MSX Daily U.S. Stock Watch: Anthropic IPO Prospectus Revealed! Revenue Grew 12x in a Year, $518 Billion Compute Commitment Tests the AI Business Model
- Key Takeaway: Anthropic has confidentially filed its S-1 prospectus. 2025 revenue stood at approximately $4.6 billion, up about 12x year-over-year, but operating losses exceeded $8 billion and compute spending reached $7.33 billion. Its IPO will serve as the public market's first major stress test of the frontier AI business model.
- Key Points:
- 2025 revenue was approximately $4.6 billion (about 12x that of 2024), total operating expenses were approximately $12.65 billion, operating losses exceeded $8 billion, and GAAP net loss approached $42 billion.
- Compute and infrastructure spending was approximately $7.33 billion (up about 3x year-over-year), roughly 1.6x full-year revenue, accounting for more than half of total operating expenses.
- Cloud computing, compute, and infrastructure payment obligations over the coming years total approximately $518 billion — these are multi-year contractual commitments, not single-year capital expenditures.
- Two customers contributed nearly a quarter of revenue, and some major customers are not bound by long-term contracts, creating significant customer concentration risk.
- The number of customers spending over $100,000 annually grew 7x in a year, with more than 500 customers spending over $1 million; Claude Code's annualized revenue run rate exceeds $2.5 billion.
- A potential IPO valuation could exceed $2 trillion, with the listing possibly delayed until after the November U.S. midterm elections.
- Amazon and Alphabet are both investors and cloud providers, while Nvidia, Akamai, and the data center and power sectors could be driven by its expansion.
[MSX Research Institute · Daily US Stock RWA Observation] is a signature daily report produced by MSX, a leading RWA trading platform. Leveraging our strong macro research capabilities, we capture the core pulse of global traditional US stocks, liquidity changes, and the RWA tokenization market, helping you position ahead for quality assets.
Today's Observation:
Anthropic may become the first frontier AI model company to face full public market scrutiny.
According to a draft IPO prospectus reviewed by Reuters, Claude developer Anthropic generated approximately $4.6 billion in revenue in 2025, a roughly 12x year-over-year increase; however, the company's operating loss for the same period exceeded $8 billion, with compute and infrastructure spending reaching $7.33 billion, already significantly exceeding full-year revenue.
Notably, Anthropic disclosed future payment obligations for cloud computing, compute, and infrastructure over the coming years amounting to approximately $518 billion. This makes the company's upcoming IPO not just a fundraising event, but potentially the first major stress test of the generative AI business model by public markets.
On June 1, Anthropic confirmed it had confidentially submitted a draft S-1 registration with the US Securities and Exchange Commission. What can currently be confirmed is that Anthropic has initiated the listing process, along with historical financial and risk data disclosed by the media; the valuation exceeding $2 trillion, the final fundraising size, and the specific listing timeline may still be subject to adjustment.
The most important information in this prospectus is not how much Anthropic has lost, but rather what it reveals about the growth formula of frontier AI companies: first sign massive compute contracts, use more compute to train stronger models, then rely on model capabilities to attract enterprise customers and developers, and ultimately use revenue growth to cover continuously expanding inference, training, and R&D costs.
Anthropic has already proven that the first half of this equation can deliver rapid revenue growth. But it has not yet proven that revenue growth can outpace the growth of compute and operating costs.
Data in a Minute:
• Anthropic confidentially submitted a draft S-1 prospectus to the SEC on June 1, 2026;
According to the prospectus reviewed by Reuters, Anthropic's 2025 revenue was close to $4.6 billion, approximately 12 times that of 2024;
• Total operating expenses in 2025 were approximately $12.65 billion;
• Of which compute and infrastructure spending was approximately $7.33 billion, roughly tripling year-over-year and accounting for more than half of total operating expenses;
• Operating loss in 2025 exceeded $8 billion;
• GAAP net loss in 2025 was close to $42 billion, of which approximately $34 billion came from accounting charges related to the increase in fair value of liabilities such as convertible financing;
• As of the end of 2025, the company held approximately $20.28 billion in cash, cash equivalents, and short-term investments;
• Anthropic's future payment obligations for cloud computing, compute, and infrastructure over the coming years total approximately $518 billion, not a single-year capital expenditure;
• Two customers contributed nearly a quarter of Anthropic's 2025 revenue, and the company also warned that some large customers are not bound by long-term contracts;
• Anthropic disclosed in February this year that its annualized revenue run rate reached $14 billion, with Claude Code's annualized revenue run rate exceeding $2.5 billion;
• Reuters said Anthropic's potential IPO valuation could exceed $2 trillion;
• The company's listing timeline may be postponed until after the November US midterm elections, with specific arrangements still subject to SEC review and market conditions;
• Anthropic recently signed a seven-year cloud computing capacity agreement with Akamai totaling approximately $11.6 billion.
MSX View:
Anthropic's prospectus gives the market its first relatively complete look at the revenue, costs, and long-term compute obligations of a frontier AI model company. The results simultaneously support two completely different conclusions.
On the optimistic side, Anthropic's business growth is indeed very fast. Revenue expanded approximately 12x in one year, indicating that Claude is no longer just a research product. The company has established actual revenue streams in enterprise API, Claude Code, Claude for Work, and cloud platform distribution. Anthropic also stated that the number of customers spending over $100,000 annually grew sevenfold in one year, customers with annualized spending exceeding $1 million have surpassed 500, and eight of the top ten Fortune 500 companies use Claude.
Claude Code in particular is transforming Anthropic from a general-purpose chatbot company into an enterprise software and developer tools provider. Code generation, software testing, and agent development are all scenarios with high usage frequency and strong enterprise willingness to pay, and are more conducive to generating sustained revenue based on Token usage.
But the financial data also reveals the core contradiction of this business today. Anthropic's 2025 revenue was approximately $4.6 billion, but compute and infrastructure spending alone reached $7.33 billion. In other words, before accounting for R&D, sales, administrative, and other expenses, compute spending already amounted to approximately 1.6 times full-year revenue.
This indicates the company is still in a phase of using large amounts of capital to purchase growth. Rapid revenue growth can improve this relationship, but only if the unit cost of model calls continues to decline, customer usage grows steadily, and price competition does not offset the margin space created by falling inference costs. Anthropic also faces a unique problem: the more capable the AI model, the longer a single task may take to execute, the more tools it may invoke, and the more Tokens it may consume.
Traditional software can serve an increasing number of users with relatively fixed server costs; agent products, however, may generate more inference costs as usage depth increases. If the price enterprise customers pay does not grow faster than per-unit compute consumption, revenue expansion will not necessarily improve gross margins in tandem. The $518 billion in compute and infrastructure obligations further amplifies this risk. This figure should not be understood as Anthropic spending $518 billion in one lump sum next year. It is a contractual obligation covering multiple future years, with some payments contingent on data centers being built on schedule, compute capacity being delivered, and services meeting agreed standards.
But even if paid over many years, it still represents an extremely massive fixed investment commitment. The logic is that Anthropic expects future AI demand to continue growing rapidly, so it must lock in chips, power, and data center resources in advance. If future Claude usage meets expectations, these contracts can ensure the company has sufficient capacity; if model competition intensifies, prices decline, or customer demand falls short of expectations, the pre-locked compute could become a heavy burden.
For the US stock market, Anthropic's listing will also affect a range of related companies. Amazon and Alphabet are both Anthropic investors and its important cloud computing suppliers; Nvidia provides GPUs and related networking equipment; Akamai, data center operators, power suppliers, and server manufacturers may all receive orders from its infrastructure expansion.
If the market accepts a valuation exceeding $2 trillion, it means investors are willing to pay a premium of hundreds of times current revenue for future AI platform status and high-speed growth. This could raise valuations for OpenAI and other AI companies, and could also drive repricing across cloud computing, chip, and data center sectors. But if investors focus on operating losses, customer concentration, and long-term compute obligations, Anthropic could also become the starting point for the market to re-examine the return on AI capital expenditure.
Anthropic has already proven that enterprises are willing to pay for Claude, and that frontier AI can build billions of dollars in revenue in an extremely short time. In the next phase, what it needs to prove to public markets is that this growth can ultimately break free from dependence on continuous massive fundraising and form a self-sustaining business model. If the answer is yes, Anthropic could become the first truly large-scale pure AI public company; if the answer is no, this prospectus could also become the turning point for the market to reassess the entire generative AI investment cycle.
About MSX:
MSX is a leading RWA trading platform dedicated to providing safe, efficient, and transparent access to global financial markets. As one of the world's earliest on-chain US stock trading platforms, MSX has always been at the forefront of the industry, leading market transformation through continuous innovation.
The platform deeply integrates blockchain technology with compliance frameworks, comprehensively offering spot and derivatives trading for nearly 400 tokenized stocks and Pre-IPO assets, perfectly bridging the gap between traditional finance and the digital asset industry.
Centered on the core mission of "enabling quality assets to circulate freely," MSX has now built a diversified digital financial services system covering US stock spot and perpetual contracts, crypto-to-crypto trading, Pre-IPO, and the Maitong Research Institute, aiming to provide global investors with round-the-clock, high-performance access to quality assets.
Official Website: https://msx.com/
Chinese TG:https://t.me/MSXOfficial
Global TG:https://t.me/MSXGlobal
Now simultaneously available on the global App Store / Google Play.
Risk Disclaimer: Macroeconomic conditions and US stock markets are highly volatile. This content is provided solely for academic and research observation reference by the Maitong Research Institute and does not constitute any investment advice.
