Anthropic Moves Towards Initial Public Offering Amid Financial Losses

Anthropic, a prominent player in artificial intelligence, has made its initial public offering (IPO) prospectus available to the public. The company, known for developing AI models such as Claude, generates revenue by providing access to these technologies for various users, including businesses and individual developers.

To create a sustainable commercial framework, Anthropic has established a range of products that function synergistically. The firm describes itself as a “research lab” that begins with AI model development, subsequently selling access to these models. Notably, Anthropic offers the Claude Platform, which allows developers to utilise these models for building custom applications and workflows.

In addition, Anthropic features ready-to-use applications like Claude Code, which assists with software development, and Claude Cowork, designed for a variety of professional tasks including research and analysis. This product suite encourages users to integrate Claude into their daily operations. As customers identify new applications for Anthropic’s offerings, their expenditure on its services tends to increase.

This growing adoption feeds into a continuous cycle where enhanced products draw in more users, thereby allowing Anthropic to further refine its offerings. The company derives income through two primary methods: first, a usage-based model where customers pay for the amount of resources consumed, often measured in tokens. Tokens refer to units of text processed by the AI models. Secondly, there is subscription revenue, where clients opt for recurring payments for predetermined levels of access.

In fiscal year 2025, which spans from January to December, Anthropic reported $4.5 billion in revenue from these sources, reflecting an increase of over 1,000% from the prior year. Revenue surged to $16 billion in the first half of fiscal year 2026, representing an approximate 1,300% growth compared to the same period in the preceding year. Currently, usage-based payments constitute about 80% of total revenue, with the remainder stemming from subscriptions.

Despite these impressive revenue figures, the company has yet to achieve profitability. In fiscal year 2025, Anthropic recorded an operating loss of $8 billion and a net loss of $41 billion. As of the first half of fiscal year 2026, while the operating loss decreased to $693 million, the net loss escalated to $80 billion, nearly double the net loss experienced throughout fiscal year 2025.

A principal factor behind this substantial net loss is the high operational costs associated with advanced AI models. In the first half of 2026 alone, the company incurred nearly $9 billion in computational and infrastructural expenses, alongside around $3 billion for employee salaries, including that of researchers, and payments to cloud service partners.

Despite holding a cash balance exceeding $130 billion, Anthropic is seeking to raise additional funds to support its ambitious growth ambitions and is not solely depending on investor financing. Another contributing factor to the large net loss is an accounting adjustment related to previous investments; as the company's valuation has risen, investors are entitled to shares that reflect this growth, thus requiring a corresponding adjustment in financial statements. In the first half of fiscal year 2026 alone, this adjustment added $72 billion to the net loss statement.

Nonetheless, the reduction in operating loss compared to the same period the previous year implies that Anthropic’s core operational economics may be improving, despite the alarming net figures.

As for the valuation that Anthropic seeks with its IPO, details regarding the IPO size and price range were not disclosed in the prospectus, making it challenging to assess its market worth accurately.

Turning to the risks associated with the IPO, Anthropic’s document outlines numerous potential challenges, including governance issues. Notably, purchasing shares does not guarantee significant influence over the company’s operations. Anthropic operates as a Public Benefit Corporation (PBC), and its governance framework restricts the control of ordinary shareholders, with major decisions being dominated by its founding members and a dedicated trust. Anthropic was founded by seven individuals, including Chief Executive Officer Dario Amodei and President Daniela Amodei, with an entity called Founder LLC maintaining 50.1% of the voting power, leaving common shareholders with only 49.9% of voting rights on shareholder matters. This unique structure is intended to focus on the responsible and long-term advancement of AI technologies rather than succumbing to short-term financial pressures.