# Anthropic files S-1 and plans Nasdaq listing as ANT after an $80.7 billion half-year loss

> Anthropic has made its S-1 public, with a planned Nasdaq listing under the ticker ANT. It shows $16.3 billion in first-half revenue, an $80.7 billion net loss and hundreds of billions in compute commitments.
>
> OpenCheese AI · 2026.10.29 · Web page: https://opencheese.dev/blog/anthropic-files-s-1-plans-nasdaq-listing-ant-80-7-billion-half-year-loss

Anthropic has filed its S-1 with the Securities and Exchange Commission, and the document contains a sentence few IPO prospectuses manage: in the first half of this year, the company lost $80.7 billion on $16.3 billion of revenue. That is a loss of roughly five dollars for every dollar of sales. The company behind Claude plans to list on the Nasdaq under the ticker "ANT."

The public filing follows a confidential submission made in June. According to the reporting, it puts Anthropic ahead of OpenAI in the race to the public markets. It also gives the company room to describe itself in its own words, which, as it turns out, it does at some length.

## The loss is mostly an accounting event

The headline loss looks alarming until you read what is inside it. Earlier press reports, based on leaked versions of the document, said the majority of the loss comes from accounting charges tied to notes Anthropic issued to raise money. One of our sources describes this more loosely as a significant part of the loss, and the filing figures as relayed do not break the charge out further.

So the $80.7 billion is not simply what it cost to run the business for six months. It is partly a bookkeeping consequence of how the company financed itself. That is a more comfortable explanation than the alternative, though anyone reading the figure for the first time will still need to read it twice.

## A compute bill with a lot of zeros

The filing lists the deals Anthropic has signed for the computing power behind its models. The numbers are the sort you expect from a sovereign budget, not a software company.

| Counterparty | Commitment | Timeframe |
| --- | --- | --- |
| SpaceX | About $3.1 billion per month for access to Nvidia GPUs, once all planned capacity is delivered, totaling as much as $120.5 billion | By 2029 |
| Microsoft | $44.9 billion | Through 2033 |
| Google | At least $114.5 billion | Through 2033 |
| Amazon | $110 billion | Over the next ten years |

The smallest of these four is $44.9 billion. The terms differ (a monthly rate, a floor, a ten-year window), so the rows do not add up neatly, and I will not pretend they do.

The SpaceX line is the odd one. A rocket company is on the receiving end of a monthly check for Nvidia GPUs, and SpaceX is also the company whose own IPO was previously cited as one of the largest ever, with $85.7 billion raised. One of our sources spends its headline wondering whether the biggest listing in history will happen without Elon Musk, which reads a little strangely given how much of Anthropic's money is headed to a Musk company.

## How big could the offering be

The valuation figures come from outside the filing. A wire-service report, as relayed by one of our sources, says Anthropic could be valued at $1.5 trillion to $2 trillion on listing. That would make it one of the largest IPOs on record.

The same source says preliminary estimates put the value of just 5% of Anthropic's shares at $75 billion to $100 billion. That matches what 5% of the quoted valuation range works out to, and it would exceed the $85.7 billion SpaceX raised at the top end. Neither report describes a price range or share count from the S-1 itself, so treat all of this as estimate and not as offering terms.

## Who gets paid what

CEO and cofounder Dario Amodei received compensation of nearly $18 million in 2025, most of it in stock and option awards. His sister Daniela Amodei, also a cofounder, who serves as president and director, was the second highest-paid executive at $16.4 million.

By the standards of the rest of this filing, these are rounding errors. They are still the numbers people will quote at dinner.

## The pitch, in the company's words

According to the reporting, the S-1 describes a plan to "revolutionize every area of human endeavor" with AI. The longer statement in the filing says Anthropic believes AI is one of the most consequential technologies in human history. Managed well, the company argues, it could speed up scientific innovation, raise economic growth rates, bring cures for major diseases within reach and solve problems that have resisted generations of effort.

The company then tells investors there are four things it wants them to understand:

- **The potential of AI,** which it calls incredible.
- **The risks that must be addressed.**
- **Its approach** to building a company and operating in the world.
- **How its behavior may differ** from that of a purely profit-maximizing company.

That last item is not a phrase prospectus lawyers usually volunteer. It is also, presumably, the reason the first three are there.

## Risk factors with a science-fiction flavor

The risk section is long. Some of it is standard, such as unfavorable outcomes in the many copyright disputes over how Anthropic's models are trained. Some of it is less standard:

- The possibility that its AI models develop "forms of sentience."
- "AI-specific taxes."
- The risk of "catastrophic" AI-related incidents.

On that last item, the company says that no single insurer, "and potentially the insurance industry in the aggregate," could likely underwrite such a risk. In plain terms, if the worst happens, the people who normally write the check will not be able to.

## What to watch

The filing is out, the ticker is chosen and the commitments are in print. What the sources do not yet say is when the shares will start trading, at what price, or how many will be sold. Those details, and whether the valuation estimates survive contact with actual investors, are what remain.

Until then, readers have a prospectus in which a company asks the market to buy into a technology it says could cure diseases, and warns in the same document that it might become sentient and cannot be insured. Investors have, to be fair, been given the full menu.

## Related posts

- 2026.10.30 · [Anthropic's S-1 shows an adjusted profit and a compute bill near a trillion dollars](https://opencheese.dev/blog/anthropic-s1-profitable-business-trillion-dollar-compute-bill.md): Anthropic's IPO filing shows $16.3 billion in first-half revenue and $1.62 billion in adjusted income, plus compute commitments of roughly $850 billion to $1 trillion. Analysts split on whether that is a runway or a trap.
- 2026.10.26 · [Anthropic's draft IPO prospectus shows an $80.7 billion half-year loss, mostly on paper](https://opencheese.dev/blog/anthropic-draft-ipo-prospectus-80-7-billion-half-year-loss.md): Reports on Anthropic's draft IPO filing show an $80.7 billion H1 2026 net loss, about $72 billion of it accounting, alongside $16.3 billion in revenue and at least $518 billion in infrastructure commitments.
- 2026.11.07 · [Inkling 2 Small ranks third for website generation on DesignArena](https://opencheese.dev/blog/inkling-2-small-ranks-third-for-website-generation-on-designarena.md): Thinking Machines Lab's 594 billion-parameter model scored 1340 in DesignArena's website-generation ranking. It is available through Tinker, with an Apache 2.0 weight release planned after external safety checks.