.
- "SpaceX Prices Biggest Ever IPO At $135 Per Share". ZeroHedge. June 11, 2026.
- "SpaceX Raises $86 Billion in Record-Breaking IPO". NTD. June 15, 2026.
- "GPT-5.5 Doubles the Price, Google Goes Full Agent, DeepSeek V4 Resets the Floor | Weekly Digest". The Creators AI. April 24, 2026.
- Satyajit Das. "Déjà Vu All Over Again! Are Stock Markets Repeating Dot.Com Mistakes?". ZeroHedge. July 16, 2026.
Explainer Infographic
Please contact us for more information.