Anthropic takes a key step toward its epic IPO! IPO filing to be published Monday, $2 trillion valuation target faces a reality check.

As the AI industry continues to face multiple questions over capital expenditure, commercial returns, and technical safety, Anthropic is preparing to take a key step toward the public capital markets. According to reports, Anthropic will publish its S-1 filing for its initial public offering (IPO) at 9:00 a.m. Eastern Time on Monday (21:00 Beijing Time). The company will formally launch its IPO roadshow as early as the week of November 9. People familiar with the matter said the company expects its IPO size to be comparable to, or possibly even larger than, that of SpaceX (SPCX.US). What to look for in the filing? Revenue quality, cash burn, and compute commitments are the three key areas For weeks, draft IPO filings and related reports have gradually leaked some of Anthropic's financial information. As a result, some of the disclosures in the S-1 may not come as much of a surprise. But the upcoming S-1 filing may help clarify a number of ambiguous points. For investors, Anthropic's IPO filing should not be read solely for revenue growth rates and valuation figures; more important is to break down the sources of its growth, its cost structure, and the price it must pay to sustain growth. In particular, rapid revenue growth does not automatically mean cash flow has improved, let alone that a high valuation is necessarily justified. First, investors need to focus on the speed and composition of revenue growth. Anthropic's commercialization is mainly built around the Claude model, with business spanning both enterprise and developer use of the model and revenue sources such as subscriptions. A usage-based pricing model allows revenue to expand as customer call volumes grow, but its sustainability also depends on customer usage frequency, inference costs, pricing power, and whether customers will switch to other providers as model competition intensifies. Second, it is necessary to distinguish between reported revenue, gross margin, and actual cash collection. AI model companies must continuously invest in expensive compute resources, R&D talent, and product development, and costs may rise in tandem with rapid revenue growth. Investors should focus on whether inference costs as a proportion of revenue are declining, whether gross margin can continue to improve, and when operating cash flow and free cash flow may turn positive. Third, compute procurement and long-term contract commitments in the filing are particularly important. AI companies need to lock in data centers, cloud computing resources, chips, and power supply in advance to avoid capacity bottlenecks when demand grows. But long-term commitments also mean that even if customer demand falls short of expectations, the company may still have to bear huge expenditures. Therefore, investors need to judge whether these contracts are strategic assets that safeguard growth or could turn into a heavy financial burden when the industry cools. Finally, governance structure, founder control, customer concentration, supplier dependence, and AI safety risks cannot be ignored. Anthropic's technological advantages and enterprise customer relationships may constitute competitive barriers, but if key customers, cloud service channels, and compute supply are concentrated among a few large technology companies, its business growth could be significantly affected by partners' commercial decisions. US IPOs froze in October; the whole of Wall Street is waiting for Anthropic to break the logjam In the first half of 2026, the US IPO market ran at full speed. Data compiled by Bloomberg showed that as of the end of September, US IPOs had raised as much as US$162.6 billion. However, the picture changed abruptly in the fourth quarter. Since the end of September, the US IPO market (excluding shell companies and other financial instruments) has raised only US$587.7 million. This directly made October the slowest month of the year for US IPO fundraising. The direct cause of the US IPO market's stall points to AI infrastructure companies. Nscale, DayOne, Data Centers, and other companies queuing up to list in the US are still struggling to sell their deals, but investors' appetite for risk has evaporatedbecause share prices of peers such as CoreWeave are plunging. The deeper reason lies in the reappearance of the "valuation gap." After a promising first half, a huge divergence has emerged between the high valuations management teams expect and the prices investors are willing to pay. Will Braeutigam, US head of capital markets transactions at Deloitte, described the bid-ask spread as having "blown out": "There is a stalemate on valuation, and that gap must be resolved before the broader market can move forward." Facing a stalled market, Wall Street is pinning all its hopes for a revival on Anthropic's expected listing. People familiar with the matter said Anthropic's debut next month is expected to raise an amount that could match or exceed the size of SpaceX's listing. This is not only about whether Wall Street can usher in a record "fat year," but also about clearing the way for other companies waiting on the sidelines. However, apart from Anthropic, the market has low expectations for other candidates in the near term. West Riggs, head of equity capital markets at Truist Financial Corp., said: "We have a lot of optimism about 2027, especially in software, but from an IPO perspective, the overall feeling is quite subdued. Even several non-AI-related projects we are working on appear to have been delayed." Will Braeutigam also expects that no more than six companies will raise more than US$500 million in the US market before year-end. "Most of our clients are focused on the second quarter of next year." The US$2 trillion valuation faces re-examination Anthropic's most closely watched valuation target is undergoing a shift from optimistic expectations to a reality check. Previously, market expectations for Anthropic's IPO valuation reached about US$2 trillion at one point. By comparison, the company was valued at about US$965 billion when it completed a funding round in May 2026. Listing at a US$2 trillion valuation would mean its valuation doubles within a few months, and that valuation also means investors must believe there is still enormous room for growth in its revenue, market share, and long-term profitability. But as more financial data emerges, investor attitudes are beginning to diverge. According to reports, after Anthropic's latest round of investor communications, some existing investors and potential IPO investors have begun discussing a valuation of about US$1.5 trillion, below the previously expected roughly US$2 trillion. One reason for investors' focus is that the company's revenue is still growing, but the quarter-on-quarter growth rate has slowed. Data shows that Anthropic's third-quarter revenue has reached US$16.8 billion, achieving adjusted profitability for the second consecutive quarter. But quarter-on-quarter revenue growth slowed to 45% from more than 140% in the second quarter. Across the entire technology industry, quarter-on-quarter revenue growth of 45% is already a remarkable achievement, but the problem is that Anthropic had previously been growing far too fast. In the first quarter of this year, Anthropic's revenue was US$4.7 billion, and in the second quarter it surged to US$11.6 billion, an increase of nearly 147%. By the third quarter, although revenue rose by another US$5.2 billion, the growth rate had clearly slowed. And Wall Street's previous valuation expectation of as much as US$2 trillion for Anthropic was built precisely on this ultra-fast growth. Therefore, after the slowdown in quarter-on-quarter revenue growth, some investors have begun to recalculate. According to three investors who attended Anthropic's investor day, some investors have begun discussing a listing valuation of around US$1.5 trillion. Jared Carmel, managing partner at Manhattan Venture Partners, said in an interview that Anthropic's IPO must set a price investors are willing to accept. He expects Anthropic may ultimately list at a valuation of US$1.4 trillion to US$1.5 trillion. At the same time, the other side of high growth is massive cost investment. Anthropic's net loss in 2025 was close to US$42 billion, five times the roughly US$8.3 billion in the prior year. However, about US$34 billion of that came from fair value changes in convertible financing instrumentsan accounting loss generated as valuations soared, not a cash expenditure. Excluding that, the actual operating loss was about US$8.06 billion, 170% wider than US$2.98 billion in 2024. The reason for the widening loss is swelling upfront compute spending. In 2025, compute and infrastructure spending reached US$7.33 billion, equivalent to 1.6 times that year's revenue. But the loss ratio is narrowingthe operating loss ratio fell to 176% in 2025 from about 750% in 2024. The US$2 trillion valuation reflects investors' high expectations for the long-term growth potential of the AI industry. Compared with traditional software companies, Anthropic not only hopes to generate revenue through model subscriptions, enterprise services, and API interfaces, but also bets that AI can gradually enter core enterprise scenarios such as programming, customer service, knowledge management, and business process automation. As enterprises move from experimental applications to production deployments, AI services have the opportunity to upgrade from auxiliary tools to part of enterprise infrastructure. In particular, the expansion of the Claude series models in enterprise applications and programming has enabled Anthropic to capture more sustainable commercial demand. Once enterprise customers embed AI models into software development, internal knowledge retrieval, or daily operations, they often need to continuously pay for model calls and services. Such demand is expected to create high revenue stickiness and also support the company's future revenue growth. However, a high valuation also means higher delivery requirements. Investors are paying not only for Anthropic's current business scale but also for its growth expectations over the coming years. If revenue growth slows after listing, profit improvement falls short of expectations, or the overall valuation of the AI industry contracts, the company's share price could come under significant pressure. Therefore, whether the US$2 trillion valuation can hold ultimately depends on whether the market believes Anthropic has the ability to translate its current technological leadership and demand growth into long-term cash flow generation. US$518 billion infrastructure commitment! A growth engine and a source of financial pressure If revenue growth is the core of Anthropic's valuation story, then massive compute and infrastructure needs are one of the cost factors in this business model that most needs scrutiny. According to reports, based on Anthropic's IPO prospectus seen by media, the company plans to incur about US$518 billion in cloud computing and infrastructure-related expenditures or commitments over the coming years. This scale shows that the company's judgment on future AI demand is very aggressive, and it also means its business expansion will be deeply tied to cloud computing services, data center construction, and advanced chip supply. It should be noted that this figure represents related expenditures and infrastructure commitments over many future years. It does not equal the full amount the company must pay in a single fiscal year, nor can it be simply equated with capital expenditure or cash outflows already incurred. But even accounting for differences in performance timing and contract structure, such long-term commitments will still have an important impact on the company's future financial flexibility. The cost structure of AI companies differs markedly from that of traditional software companies. For traditional software, after product development is completed, the marginal delivery cost of adding a user is usually relatively limited; large model services, by contrast, require continuous inference computing, and the more frequently users call the model and the more complex the tasks the model handles, the higher the related compute costs. For products such as coding agents that can perform complex tasks over long periods, increased user usage may bring in more revenue but may also drive up computing resource consumption. Therefore, whether Anthropic can build attractive profit margins in the future depends on whether multiple factors can improve simultaneously: whether model inference efficiency can rise, whether unit computing costs can fall, whether enterprise customers are willing to accept higher prices, and whether the company can reduce compute procurement costs through long-term contracts and resource scheduling. If the commercial value of AI models continues to rise and enterprise customers are willing to pay for higher productivity, then massive infrastructure investment may become a necessary condition for building competitive barriers. Conversely, if model capabilities gradually converge, service prices continue to fall, or enterprise customers begin to cut AI budgets, long-term compute commitments could become a heavy cost burden. This also means the market cannot simply view the US$518 billion infrastructure commitment as a positive signal for the chip and data center industries. For upstream suppliers, it may represent future demand; for Anthropic itself, it simultaneously represents performance obligations, resource dependence, and future profitability pressure. With the listing valuation already highly dependent on future growth, Anthropic needs to prove to investors that these expenditures are not merely spending to chase model scale, but long-term investments capable of generating sufficient commercial returns. Customer concentration and cloud platform dependence may limit revenue quality Another important risk for Anthropic comes from its revenue structure. According to Anthropic's prospectus seen by media, the company generated US$2.16 billion in revenue in 2025 through the cloud platforms of Amazon.com, Inc. and Alphabet Inc. Class C, accounting for 47% of full-year revenue. That proportion was only 11% in 2023 and rose to 32% in 2024, indicating that the company's dependence on cloud platform channels is gradually deepening. For Anthropic, selling through cloud platforms can significantly reduce the difficulty of customer acquisition. Many large customers purchase Claude services directly through cloud marketplaces, which can reduce contract, settlement, and technical deployment steps. Anthropic said in its IPO materials that these cloud platforms have sales networks that a single enterprise would find difficult to replicate independently, helping Claude enter the global enterprise market faster. However, the problem is that Amazon.com, Inc. and Alphabet Inc. Class C are not ordinary channel partners: they both invest in Anthropic and sell it the computing resources needed for training and inference, while also owning their own AI models and cloud computing businesses. This means these competitors can access some of Anthropic's pricing and commercial terms and may influence compute allocation, product promotion, and customer relationships. Anthropic also acknowledged in its prospectus that dependence on a small number of partners and suppliers may give rise to conflicts of interest and adversely affect the company's access to computing resources. High customer concentration does not necessarily mean there is a problem with the business model. For AI companies in a rapid expansion phase, large technology companies, enterprise customers, and cloud platform partners can often provide considerable procurement scale in the early stages and also help drive products into broader application scenarios. But when a small number of customers contribute a large share of revenue, the company's sensitivity to customer procurement decisions also increases. Once an important customer adjusts its procurement budget, shifts to in-house models, reduces call volumes, or demands lower prices in contract negotiations, Anthropic's revenue growth and profit margins could be affected. More complicated still, some large partners in the AI industry are simultaneously customers, compute suppliers, or investors. This relationship helps Anthropic obtain computing resources and commercial channels, but it may also leave the company facing a degree of dependence. If partners change commercial terms, adjust resource allocation, or devote more resources to promoting their own models, Anthropic will need to cope with potential cost and competitive pressure. Cloud platform channels also involve a difference between revenue recognition and economic returns. Sales revenue obtained through partners can expand a company's customer coverage, but a portion of it must be paid to cloud service providers. Investors therefore need to focus on revenue after deducting channel sharing, gross margin, customer renewal rates, and customer retention, rather than looking only at unadjusted revenue run rates. After listing, whether Anthropic can continue to expand its direct enterprise customer base, reduce dependence on a small number of customers, and improve the diversity of its revenue sources will become important indicators for measuring the resilience of its business model. Overall, Anthropic's upcoming IPO comes at a critical stage when the AI industry narrative is shifting from "capability breakthroughs" to "commercial returns." On the one hand, Claude and related products are driving rapid revenue growth, and enterprise customers' demand for AI tools provides the company with a realistic commercial foundation. If the listing goes smoothly, Anthropic is expected to gain new financing channels and establish a more transparent valuation reference for the entire industry. On the other hand, massive operating investment, long-term compute contracts, customer and supplier concentration, and AI safety risks all mean its valuation cannot be supported solely by long-term revenue forecasts. What the market ultimately needs to see is whether revenue growth can continue to profit and cash flow, and whether the company can control costs and risks while continuously enhancing model capabilities. Therefore, what is most worth watching about Anthropic's IPO is not whether it can set a new listing-size record, but how the public market will price the growth potential, capital consumption, and safety risks of frontier AI at the same time. If it can prove that heavy investment is translating into sustainable commercial returns, this IPO could become an important milestone in the AI industry's march toward maturity; if it cannot deliver on long-term growth expectations, it could also become a turning point for the market to re-examine AI valuations, capital expenditure, and financing models.