Anthropic going public: what do we know from the IPO filing

The company is testing investor appetite amid market volatility

Dario Amodei, Anthropic co-founder and CEO, speaking at TechCrunch Disrupt 2023
AI & BusinessLast Update 4 Min

Explore the interactive infographic

An interactive infographic i worked on in Tableau, with help from Claude, to explore the prospectus visually.

Open infographic ↗

Anthropic, the company behind the Claude AI models, is about to find out how much appetite Wall Street still has for AI. Its listing on Nasdaq, under the ticker ANT, is set to be the biggest IPO on record, according to Bloomberg.

It comes at a pivotal moment. AI stocks have been swinging hard, and more investors are openly asking whether the boom has turned into a bubble.

We will explore some of the main key highlights from the prospectus and how it reflects on investors and organizations. You can refer to the Anthropic IPO filing here.

Growth few companies have seen

Just after 3 years since it’s launch, Anthropic revenue reached $17.0 billion in the third quarter of 2026 alone. That is about 12 times the same quarter a year earlier.

More than 11,900 businesses now pay Anthropic over $100,000 a year, up from about 20 at the end of 2023.

202009501,50011,900+8x in nine monthsDec 23Dec 24Dec 25Sep 26
Source: Anthropic S-1 filing.

The bill is growing too

Most of that money goes straight back out. Training and running the models takes enormous computing power, and in the latest quarter Anthropic’s costs were still slightly higher than its revenue.

$376MQ1 25$787MQ2 25$1.4BQ3 25$2.0BQ4 25$4.7BQ1 26$11.6BQ2 26$17.0BQ3 26
Quarterly revenue
Total operating costs
Source: Anthropic S-1 filing.

The bigger number is what it has promised to spend. Anthropic has committed $900.9 billion to future spending, mostly on chips and data centres, with $70.7 billion due within a year. Its cash plus the money it expects from the IPO covers about a quarter of that.

Committed
$900.9B
Covers about 26%
Available
$231.2B
Leases not started $451.2B
Signed contracts $320.0B
Cash $95.6B
IPO proceeds $84.6B (est.)
Partner funding, up to $67.0B
Spending Anthropic has committed to, mostly computing capacity, against the cash and expected IPO proceeds it has available. Source: Anthropic S-1 filing.

The 1 Trillion Company

At the $74 midpoint of the price range, the company would be valued at around $968 billion, roughly 27 times its sales over the past year.

Retail investors would pay about 6.4 times what existing shareholders paid on average, and would own about 6% of the company. The founders keep just over half of the votes to control execution.

IPO buyers pay
$74.00
Existing holders paid, on average
$11.50
Book value per share after the IPO
$8.21
Source: Anthropic S-1 filing.

A very big market, on paper

The filing indicates knowledge work was worth more than $40 trillion worldwide in 2025, and quotes an IDC estimate that AI will add $22.5 trillion to the global economy by 2031.

$40T
Value of knowledge work worldwide in 2025 (the company's estimate)
$22.5T
IDC's estimate of AI's total economic impact by 2031
$35.3B
Anthropic's revenue over the last 12 months, under 0.1% of $40T

In my opinion, AI company is way overestimating their potential markets which is evident reading SpaceX prospectus that indicated similar outlandish claims. Anthropic revenue over the past year was about $35 billion, less than 0.1% of the headline number. Competition is fierce, cheaper models keep improving and even open models are becoming way more competitive and accessible. Adding to that, Anthropic has already cut prices, which cost it about $600 million in the third quarter.

Politics after the midterms

The timing adds another layer to the mix. Democrats won both the House and the Senate on 3 November, and Congress is expected to start with hearings on AI companies.

Voters in general are also pushing back on data centers as one of the main concerns during this election cycle. By the Harvard Ash Center’s count, 40 of 41 local ballot measures on data centers went in favor of limits. For a company that needs to rapidly expand its computing capacity, that matters.

What it means for businesses

Whether or not you are buying any shares, the filing shows some indicators where the next phase is heading. Anthropic is building versions of Claude for banks, healthcare, law, education and government, along with agents that handle research, drafting and coding on their own. Other AI companies might be already following up with the same plan.

Many businesses are already on board. About one in five U.S. companies now uses AI somewhere in its operations, according to the U.S. Census Bureau’s Business Trends and Outlook Survey.

Business uses can vary from offering new products and services, streamline existing processes and help employees get more done. For most organizations, the question is no longer whether to adopt AI. It’s what AI can actually contribute.

Where I stand as an investor

Personally, I am usually cautious about investing in any IPO. A listing is often an exit for early investors and staff, and about 15.5 billion Anthropic shares unlock 180 to 365 days after the debut. I’d rather see how the price holds up once that happens.

SpaceX is a useful reminder. It went public in June under the ticker SPCX amid similar excitement, and has since leaned heavily into AI, merging with xAI, the maker of Grok, and buying the coding tool Cursor. Its shares now trade about 30% below their post-IPO peak.

Long story short, the business is impressive. I just want to see a few quarters of it as a public company first.




This article is my personal opinion, written for educational purposes. It is not investment advice.

Cover photo: Dario Amodei at TechCrunch Disrupt 2023. Photo by TechCrunch, CC BY 2.0, via Wikimedia Commons.

Read More