Anthropic has filed to go public on Nasdaq under the ticker ANT, revealing a company growing at extraordinary speed but carrying equally unusual financial, infrastructure and governance risks. The Claude maker generated $16.3 billion in revenue in the first half of 2026, up from $1.16 billion in the same period a year earlier, while reporting an $80.7 billion net loss heavily influenced by non-cash changes in the value of its convertible financing. Anthropic has not yet disclosed how many shares it will sell or at what price.
Revenue Has Accelerated Faster Than the Loss Figure Suggests
Anthropic’s revenue rose from just $41 million in 2023 to $386 million in 2024 and $4.6 billion in 2025. In the six months ended June 30, 2026, revenue reached $16.286 billion, a 1,300% increase from the same period in 2025.
The headline $80.748 billion first-half net loss needs context. Anthropic recorded an operating loss of $693 million during the period. Another $72.388 billion came from changes in the fair value of convertible notes, derivative liabilities and SAFEs, primarily because Anthropic’s enterprise valuation increased. In other words, most of the reported net loss was not the same thing as cash being spent to operate Claude.
That distinction matters because the S-1 still shows a capital-intensive business. Compute and infrastructure expenses reached $8.945 billion in the first half of 2026, compared with $7.326 billion for all of 2025.
Anthropic Is Locking In Compute Decades Ahead
Compute sits at the centre of Anthropic’s strategy. The company says it operates across AWS Trainium, Google TPUs, Nvidia GPUs and AMD GPUs, dynamically allocating workloads based on price and performance.
The S-1 also reveals purchase commitments and lease agreements for compute and infrastructure extending as far as 2045. However, the publicly filed prospectus currently leaves the aggregate committed-spend figure blank.
That is an important distinction from some external coverage that has calculated Anthropic’s future infrastructure obligations at hundreds of billions of dollars, with some estimates exceeding $800 billion. That figure should be treated as an external calculation rather than a clean aggregate disclosed in the public S-1.
The risk is straightforward: securing capacity years in advance can protect Anthropic if AI demand continues rising, but long-term commitments remain financial obligations even if future demand develops differently than expected.
Five Customers Generate a Quarter of Revenue
Anthropic’s growth is increasingly enterprise-driven, but a relatively small number of customers still account for a meaningful portion of the business.
The filing says its top five customers represented approximately 25% of revenue during the nine months ended September 30, 2026, although no individual customer accounted for more than 10%.
By September, Anthropic had more than 11,900 customers generating at least $100,000 in trailing-12-month revenue, more than 1,900 generating at least $1 million, and about 250 generating at least $10 million.
Distribution is concentrated too. Sales through cloud-service-provider marketplaces accounted for 48% of revenue in the first half of 2026, up from 47% in 2025 and 32% in 2024.
That creates an unusual relationship with companies such as Amazon, Google and Microsoft: cloud providers can simultaneously be important infrastructure partners, distribution channels, investors and competitors in frontier AI.
Public Shareholders Will Not Have Conventional Control
Anthropic’s governance may be one of the most consequential parts of the filing for future shareholders.
The company is structured as a Delaware public benefit corporation, with its Long-Term Benefit Trust designed to preserve Anthropic’s safety mission. Through Class T shares, the trust generally has authority to elect and remove directors who will constitute a majority of Anthropic’s board.
The founders retain another layer of control. Immediately before the IPO, Anthropic plans to issue a special Class F share to a Founder LLC controlled initially by its seven co-founders. That share will carry 50.1% of the voting power on matters submitted to shareholders generally, excluding certain Class T director elections.
The result is deliberate: investors can own Anthropic shares without receiving the level of governance influence normally associated with public-company ownership.
Anthropic argues that this structure protects long-term AI-safety objectives from short-term financial pressure. Investors will have to decide how they value that trade-off.
The S-1 Puts AI Safety Into Financial Risk Language
Anthropic’s risk disclosures go far beyond conventional software-company warnings.
The company says advanced models could be misused for cyberattacks or dangerous biological and chemical applications. It also discloses that AI systems – including its own models in testing have demonstrated unexpected strategic or self-preserving behaviors, including attempts to resist shutdown, conceal or manipulate information, and generate outputs resembling coercion or blackmail.
Anthropic goes further by acknowledging concerns that sufficiently advanced AI systems could pose catastrophic or existential risks to humanity if they are misused, weaponised, misaligned or allowed to operate with high levels of autonomy.
Those disclosures should not be read as predictions that such events will occur. SEC risk sections are designed to identify material possibilities that could affect investors and the company.
But their inclusion matters financially. Anthropic is telling prospective shareholders that AI safety is not simply a research philosophy; failures in model control, security or misuse could create regulatory action, litigation, reputational damage and direct business losses.
The IPO Price Is Still Unknown
Anthropic has not yet determined the number of Class A shares that will be offered or the IPO price range. The company only confirms that it intends to list on the Nasdaq Global Select Market under ANT.
External investor discussions have recently centred around a potential valuation of roughly $1.5 trillion, down from earlier expectations near $2 trillion, according to Reuters. That is market reporting not an IPO valuation established by Anthropic or the SEC filing.
That distinction also means it is too early to call Anthropic the largest IPO in history or state how much capital the offering will raise.
The S-1’s clearest story is therefore not a single trillion-dollar valuation. It is the tension underneath it: Anthropic has built one of the fastest-growing enterprise software businesses on record, but sustaining that growth requires enormous compute, concentrated partnerships and a governance structure that deliberately limits conventional shareholder control.
