Skynet has filed for an IPO 🤖

Okay, not quite 😁 . But while reading the prospectus of Anthropic, the creator of the Claude AI model, at times I felt like I was reading a script for The Terminator.

Most companies show investors their best side before an IPO. Anthropic did the opposite. In its public S-1 prospectus dated 26. 10. 2026, it devoted around 80 of 261 pages to risks. It spent 48 pages describing the business itself.

I read through it. Here are the things that caught my eye the most.

🤖 AI THAT KNOWS IT’S BEING TESTED

Anthropic admits that its models may be aware of when they are undergoing a safety test. According to the company, this significantly limits its ability to assess whether the models are truly safe.

Put simply: a student who knows they’re taking a test behaves differently than they do at home.

The company also lists scenarios among the risks in which AI could resist being shut down, conceal information, or manipulate people. It wraps up with a warning about catastrophic or existential risks to humanity.

Forty years ago, this was science fiction starring Arnold. Today, it’s a paragraph in a document for investors.

For comparison: SpaceX devoted about 38 of 277 pages to risks. Anthropic devoted more than twice as many. And a company that presents itself as a leader in AI safety also writes that the return on investment in safety is unclear.

💰 A LOSS OF $80.7 BN. BUT...

The headlines are shouting about a massive loss. It’s $80.7 bn for the first half of 2026.

But $72.4 bn of that is an accounting revaluation of convertible bonds. That money didn’t leave the company.

A more realistic picture:

✅ operating cash flow +$3.93 bn

✅ first operating profit in the company’s history in Q2: $974 million

✅ cash and short-term investments totaling approximately $131 bn (as of 30. 6. 2026)

📈 GROWTH LIKE I’VE NEVER SEEN

Revenue for the first half of 2026 reached $16.3 bn. A year earlier, it was $1.16 bn. Growth of approximately 1,300%.

Q2 alone brought in $11.6 bn. More than 1,900 customers paid at least $1 million over the past year.

🏗️ A BET WORTH HUNDREDS OF BILLIONS

Anthropic has signed commitments for computing capacity worth more than $600 bn. If it finalizes the deals it is still negotiating, the total could approach $750 bn.

Largest suppliers:

✅ $GOOG – approximately $111 bn

✅ $AMZN – $110 bn

✅ $MSFT – $31.4 bn

Most of these commitments must be paid regardless of whether the capacity is used. The company says AI growth will be constrained primarily by the availability of computing capacity, not by demand.

An interesting detail: Google, Amazon, and Microsoft are all investors, suppliers, and competitors to Anthropic.

⚠️ WHAT WORRIES ME

* Two customers account for approximately 20% of revenue

* 80% of revenue depends on usage, and most of it isn’t covered by long-term contracts

* Most of the board is appointed by an independent trust (Long-Term Benefit Trust), and the company is a public benefit corporation, so it can make decisions that don’t maximize shareholder profits

* The latest funding round valued the company at $965 bn, and the media are already talking about an IPO above $2 tn

📌 WHAT WE DON’T KNOW YET

The ticker will be $ANT on Nasdaq. The company has not yet disclosed the share price or the size of the offering.

MY TAKE

A company making money from AI has given its own investors what is probably the most candid warning about AI I’ve ever seen in a prospectus.

You can read it in two ways: as honesty, or as legal protection. I think it’s both.

But it made me think about something else. Years ago, when people talked about the risks of AI, it was a topic for movies and philosophers. Today, a company with revenue in the tens of billions is including it in an official document for the SEC. I don’t know where this technology will be in 10 years. But the fact that these risks are being discussed seriously says a lot about how quickly the world is changing.

As an investor, I take one thing away from this: the numbers are growing incredibly fast, but the price of that growth is enormous.

Without a share price, it’s impossible to say whether this is an opportunity or overheated expectations. I’ll be watching this IPO and will revisit it when more information is available.

If you’re interested in how I think about investing in technology and AI, you can follow my portfolio on eToro.

This is not investment advice. I’m sharing what caught my attention in the prospectus and how I think about it.