A pivotal showdown before the IPO! Anthropic is set to unveil its revenue figures, as its $2 trillion valuation faces a major test.
Against the backdrop of Anthropic's ongoing efforts to go public, these latest revenue figures are expected to provide the market with insights into its growth and profitability, help assess whether its proposed valuation of up to $2 trillion is justified, and potentially have a significant impact on AI‑focused trading in the tech sector.
Zhitong Finance APP has learned that, according to reports, Anthropic plans to hold a pre-IPO investor day at its San Francisco headquarters on October 14 and disclose its third-quarter revenue figures to a select group of institutional investors. Against the backdrop of Anthropic's ongoing efforts to go public, these latest revenue numbers are expected to provide the market with insights into its growth and profitability, help assess whether its proposed valuation of up to $2 trillion is justified, and potentially have a significant impact on AI‑focused trading in the tech sector.
Notably, Anthropic's disclosure comes at a time when investors are scrutinizing the financial performance of private AI companies more closely to assess the sustainability of growth across the industry. Last week, a report revealing OpenAI's revenue and accounting practices sparked confusion, dragging down the U.S. equity AI sector as a whole. OpenAI disclosed to investors that, as of the end of September, its annualized revenue stood at roughly $50 billion—lower than the $70 billion widely reported at the end of the previous month. Although sources familiar with the matter said this revised figure excludes revenue from cloud partners and does not reflect any changes in underlying demand, it remains unclear whether Anthropic will adopt the same accounting approach, namely whether it will include partner‑generated revenue in its calculations.
Anthropic's planned IPO is increasingly becoming the market's focal point. Recent reports indicate that the company could launch its IPO roadshow as early as the week of November 9, with trading potentially beginning before Thanksgiving on November 26. The firm is reportedly seeking a valuation of $2 trillion. If this target is achieved, Anthropic would surpass SpaceX's $1.77 trillion valuation at its June IPO, setting a new record for the largest IPO in history. It is important to note, however, that the $2 trillion figure should currently be viewed as a potential valuation under discussion, rather than an official offering price or a formal commitment from the company.
Anthropic, an AI upstart founded just five years ago, is surging from private unicorn status to the public markets at a pace that has left Wall Street astounded, driven most directly by the explosive growth of its financial metrics.
As of the end of July 2026, Anthropic's annualized recurring revenue (ARR) has surpassed $65 billion. This represents more than a sevenfold increase from approximately $9 billion at the end of 2025 and a further 38% rise from $47 billion in May. In the second quarter of 2026, the company's preliminary revenue exceeded $11.5 billion, compared with just $787 million in the same period of 2025—more than a 14-fold year-over-year jump. Even more encouraging for the market, Anthropic posted positive adjusted operating income in the second quarter, amounting to roughly $559 million.
This growth trajectory is nothing short of steep: at the beginning of 2024, ARR stood at roughly $87 million; by year-end 2024, it had climbed to about $1 billion; by the end of 2025, it reached approximately $9 billion; by April 2026, it surpassed $30 billion; by May, it hit $47 billion; and by late July, it surged to $65 billion. The most aggressive investor forecasts project annualized revenue of between $100 billion and $120 billion by the end of 2026.
Behind the explosive revenue growth lies Anthropic's clear commercialization roadmap. Unlike OpenAI, which relies on consumer‑level subscriptions, over 80% of Anthropic's revenue comes from API‑usage fees paid by enterprises and developers. The company now serves more than 300,000 enterprise customers, with roughly 70% of Fortune 100 firms having adopted its solutions at scale. In recent months, the number of customers with annual spend exceeding $1 million has doubled, surpassing 1,000. Its AI coding tool, Claude Code, is the primary growth driver: by early 2026, its annual recurring revenue (ARR) had exceeded $2.5 billion, and by August, estimates put ARR as high as $15.12 billion.
The substantial improvement in inference efficiency has further reinforced the positive feedback loop of the business model. Anthropic has adopted an API‑centric, pay‑per‑use pricing model, driving its overall gross margin into the mid‑60% range and lifting inference‑specific gross margins from 38% a year ago to between 70% and 85%. For cutting‑edge model companies that remain in a phase of heavy investment in computational resources, this marks a critical signal: the scale effects are beginning to translate into operating leverage.
The chief strategist at Bank of America Securities pointed out that a $2 trillion valuation implies expectations for 2028 revenue and requires growth rates to remain near triple-digit levels. According to reports, Anthropic internally projects 2028 revenues of roughly $190 billion to $200 billion. Based on $100 billion in revenue by the end of 2026, the $2 trillion valuation corresponds to a forward price-to-sales ratio of about 20 times, lower than the 50-plus times P/S ratios often seen among AI hot stocks like Palantir and Nebius. Some investors even remarked: "If Anthropic can sustain an 800% annual growth rate, even using a conservative 30x revenue multiple, the company's valuation would approach $3 trillion."
Notably, the flip side of rapid growth is substantial cost investment. In 2025, Anthropic's net loss approached $42 billion, roughly five times the approximately $8.3 billion reported the previous year. However, about $34 billion of this figure stems from changes in the fair value of convertible financing instruments—accounting losses arising from soaring valuations, rather than actual cash outlays. Excluding these adjustments, the company's underlying operating loss stood at roughly $8.06 billion, a 170% increase from $2.98 billion in 2024.
The widening loss is attributable to soaring upfront expenditures on computing power. In 2025, spending on computing power and infrastructure will reach $7.33 billion, equivalent to 1.6 times the company's revenue for that year. However, the loss margin is narrowing—operating loss as a percentage of revenue has declined from approximately 750% in 2024 to 176% in 2025.
A $2 trillion valuation reflects investors' strong confidence in the AI industry's long-term growth potential. Compared with traditional software companies, Anthropic aims not only to generate revenue through model subscriptions, enterprise services, and APIs, but also to position AI as a cornerstone for core business functions such as programming, customer service, knowledge management, and business process automation. As enterprises transition from pilot deployments to production‑grade implementations, AI‑powered solutions stand to evolve from auxiliary tools into integral components of corporate infrastructure.
In particular, the expansion of Claude‑series models into enterprise applications and programming has enabled Anthropic to secure more sustainable commercial demand. Once corporate clients integrate AI models into software development, internal knowledge retrieval, or routine operational workflows, they typically incur ongoing costs for model calls and related services. This type of demand is expected to generate strong revenue stickiness, providing a solid foundation for the company's future growth.
However, a high valuation also entails higher expectations for realization. Investors are not just paying for Anthropic's current scale of operations; they are also factoring in growth prospects over the next several years. Should revenue growth slow post‑IPO, profit improvements fall short of expectations, or the broader AI sector experience a broader valuation contraction, the company's stock could come under significant pressure. Consequently, whether a $2 trillion valuation is sustainable ultimately hinges on whether the market believes Anthropic can translate its technological leadership and robust demand into sustained, long‑term cash‑flow generation.
$518 billion in infrastructure commitments! The growth engine is also a source of financial pressure.
If revenue growth is at the heart of Anthropic's valuation story, then its massive demand for computing power and infrastructure stands as one of the most critical cost factors in this business model.
According to reports, based on Anthropic's IPO prospectus seen by the media, the company plans to incur or commit to roughly $518 billion in cloud‑computing and infrastructure‑related expenditures over the next few years. This scale underscores the company's highly aggressive outlook on future AI demand and suggests that its commercial expansion will be closely tied to cloud‑service offerings, data‑center construction, and the supply of advanced semiconductors.
It should be noted that this figure represents the associated expenditures and infrastructure commitments over the coming years; it does not equate to a one-time payment of the full amount in any single fiscal year, nor can it be simply equated with capital expenditures or actual cash outflows. Nevertheless, even accounting for differences in performance timelines and contract structures, such long-term commitments can still have a significant impact on a company's future financial flexibility.
The cost structures of AI companies differ markedly from those of traditional software firms. For conventional software, once product development is complete, the marginal cost of onboarding an additional user is typically quite limited; by contrast, large‑model services require continuous inference computations, and the more frequently users invoke the model and the more complex the tasks it processes, the higher the associated compute costs. For products like programming agents that can perform intricate tasks over extended periods, increased usage may both generate greater revenue and drive up consumption of computing resources.
Accordingly, whether Anthropic can achieve attractive profit margins in the future hinges on the concurrent improvement of several factors: enhanced model inference efficiency, reduced per‑unit computing costs, willingness among enterprise customers to accept higher pricing, and the company's ability to lower its compute‑procurement expenses through long‑term contracts and optimized resource allocation.
If the commercial value of AI models continues to rise and enterprise customers are willing to pay for higher productivity, substantial investments in infrastructure may become a necessary condition for establishing competitive barriers. Conversely, if model capabilities increasingly converge, service prices keep falling, or enterprise customers begin to cut their AI budgets, long-term commitments to computing power could turn into a heavy cost burden.
This also means that the market should not simply interpret the $518 billion infrastructure commitment as a positive signal for the chip and data‑center industries. For upstream suppliers, it may signal future demand; for Anthropic itself, it entails contractual obligations, resource dependencies, and mounting pressure on profitability going forward.
With its market valuation already heavily reliant on future growth, Anthropic must demonstrate to investors that these expenditures are not merely aimed at expanding model size, but rather constitute long-term investments that will generate sufficient commercial returns.
Customer concentration and reliance on cloud platforms may constrain revenue quality.
Another significant risk for Anthropic stems from its revenue structure. According to reports, Anthropic's prospectus reveals that in 2025, the company will generate $2.16 billion in revenue through Amazon and Google's cloud platforms, accounting for 47% of its annual revenue. This share was only 11% in 2023 and rose to 32% in 2024, indicating that the company's reliance on cloud platform channels is steadily increasing.
For Anthropic, the cloud‑platform‑based sales model helps the company significantly reduce customer acquisition costs. Many large customers procure Claude services directly through cloud marketplaces, streamlining contract negotiations, billing, and technical deployment. In its IPO filing, Anthropic noted that these cloud platforms boast sales networks that a single enterprise would find difficult to replicate on its own, enabling Claude to penetrate the global enterprise market more rapidly.
However, the issue is that Amazon and Google are not ordinary channel partners: they both invest in Anthropic and sell it the computing resources needed for training and inference, while also maintaining their own AI models and cloud computing businesses. This means these competitors can access certain aspects of Anthropic's pricing and commercial terms, and may influence its allocation of computing power, product promotion, and customer relationships. In its prospectus, Anthropic also acknowledged that its reliance on a small number of partners and suppliers could give rise to conflicts of interest and adversely affect the company's ability to secure computing resources.
High customer concentration does not necessarily indicate a flawed business model. For AI companies in a rapid expansion phase, large technology firms, enterprise clients, and cloud-platform partners often provide substantial procurement volumes early on, helping to broaden the adoption of their products across a wider range of use cases.
However, when a small number of customers account for a significant share of revenue, the company becomes more sensitive to their purchasing decisions. Should a key customer adjust its procurement budget, shift to in-house models, reduce usage volumes, or demand lower prices during contract negotiations, Anthropic's revenue growth and profit margins could both be adversely affected.
Adding further complexity, some of the AI industry's major partners serve simultaneously as both customers and providers of computing power or investors. While this arrangement helps Anthropic secure computational resources and access to commercial channels, it also exposes the company to a degree of dependency. Should these partners alter their business terms, reallocate resources, or prioritize the promotion of their own models, Anthropic would have to contend with potential cost pressures and competitive challenges.
Cloud‑platform channels also involve a distinction between revenue recognition and economic profit. Sales revenue generated through partners can expand a company's customer base, but a portion of it must be paid to the cloud service provider. As a result, investors should focus on revenue after deducting channel commissions, gross margin, customer renewal rates, and customer retention—rather than relying solely on unadjusted revenue run rate.
Following its IPO, whether Anthropic can continue to expand its base of direct enterprise customers, reduce its reliance on a small number of clients, and enhance the diversification of its revenue streams will serve as a key indicator of the resilience of its business model.
AI competition is accelerating; technological leadership does not necessarily translate into excess profits over the long term.
Anthropic's competitive landscape is equally formidable. Competition in the generative AI market stems not only from direct rivals like OpenAI, but also from in-house models developed by major tech firms, open-source models, and an ever-growing array of industry-specific solutions. As model capabilities continue to advance, enterprise customers are increasingly likely to weigh performance, cost, data security, system compatibility, and deployment expenses—rather than relying solely on the brand appeal of a single model.
For Anthropic, the Claude series of models' strengths in enterprise services and programming scenarios form a key foundation for winning customers. However, the model‑building industry is characterized by rapid iteration, meaning today's technological edge may not endure over the next few years. Even as the company continues to roll out new models, it must sustain ongoing investments in R&D, computing power, and talent to remain competitive. Reportedly, the prediction market is currently betting that Google (GOOGL.US) could overtake Anthropic and take the lead in the 2026 AI‑model race.
Moreover, improvements in model capabilities do not always align with commercial profitability. More powerful models can help companies tackle more complex challenges and boost customers' willingness to pay; however, if competitors offer comparable performance at lower prices, Anthropic may face pressure to cut its own prices. If inference costs decline faster than selling prices, profit margins could improve; conversely, price competition might erode the gains from increased efficiency.
The development of AI agents presents new growth opportunities for Anthropic. Compared with chatbots that merely answer questions, agents capable of performing programming, research, and business‑process tasks can deliver more tangible productivity gains for enterprises, while also fostering higher usage rates and deeper workflow integration.
However, agents may also entail higher computational costs, stricter reliability requirements, and greater security liabilities. The fact that companies are willing to pay for AI systems that can get the job done does not mean they are prepared to shoulder unlimited invocation expenses. Anthropic must demonstrate that the commercial value of its agent‑based products can offset their resource consumption and establish a replicable profit model.
Therefore, technological leadership is only one component of Anthropic's valuation. A more sustainable competitive advantage will also depend on factors such as customer relationships, integration into enterprise workflows, model reliability, model efficiency, and cost control.
Legal conflicts with the U.S. government add an unquantifiable policy risk to the IPO.
In addition to financial and competitive challenges, Anthropic also faces a unique risk closely tied to corporate governance, business development, and technological security: a dispute with the U.S. Department of Defense over restrictions on AI usage.
According to reports, Anthropic has petitioned the U.S. Court of Appeals for the District of Columbia to reconsider a previous ruling that upheld the U.S. Department of Defense's decision. Earlier, the court ruled 2–1 in favor of designating Anthropic as a company posing risks to the national security supply chain. At the heart of the dispute is Anthropic's refusal to allow its AI technologies to be used in autonomous weapons or for mass surveillance of the American public, while the Department of Defense contends that such restrictions could undermine military operations.
Anthropic, for its part, argues that the government's decisions exceed the scope of its legal authority and are motivated by retaliation against its AI safety stance. The company has previously stated that this determination has already resulted in billions of dollars in lost business. Meanwhile, another related legal dispute took a different turn in August: a federal judge in California overturned a parallel ruling based on a different statute. These developments underscore that the underlying legal issues remain complex and cannot be dismissed as settled.
The impact of this conflict on Anthropic extends beyond potential revenue from government contracts. Government determinations could affect its eligibility to participate in federal programs, its collaborations with government contractors, and certain customers' assessments of supplier compliance risks. Even if the company enjoys strong demand in the commercial enterprise market, policy-related risks may still alter some customers' procurement decisions.
Meanwhile, Anthropic's stance also reflects its brand positioning and long-term strategy. The company has consistently emphasized AI safety and adopted a relatively cautious approach to the risks and regulatory challenges posed by advanced AI systems. If it were to relax these safeguards in pursuit of government contracts, it could invite scrutiny over its corporate principles and commitment to safety; conversely, maintaining stringent restrictions might result in the loss of certain government contracts or expose the company to adverse policy implications.
From a capital markets perspective, this is a type of risk that cannot be readily assessed using traditional financial metrics. It can affect both revenue and the company's reputation, regulatory environment, and future business scope. Offering documents are required to disclose relevant litigation, potential losses, and their impact on the business, while investors must evaluate whether these disputes represent short-term uncertainties or structural factors that could materially shape the company's growth trajectory over the long term.
Anthropic's IPO will serve as a major stress test for the AI market.
Anthropic's IPO not only affects an AI company's fundraising and its shareholders' ability to exit, but could also serve as a crucial litmus test for the broader AI industry's business models.
If the company can continue to demonstrate rapid revenue growth following the disclosure of its financial results, while simultaneously improving profit margins, controlling infrastructure costs, expanding its customer base, and mitigating policy risks, its IPO could further bolster market confidence in the commercialization prospects of enterprise‑level AI and generate new demand expectations for the AI infrastructure and semiconductor industries.
However, if financial disclosures reveal that revenue growth is heavily reliant on a small number of customers, that cloud‑platform channel costs are elevated, or that long‑term computing‑power commitments place significant strain on cash flow, investors may reassess the company's valuation fundamentals. Moreover, if legal disputes with government authorities further disrupt operations, the market may demand a higher risk premium, reflecting greater uncertainty about the company's long-term growth trajectory.
For potential investors, what matters most about Anthropic is not any single valuation figure, but rather the interplay among revenue, costs, profits, cash flow, and risk. Rapid growth can demonstrate market demand, but it does not automatically prove that the business model has reached maturity; substantial investments in infrastructure may underpin future competitiveness, yet they can also exacerbate financial pressures if demand falls short of expectations.
Ultimately, Anthropic is grappling with the AI industry's most fundamental challenge: as large models transition from technological breakthroughs to large-scale commercial deployment, can companies sustainably generate economic returns while continuing to innovate?
The $2 trillion valuation reflects not only investors' expectations for Anthropic, but also a commitment that must be delivered through future performance. What will ultimately determine the success of its IPO is not how enthusiastic the market is about its AI vision, but whether the company can demonstrate that revenue growth can translate into stable profits, sustainable cash flow, and competitive barriers strong enough to sustain its long-term valuation.
Editor/Deng