Anthropic’s confidential draft registration statement points to an IPO candidate with exceptional commercial growth and an equally exceptional dependence on computing infrastructure: the Claude developer generated about $4.6 billion in 2025 revenue while spending roughly $7.33 billion on compute and infrastructure alone, according to a draft prospectus reviewed by Reuters.
The figures place Anthropic among the largest prospective technology listings in years, while exposing the financial burden behind the race to build and operate frontier AI models. Revenue was about 12 times higher than in 2024, but the company reported operating losses exceeding $8 billion for 2025 as its infrastructure, research, personnel and other costs expanded rapidly.
Reuters reported that Anthropic has confidentially filed a draft S-1 registration statement with the U.S. Securities and Exchange Commission. The company’s eventual valuation could exceed $2 trillion, Reuters said, though the offering’s timing remains subject to SEC review and market conditions. A listing could also be postponed until after the U.S. midterm elections in November.
Compute costs outrun annual revenue
Anthropic’s $7.33 billion in 2025 compute and infrastructure spending was roughly three times its prior-year level and represented more than half of total operating expenses of about $12.65 billion, the draft prospectus showed. The outlay equaled about 1.6 times the company’s annual revenue.
That gap illustrates the commercial challenge facing companies developing large AI systems. Enterprise demand can scale quickly once models are integrated into software products and internal workflows, but each expansion in usage requires access to expensive chips, data-center capacity, networking equipment and electricity. Training newer models adds another layer of costs before those systems begin generating customer revenue.
Anthropic reported a GAAP net loss near $42 billion for 2025. About $34 billion of that result came from accounting charges related to higher fair values for liabilities, including convertible financing arrangements, according to the prospectus. Those non-cash valuation effects sharply increased the reported net loss beyond the company’s operating loss, although they do not remove the underlying pressure created by multi-billion-dollar annual infrastructure spending.
The company ended 2025 with about $20.28 billion in cash, cash equivalents and short-term investments. That balance gives Anthropic a substantial funding cushion, but its disclosed future commitments show why a public offering could become strategically useful even after large private fundraising rounds.
More than $500 billion in contractual commitments
Anthropic disclosed about $518 billion in future cloud, compute and infrastructure payment obligations over coming years. The company described those amounts as multi-year contractual commitments rather than a single-year capital expense.
The total dwarfs its current annual revenue and indicates that the company is reserving access to computing resources far in advance. Such agreements can protect an AI developer from chip and data-center shortages, but they also reduce flexibility if customer growth slows, model costs change or competing technologies make existing capacity less valuable.
The prospectus also refers to a seven-year cloud-capacity agreement with Akamai valued at roughly $11.6 billion. Amazon and Alphabet appear in the filing as both financial backers and cloud suppliers, creating relationships in which major partners provide capital as well as critical operational infrastructure. Nvidia is identified as a supplier of GPUs and networking equipment.
Those dependencies place Anthropic’s financial plans alongside a small group of companies controlling the hardware and cloud systems needed to run advanced AI. Its ability to meet customer demand will depend partly on suppliers whose services are central to its product delivery, rather than merely routine operating vendors.
Enterprise growth brings concentration risk
Anthropic’s customer data shows strong adoption among large organizations. The number of customers spending more than $100,000 annually rose sevenfold year over year, according to the filing. More than 500 customers had annualized spending above $1 million, while eight of the 10 largest Fortune 500 companies used Claude.
The company had previously said in February that its annualized revenue run rate had reached $14 billion. Claude Code, its programming-focused product, had an annualized revenue run rate above $2.5 billion.
These measures suggest that Anthropic is expanding beyond experimentation budgets and into larger enterprise deployments, particularly for software development and workplace automation. Annualized run-rate figures, though, measure current revenue pace rather than booked full-year sales, so they should not be treated as equivalent to reported 2025 revenue.
The filing also reveals a risk common to rapidly growing enterprise software businesses: two customers accounted for close to one-quarter of 2025 revenue. Anthropic warned that some large customers do not have long-term contractual commitments, leaving a portion of its revenue exposed to changing technology budgets and vendor decisions.
A few large enterprise clients can accelerate growth quickly, especially when their usage increases across large workforces. They can also create abrupt revenue swings if one customer shifts workloads to a rival model provider, builds internal systems, or reduces spending on AI tools.
IPO would test the market’s appetite for AI infrastructure risk
A public listing would give traders their first detailed view of Anthropic’s economics, including the degree to which rising Claude revenue can eventually absorb its infrastructure obligations. The prospectus presents a company that is growing at a rare pace but has locked itself into an unusually capital-intensive operating model.
That distinction also matters for cryptocurrency markets, where AI-linked tokens are often traded as a simple proxy for enthusiasm around artificial intelligence. Anthropic’s disclosures show that demand for AI services does not automatically translate into straightforward profitability for companies supplying or building the technology. Revenue growth, customer retention, access to chips and cloud capacity, and the cost of serving each additional user remain separate variables.
Anthropic’s prospective IPO would therefore offer a public-market benchmark for the economics underlying the AI boom: rapidly expanding enterprise sales on one side, and a multi-year commitment to vast physical computing capacity on the other.
Explore how AI and finance intersect—learn about tokenized equities and their impact on future capital markets.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
