Anthropic Files for IPO Revealing $16B Revenue and $80B Loss
The Claude maker is heading to Nasdaq under ticker ANT with $16.3B in half-year revenue, an $80.7B loss, and a potential $2T valuation.
- Anthropic publicly filed its S-1 to list on Nasdaq under ticker ANT.
- H1 2026 revenue hit $16.3B, with a reported $80.7B net loss from convertible note markups.
- Goldman Sachs and Morgan Stanley lead the offering; share count and price not yet set.
- Reports point to a valuation matching or exceeding SpaceX's $1.8 trillion, with marketing starting Nov 9.
- Prospectus discloses roughly $518B in compute obligations to Google, Amazon, Microsoft, Broadcom and SpaceX.
- Last private valuation was $965B in a May 2026 Series H round.
Anthropic has filed its S-1 publicly with the Securities and Exchange Commission after submitting a confidential draft in June 2026. The AI company plans to list Class A common stock on the Nasdaq Global Select Market under the symbol ANT.
The preliminary prospectus leaves the share count, price range, proceeds, valuation and first trading date unset. Goldman Sachs and Morgan Stanley are leading the offering.
Why net income looks so extreme
Anthropic reported $16.3 billion in revenue and an $80.7 billion net loss for the first half of 2026. Most of the loss came from the rising fair value of convertible notes as the company’s private valuation increased.
Fair-value accounting can produce a large expense when convertible securities are recorded as liabilities. As Anthropic’s valuation rises, the amount owed to noteholders increases on paper, and the company recognizes that change through net income. Operating cash flow and operating loss offer clearer measures of cash consumption than the headline net loss.
Revenue has accelerated sharply from approximately $4.6 billion for all of 2025. The first-half 2026 figure exceeds that total by more than three and a half times. Enterprise adoption of Claude and demand for the Claude Code developer tool have driven much of the growth.
The $65 billion revenue run rate cited in reports around the offering is a point-in-time annualization, usually based on a recent month or quarter. It exceeds the $32.6 billion annualized pace implied by first-half revenue, so the measurement period matters when comparing Anthropic with competitors.
The $518 billion compute question
Media accounts of the filing describe approximately $518 billion in planned cloud, hardware and infrastructure commitments. The reported supplier amounts are substantial:
| Supplier | Commitment |
|---|---|
| $111.1 billion | |
| Amazon | $110 billion |
| Microsoft | $31.4 billion |
| Broadcom | $161.2 billion |
| Total for named suppliers | $413.7 billion |
The four named commitments total $413.7 billion, leaving $104.3 billion between that subtotal and the reported $518 billion. CNBC also reported a separate commitment to pay SpaceX $3.1 billion per month for compute, equivalent to $37.2 billion a year. The filing’s footnotes need to clarify the contract periods, potential overlap, minimum purchase requirements and termination rights before the totals can be reconciled.
Long-term compute agreements spread payments across several years and may secure capacity that Anthropic expects to resell through model access. They still create fixed obligations that can outlast demand forecasts. A public listing would give Anthropic another source of equity, acquisition currency and potential borrowing capacity, although the preliminary filing does not disclose how much capital the offering would raise.
Valuation outruns the deal terms
- May 2026: Anthropic’s Series H financing valued the company at $965 billion.
- June 2026: The company submitted its confidential S-1.
- November 9: Bloomberg reported that offering marketing could begin on this date.
- Week of November 16: Reports identified this period as the target for the stock-market debut.
Bloomberg reported that Anthropic was considering an IPO valuation at or above $1.8 trillion, the figure associated with SpaceX’s planned offering. Other reports have placed the upper end near $2 trillion. Reuters calculated that such a valuation would put Anthropic among the world’s 10 most valuable public companies. Those figures remain reported ambitions because the S-1 contains no proposed valuation.
Anthropic has raised approximately $118.15 billion across nine primary funding rounds. That fundraising cadence reflects the capital required to train models, reserve data-center capacity and prepay infrastructure suppliers.
Cloud vendors own part of their customer
Amazon and Google combine major equity positions in Anthropic with large contracts to supply its computing infrastructure. Amazon’s stake has been estimated in the mid-to-high teens, implying a value of roughly $135 billion to $160 billion at the $965 billion private valuation. Amazon recorded $16.8 billion in pretax gains from Anthropic valuation increases during the first quarter of 2026, while Anthropic has committed more than $100 billion to Amazon infrastructure over the next decade.
Google holds a smaller stake and has its own multiyear compute agreement. The structure sends invested capital back to cloud vendors through infrastructure purchases. Anthropic records the resulting costs, the suppliers record revenue, and their equity stakes add further exposure to Anthropic’s valuation. Public filings from all three companies should make those flows easier to track.
Two AI labs approach the market
OpenAI reportedly submitted a confidential S-1 shortly after Anthropic’s June filing. Public investors will therefore compare two frontier-model developers with different product mixes: Anthropic has concentrated on enterprise deployments and coding tools, while OpenAI retains a broader consumer footprint through ChatGPT.
| Metric | Anthropic | OpenAI |
|---|---|---|
| Flagship product | Claude | ChatGPT |
| Estimated revenue run rate | About $65 billion | About $40 billion |
| Last reported private valuation | $965 billion | $852 billion |
| Reported IPO valuation discussion | Up to $2 trillion | Up to $1 trillion |
The run rates and prospective valuations come from reported estimates rather than audited annual results or stated offering terms. Differences in measurement periods, revenue recognition and cloud-contract structures may limit direct comparisons.
The gaps investors will price
The preliminary S-1 omits the offering size, target valuation and an estimate of Anthropic’s total addressable market. Registration statements do not require a market-size estimate, but the omission leaves investors to build their own assumptions about demand for model training, inference, coding agents and enterprise AI services.
Several disclosures will determine how the offering is valued:
- The payment schedules, cancellation provisions and minimum spending requirements attached to compute contracts.
- Operating cash flow, consolidated gross margin and the cost of serving additional model requests.
- Revenue concentration, customer retention and the share of sales tied to Claude Code.
- Dependence on individual cloud, chip and networking suppliers.
- Regulatory, copyright, security and frontier-model safety exposure.
- Voting control and dilution across Anthropic’s share classes.
The market backdrop adds financing risk. Shares of infrastructure providers CoreWeave and Nebius fell more than 30% in October amid concerns about data-center leverage and the pace at which AI demand will produce cash returns. Anthropic’s reported commitments expose it to the same questions on a much larger contractual base.
What Claude customers can monitor
Existing Claude API and Claude Code terms remain in force unless Anthropic announces changes. The registration process itself does not alter service-level agreements, prices, rate limits, data-use terms or model-deprecation schedules.
SEC reporting should improve vendor assessments by providing regular information about liquidity, consolidated margins, infrastructure commitments and supplier concentration. Those reports may still stop short of revealing profitability for individual model families or products.
- Capacity: Reserved compute could support higher limits and more predictable availability, while supplier concentration leaves Anthropic exposed to outages and contract disputes.
- Pricing: Fixed infrastructure commitments and public margin targets may influence API rates, volume discounts and free-tier limits.
- Roadmap: Enterprise revenue may direct more investment toward security controls, governance, reliability and administrative tooling.
- Model lifecycle: Customers should track notice periods for model retirement, migrations and changes to safety policies.
- Procurement: Quarterly reports can support counterparty reviews, although customers will still need contractual protections for continuity and data handling.
Anthropic’s S-1 reportedly retains extensive language about the long-term risks of frontier models. Those policies can affect model availability, permissible uses, evaluation requirements and release timing, making the safety disclosures operationally relevant to teams building on Claude.
The eventual price range will show how public investors value rapid AI revenue growth alongside exceptionally large infrastructure obligations. For Claude customers, the practical effects will emerge through capacity, pricing, product policy and the financial disclosures that follow each quarter.