---
title: "What Does Anthropic's IPO Filing Reveal About Its Finances?"
url: "https://goremy.ai/blog/articles/anthropic-ipo-filing-revenue-compute-commitments"
author: Luis Chavez-Mattos
editor: Luis Chavez-Mattos
editor_title: Director, Product
published: "2026-10-29T16:20:07.893Z"
updated: "2026-10-29T16:20:07.893Z"
topic: AI Industry & Strategy
description: "Anthropic's IPO filing cites $16.3B in half-year revenue, up to $750B in compute commitments and a class F share with 50.1% of votes. Key figures."
---

# What Does Anthropic's IPO Filing Reveal About Its Finances?

> Anthropic's IPO filing cites $16.3B in half-year revenue, up to $750B in compute commitments and a class F share with 50.1% of votes. Key figures.

Edited by Luis Chavez-Mattos, Director, Product · Published 2026-10-29

## What does Anthropic's IPO filing reveal about its revenue, spending and governance?

According to figures quoted from the filing, Anthropic's revenue rose from $41 million in 2023 to $386 million in 2024 and $4.6 billion in 2025. It then hit $16.3 billion in the six months ended June 30, 2026, a 1,300% year-over-year increase. The filing also cites compute commitments of up to roughly $750 billion, says over 90% of codebase contributions were written by Claude, and creates a single class F share carrying 50.1% of the voting power.

A caveat: these numbers come from a creator's walkthrough that quotes the filing. They were not checked against the primary document here, so confirm them against the filing itself before relying on them.

## TL;DR

- Revenue went from **$41 million in 2023 to $16.3 billion in the first half of 2026**, with the filing citing 1,300% year-over-year growth for that period.
- On the creator's comparison, Anthropic reached **$10 billion in revenue in about three years**, versus roughly 9 for OpenAI, 8 for Google, 10 for Meta and 12 for Amazon.
- Anthropic is described as securing **compute commitments of up to about $750 billion** over the next few years.
- The filing says **over 90% of contributions to the codebase** in September 2026 were written by Claude.
- Internal Claude usage valued at list prices grew **more than 250 times between January and September 2026**, and is said to rival or exceed personnel costs.
- A single **class F share equals 50.1% of total voting power**, which keeps control with the co-founders regardless of new investment.

## How fast has Anthropic's revenue grown?

The filing, as quoted, gives these annual figures and one half-year figure:

| Period | Revenue |
|---|---|
| 2023 | $41 million |
| 2024 | $386 million |
| 2025 | $4.6 billion |
| Six months ended June 30, 2026 | $16.3 billion |

The half-year number is the striking one. Anthropic booked more than three times its entire 2025 revenue in six months. The filing frames the period as 1,300% year-over-year growth.

The creator also compared time to $10 billion in revenue. His figures: Anthropic about three years, OpenAI nine, Google eight, Meta ten, Amazon twelve. These are his comparisons, not statements from the filing, and the definition of "years to $10 billion" varies by company (founding date, first revenue, product launch). Treat the ranking as directional. The gap is large enough that the direction is not in doubt, but the exact multiples depend on where you start the clock.

## Why compare Anthropic to old software companies at all?

The creator argues you cannot judge Anthropic with the yardstick used for earlier companies. Critics point to losses and debt and say the company makes no sense on paper. His counter is that the economics differ: a product that behaves like software (fast growth, high demand) but also like a utility (near-universal need, heavy infrastructure spend).

That is an opinion, not a finding from the filing. It is a useful lens, though. A company whose main cost is compute capacity looks more like a utility or cloud provider than a classic SaaS vendor. Investors will likely weigh margins, commitments and customer concentration, not just growth rate.

## How big are Anthropic's compute commitments?

The figure quoted is up to roughly three quarters of a trillion dollars, or $750 billion. The creator describes these as contracts that obligate Anthropic to pay that amount over the next few years. His explanation for the scale is that Anthropic expects compute to become its main bottleneck.

Some points to keep in mind when reading a number like this:

- A commitment is an obligation to pay, not money already spent.
- "Up to" means the total depends on terms, options and usage thresholds that the summary does not detail.
- The size relative to revenue is the central risk and the central bet. Anthropic is betting that demand keeps growing fast enough to cover the bill.
- The creator also notes the filing's risk language about depending on a limited number of compute providers, though that exact wording was not read out as a verified quote.

## What does "90% of code written by Claude" mean?

The filing is quoted as saying: "Over 90% of the contributions to our codebase in September 2026 were written by Claude." That is a statement about contributions to Anthropic's own codebase, not about the industry.

It matters for two reasons. First, it shows how far a frontier lab now uses its own product in production engineering. Second, it ties to the spending story. If Claude does most of the coding, then Claude usage is a core input to how the company operates, and compute demand rises even from internal use.

The summary does not say how "contributions" are measured, or how much human review and direction sits around the generated code. Those details decide how to read the 90%.

## Is internal Claude usage really costing more than staff?

The creator reports that the value of Anthropic's internal Claude usage at list prices grew more than 250 times from January 2026 to September 2026, and that it now exceeds what the company pays its staff. His phrasing of the exact comparison point is muddled in the source, mentioning April 2026 as the point where usage approximately equaled personnel expense. So the safest reading is this: internal usage, valued at list prices, rose over 250x in nine months and reached roughly the scale of personnel costs.

Two cautions apply. "At list prices" is not the same as actual cost to Anthropic, which pays for its own compute, not retail rates. And the creator's point that Anthropic salaries are high (he mentions listings in the range of a quarter million to half a million dollars a year plus equity) is his observation, not a filing figure. Still, the direction is clear. Usage by the company itself is now a major line item in how it thinks about its economics.

## What is the class F share and why does it matter?

The filing creates a single share of class F common stock. As quoted, it carries an aggregate number of votes equal to 50.1% of total voting power. The creator says it is held collectively by the co-founders.

The effect is simple. Whatever new equity Anthropic sells, however many shares it issues, that one share keeps a majority of votes. Outside investors can buy economic exposure through class A stock, but they cannot dilute the founders' control. The creator links this to the founders seeing themselves as stewards of the technology.

Dual-class and special-share structures are not new in tech IPOs. What stands out here is the mechanism: one share, fixed at a majority of the vote, instead of a multiple-votes-per-share design.

## What should builders and investors take from this?

For people building with Claude, the filing signals a company that expects to spend heavily on capacity and to keep shipping fast. That points toward continued model releases and continued pressure on pricing and rate limits as demand meets supply, though the filing summary does not say anything specific about either.

For investors, the open questions are the ones any analyst would ask:

1. How much of the $750 billion in commitments is firm, and on what schedule?
2. What are the margins on that $16.3 billion once compute costs are counted?
3. How durable is growth at this pace, and how concentrated is it among a few customers?
4. What does founder voting control mean for outside shareholders in practice?

The creator's more sweeping predictions, such as a future valuation in the quadrillions or Anthropic owning energy and chip makers, are personal speculation. Nothing in the quoted filing supports them. Neither is it financial advice.

## Frequently Asked Questions

### How much revenue did Anthropic report?

The quoted figures are $41 million in 2023, $386 million in 2024 and $4.6 billion in 2025. For the six months ended June 30, 2026, revenue was $16.3 billion, which the filing describes as 1,300% year-over-year growth.

### How big are Anthropic's compute commitments?

They are described as up to about $750 billion, payable over the next few years. These are contractual obligations, so they represent future spending rather than money already paid. Details on timing and terms were not part of the summary.

### What is Anthropic's class F stock?

It is a single share of common stock whose votes equal 50.1% of total voting power. The creator says the co-founders hold it collectively. It lets them keep control even as new investors buy in.

### Did Claude really write over 90% of Anthropic's code?

The filing is quoted as saying over 90% of contributions to its codebase in September 2026 were written by Claude. The summary does not explain how contributions are counted or how much human review is involved, so read it as a company statement about its own repository.

### Are these numbers verified?

They come from a creator quoting the filing, and some phrasing in the source was unclear, especially around internal usage costs. Check the filing directly before using any figure in research or investment decisions.