Inside Anthropic’s IPO Filing: Rapid Growth, Rising Stakes

Anthropic’s proposed US stock market listing offers investors a detailed look at the business behind Claude, its family of AI models and applications. The company filed a registration statement with the US Securities and Exchange Commission on 26 October 2026, proposing to list on the Nasdaq Global Select Market under the ticker ANT. The filing sets out its financial performance, business model, ownership structure and risks, although key offering terms remain undisclosed.

Revenue has grown rapidly, and the latest quarterly results show Anthropic’s first operating profit in the periods covered by the filing. But the accounts also reveal a large reported net loss, significant financing arrangements and long-term commitments to secure computing capacity. Can Anthropic turn its rapid growth into sustained earnings and cash flow while meeting the costs of expanding its AI infrastructure?

The IPO: Confirmed Terms Versus Market Expectations

Anthropic’s preliminary S-1 leaves the proposed share price, number of shares and total fundraising amount undisclosed. Media reports have placed its potential valuation between USD 1.5 trillion and USD 2 trillion, with speculation that it could raise up to USD 100 billion. These figures are not confirmed IPO terms.

Anthropic’s reported May 2026 private funding round valued the company at USD 965 billion. That provides a reference point for the expectations surrounding the listing, but a private funding valuation does not automatically establish what public market investors will pay. The eventual valuation will depend on the offer price, the shares issued, the company’s financial outlook and the risks investors are willing to accept.

The next milestone is an amended registration statement, known as an S-1/A, which should disclose the proposed price range, share count and offering size. Until then, investors can assess the business and its financial commitments, but cannot determine the valuation implied by the final offering terms.

Revenue Growth: Enterprise Spending Drives the Business

Anthropic generated USD 16.286 billion in revenue in the first half of 2026, up from USD 1.163 billion a year earlier, representing growth of approximately 1,300%. The increase reflects wider adoption of Claude across software development, enterprise applications and business workflows, where usage-based pricing allows spending to grow as customers rely on the models for more tasks.

The filing reports net revenue retention of approximately 900% as at 30 September 2026, indicating substantial spending growth among existing customers. The metric includes expansion, contraction and churn, but excludes Claude Pro and Max subscribers because their fixed-price subscriptions do not capture the usage changes it is designed to measure. This figure is not the company’s overall revenue growth rate and does not guarantee that spending will continue at the same pace.

Enterprise adoption has also expanded sharply. As at 30 September 2026, Anthropic had more than 11,900 customers generating at least USD 100,000 in revenue over a trailing 12-month period, compared with more than 950 a year earlier. More than 1,900 customers generated at least USD 1 million over the same period.

The challenge is sustaining this momentum as the business grows. Consumption-based revenue remains sensitive to customer demand, pricing and competition, while large customers may reduce usage without being bound by long-term spending commitments. Anthropic’s growth is substantial, but its durability will matter more than the headline figures alone.

The USD 80.7 Billion Loss: Understanding the Accounts

Anthropic reported a net loss of USD 80.748 billion for the first half of 2026. Most of that loss came from financing-related accounting adjustments rather than the ordinary costs of running its AI business.

The largest item was a USD 72.388 billion loss arising from changes in the fair value of convertible notes, derivative liabilities and simple agreements for future equity. The increase in Anthropic’s valuation affected the accounting value of these financial instruments. A further USD 8.757 billion in other expenses contributed to the result, largely through the initial recognition and revaluation of other financial instruments.

Fair-value adjustments can affect reported earnings without creating an equivalent cash outflow in the same period. That distinction helps explain why the net loss is so large relative to revenue. However, the charges should not simply be dismissed. The financing arrangements behind them can affect future obligations, ownership and the company’s capital structure.

Operating results provide a different view of the business. Anthropic recorded an operating loss of USD 1.667 billion in the first quarter of 2026, followed by operating income of USD 974 million in the second quarter. Across the first half, it reported an operating loss of USD 693 million. Operating cash flow was positive at USD 3.929 billion over the six months.

The second-quarter profit marks a significant improvement, but one quarter does not establish a lasting trend. Investors will need to monitor operating expenses, stock-based compensation, capital requirements and financing arrangements alongside reported earnings.

The Cost of Scaling AI

Training and serving advanced AI models requires substantial computing capacity, much of it secured years in advance. Fortune’s aggregation of eight infrastructure service agreements and equipment leases with providers including Microsoft, Amazon and SpaceX puts their combined estimated value at USD 657 billion, of which USD 498 billion, or around 76%, is classified as non-cancellable. These figures are Fortune’s tally of individual agreements rather than a single total stated in the S-1, and the contracts differ in their terms.

The structure creates a potential mismatch between costs and revenue. Infrastructure obligations can extend for years, with some payments due regardless of how fully the capacity is used, while much of Anthropic’s revenue depends on customer consumption. If businesses reduce their usage, negotiate lower prices or move to competing models, revenue could weaken while a significant portion of the associated costs remains in place.

There are signs that the economics have improved as the business has scaled. Computing and infrastructure expenses fell from 201% of revenue in the first half of 2025 to 55% in the first half of 2026, and to around 48% in the second quarter. That does not mean infrastructure has become cheaper: these expenses rose 283% to USD 8.945 billion over the same period. The lower ratio reflects revenue growing faster than a cost base largely committed in advance.

Anthropic also identifies pricing pressure as a risk, warning that competition could force it to reduce the prices charged for AI model usage. Lower prices could stimulate demand, but they could also reduce revenue per unit of computing consumed. The outcome will depend on whether additional usage and efficiency gains offset that pressure as more contracted capacity comes online.

An Unusual Governance Structure

Anthropic is not structured like a typical listed technology company. Its seven co-founders will hold 50.1% of general voting power through a single Class F share held by a Founder LLC, while the Long-Term Benefit Trust has the power to appoint and remove a group of directors that will constitute a majority of the board.

Dual-class and controlled structures are not new. Several large US technology companies use them to retain founder control after listing. The practical effect is that outside shareholders may own a substantial economic interest but have limited influence over board composition and major decisions. Investors are effectively placing greater reliance on the judgement of the controlling group.

Anthropic is also incorporated as a public benefit corporation, requiring it to consider its stated public benefit alongside shareholders’ financial interests.

Google holds an option to purchase up to USD 10 billion of Anthropic stock through a private placement alongside the IPO, at the offering price. The arrangement adds another dimension to the relationship between Anthropic and one of its major investors and infrastructure partners.

Reading the Risk Factors

Anthropic’s risk factors include the possibility that AI-related harm could exceed available insurance capacity.

The filing also discloses a civil investigative demand from the Federal Trade Commission in October, following earlier subpoenas to AI companies.

Another key issue is Anthropic’s dispute with the US Department of War. The filing outlines the Pentagon’s supply-chain risk designation, a district court ruling largely in Anthropic’s favour in August, and a 2-1 appeals court ruling against the company in September. Anthropic has sought a rehearing before the full court. The company argues the designation was retaliation for its position on military uses of its technology, while the Pentagon maintains the dispute concerns operational control rather than speech.

What to Watch Next

Anthropic’s proposed IPO comes amid a broader wave of AI fundraising. Meta Platforms is reportedly pursuing a USD 95 billion equity raise, while SpaceX has reportedly secured USD 40 billion in debt financing to purchase AI chips. These deals highlight the scale of capital required to build AI infrastructure and the growing reliance on external funding to support long-term investment.

For Anthropic, the next step is the S-1/A, which should set out the price range, share count and offering size. Investors will also be watching management’s roadshow commentary, hyperscaler earnings for signs of sustained AI infrastructure demand, and developments in Anthropic’s legal and regulatory disputes.



Anthropic’s filing provides a clearer view of frontier AI economics, with rapid revenue growth set against substantial infrastructure commitments, financial complexity and regulatory uncertainty. How investors weigh these factors will shape the reception of the IPO and could influence the prospects for other AI companies seeking public market funding.