TL;DR
· SemiAnalysis's model predicts that Anthropic's adjusted operating profit will exceed $1 billion in the third quarter, but this is not official financial disclosure.
· Claude Code and enterprise clients are seen as sources of profit, with B2B pricing, Inference Cost, and renewal expansion decisions determining realization.
· Both Anthropic and OpenAI have secretly submitted S-1 filings, with no set date for going public; public financials will be the first litmus test.

Anthropic has pushed the IPO process to the SEC's doorstep, and market attention on the cutting-edge AI company has shifted from "who has the stronger model" to "who will make money first." According to Anthropic's announcement on June 1, the company has secretly filed a Form S-1 draft with the U.S. Securities and Exchange Commission for a proposed IPO, but the issuance size, price, and timetable have not been disclosed. SemiAnalysis's latest model provides a more aggressive assessment: Anthropic could achieve over $1 billion in adjusted operating profit in the third quarter of 2026. This figure is not from company financial reports or SEC data, but it has hit on the most sensitive issue in the AI capital market: whether a large model laboratory is simply a financing machine for ongoing compute purchases or is already close to being a publicly-listed software company.

$1 Billion Profit Forecast, Look at the Metrics First

SemiAnalysis's core assessment is that Anthropic's adjusted operating profit in the third quarter of 2026 could exceed $1 billion. This assessment comes from its bottom-up Tokenomics model, estimating revenue, token usage, inference costs, and profit across dimensions such as product, customer type, price tier, API calls, and enterprise subscriptions.

The key limitations are also clear. Without seeing Anthropic's public S-1, the external world cannot verify revenue, losses, equity incentives, cloud service costs, depreciation, and cash flow metrics in the formal documents. If SemiAnalysis's claims come from paid communication, they should also be seen as model predictions, not company-confirmed operating results.

"Adjusted operating profit" also cannot be directly equated with GAAP net profit. This metric typically excludes some non-cash expenses, one-time items, or equity incentives, making it more suitable for observing the business's operational status. For an AI company that heavily relies on compute procurement, cloud service agreements, and equity incentives to compete for talent, which costs are excluded from adjustments will directly affect the quality of profit.

In its official announcement on May 28, Anthropic stated that the company had completed a $65 billion Series H financing, reaching a post-money valuation of $965 billion. The announcement also mentioned that the company's annualized revenue had already exceeded $47 billion earlier in May. The high valuation and rapid revenue growth gave SemiAnalysis's model greater room for imagination, but what truly matters for the public market is still the revenue structure, gross margin, costs, and cash flow in the formal documents.

Why Betting on Anthropic to Make Money Before OpenAI

SemiAnalysis's bullish view on Anthropic focuses not on consumer-facing chatbots, but on the enterprise and developer market.

Claude Code is seen as the key product in Anthropic's commercialization efforts. It targets programmers and enterprise R&D teams, helping with code generation, debugging, refactoring, and software development processes. Compared to regular chat subscriptions, enterprise development scenarios have a stronger willingness to pay, a more stable budget, and usage frequency that is more easily converted into recurring revenue.

This is also why B2B revenue plays a key role in the model. Enterprise customers may pay per seat, API call, or workflow, or they may embed AI tools into their internal development processes. Once a dependency is formed, there is usually more room for package upgrades and price increases compared to consumer subscriptions. Therefore, SemiAnalysis believes that Anthropic's revenue quality, gross margin, and pricing power in the enterprise AI and coding scenes may be stronger than OpenAI's.

This is still an opinion rather than an official financial comparison between the two. OpenAI has a stronger consumer entry point, brand influence, and developer ecosystem, but it also has a more complex cost structure. The expansion into free users, consumer subscriptions, developer APIs, enterprise products, model training, and inference workloads all consume a significant amount of funds.

OpenAI also announced on June 8th that it had secretly submitted an S-1, emphasizing that the time of going public has not yet been decided and may take some time. Both companies are preparing to enter the public market, but whoever reveals their finances first will face more specific questions: whether revenue comes from high-visibility enterprise customers, if inference cost reduction can keep up with usage growth, how much additional capital is needed to train the next generation of models, and how the funding and computing power support from cloud providers will affect actual profit.

Nearing Trillion-Dollar Valuation, Putting AI Labs to the Financial Test

Anthropic's post-money valuation of $965 billion is approaching the trillion-dollar company threshold. For such a valuation, an IPO is not just a fundraising event, but more like a public examination of the AI lab's business model.

Secretly submitting an S-1 is a common process in the U.S. market. Companies can first communicate with the SEC on a draft registration statement, then decide when to make the filing public, start the roadshow, and list. As of now, Anthropic has only submitted a draft, which does not mean that the timing of the IPO has been set.

However, the capital market window will not stay open indefinitely. For Anthropic, submitting the filing early can allow the company to showcase a narrative of "high growth plus visible profit" to investors while AI funding is still hot and revenue is growing rapidly. For OpenAI, the same process will shift external focus from product releases, model capabilities, and user scale back to revenue, costs, and losses.

Over the past few years, leading AI companies have often been understood within the same loop: fundraising, acquiring compute power, training models, scaling inference workloads, and then continuing to raise funds. SemiAnalysis's pre-IPO chart for Anthropic offers another version, suggesting that large AI model companies may initially achieve profitability in the enterprise software and developer tools market, rather than staying in a prolonged capital consumption phase.

Whether this vision can materialize will depend on how the S-1 filing discloses key information. Factors such as revenue concentration among a few large customers, enterprise contract renewal rates, whether cloud service costs scale up proportionally with usage, and the extent to which equity incentives dilute profits will all influence the public market's evaluation of Anthropic's profitability.

A $6 Trillion Vision is Distant; S1 Is the First Hurdle

SemiAnalysis has also presented a more optimistic outlook, suggesting that if Anthropic continues to execute successfully, it has the potential to become the first $6 trillion company. However, this assessment appears more like an emotional upper limit rather than a current valuation basis.

Reaching this level would necessitate a series of conditions aligning: Claude Code maintaining a dominant position in developer workflows, long-term expansion of enterprise AI spending, continual reduction in inference costs, ongoing leadership in new models, and open-source models and other proprietary competitors not rapidly driving prices down.

Real-world constraints are also apparent. Will enterprise customers be willing to sustain high AI bills over the long term, requiring renewal and data-backed expansion? The larger the inference demand, the greater the pressures on cloud costs, chip supply, and data center capacity. If adjusted profits heavily rely on exclusions, the public market may discount the quality of earnings.

Both Anthropic and OpenAI have entered the confidential filing stage, but a true public listing is still some distance away. SemiAnalysis's $10 billion profit forecast has given Anthropic a strong starting point, but whether it can ultimately support a valuation close to a trillion dollars will depend on the transparency of revenue, costs, cash flow, and risk factors in the S-1 filing.

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia