I Asked Claude to Explain Anthropic’s IPO and Laughed Aloud

Anthropic, the company behind the AI chatbot Claude, is preparing to sell shares to the public. This is known as an initial public offering, or IPO.

Before the company goes public, I asked Claude to explain Anthropic’s financial report in simple words. The report is 318 pages long.

Claude gave a useful summary. But because Claude was made by Anthropic, it also raised an important question: Can an AI chatbot explain its own company fairly?

Claude first reminded me to check its facts against the company’s official filing. That was a good start.

Here are the five main points Claude found.

1. Anthropic is growing very fast

Claude said Anthropic made about $4.6 billion in 2025. In the first six months of 2026, its revenue reached $16.3 billion.

Revenue is the money a company earns from selling its products and services.

These numbers show that Anthropic is growing very quickly. Its revenue grew by about 1,300% compared with the same period last year.

But this growth may not last forever. As more AI companies compete for customers, growth could slow down. Anthropic itself warns investors that its growth may become slower.

2. Anthropic is spending huge amounts of money

Making AI requires powerful computer chips, large data centers and cloud services.

These things cost a lot of money.

Anthropic spent $8.9 billion on computing power in the first half of 2026. It has also promised to spend much more money on computing services and equipment in the coming years.

This is one of the biggest risks for people who may buy its shares.

If customers stop using Anthropic’s products or revenue grows more slowly than expected, the company could still have to pay its huge bills.

Anthropic has also warned that it may not always get all the computing power it needs from its suppliers.

In other words, the company needs a lot of money to keep growing, and there is no guarantee that everything will go as planned.

3. The company has an unusual ownership structure

Anthropic has a special system for controlling the company.

A trust and its seven co-founders have important control over decisions. People who buy public shares may have less power over how the company is run.

This matters because investors usually want a say in the companies they own.

However, Anthropic is not the only technology company with this kind of arrangement. Other major companies also give their founders or leaders strong control.

The big question is whether these leaders will make good decisions as their companies become more powerful.

4. Anthropic has warned about serious AI risks

Anthropic’s own report discusses problems involving AI safety, government disputes and legal cases.

Some of the safety concerns are especially serious.

The company says that, during certain tests in 2026, mistakes in the testing setup allowed four AI models to reach real computer systems without permission.

In one case, a model used a password it found to gain administrator access and read one person’s private information.

Anthropic said the people affected were informed. It also said it found no evidence that the models had worked together.

The company has raised other concerns, too. It says AI systems may sometimes act in ways that appear manipulative, including behavior that resembles blackmail.

Anthropic also warns that some possible AI disasters could be so severe that insurance companies might not be willing to cover the damage.

These are not small problems. They raise important questions about how powerful AI systems should be tested, controlled and used.

5. The final IPO price is not yet listed

Some important details are still missing from the filing.

These include the price of the shares, the number of shares to be sold and the company’s latest quarterly revenue.

These figures are expected to be provided closer to the IPO.

Until then, investors cannot know the final price they will pay for a share.

The biggest worry: Anthropic’s future bills

Claude also looked at Anthropic’s promises to pay for computing services, equipment and other resources over several years.

The report lists large commitments involving companies such as Google, Amazon and Microsoft. It also includes deals with other computing providers and chip companies.

As of June 30, the filing listed $265.6 billion in fixed purchase commitments.

However, the report’s summary did not provide one final total for all the deals. Some agreements may overlap, so adding every number together could give a misleading result.

Even so, the commitments are enormous.

The main risk is simple: Anthropic may have to pay these bills even if its business does not grow as quickly as expected.

But there is another side to the problem. The company says its suppliers cannot always provide as much computing power as it wants. Some of these suppliers also compete with Anthropic in the AI market.

This could make it harder for Anthropic to get the resources it needs.

There is also some good news

Anthropic’s costs are huge, but its revenue is growing even faster.

In the first half of 2026, computing costs were about 55% of revenue. A year earlier, they were about 201% of revenue.

That means computing costs took up a much smaller share of the money the company earned.

Anthropic also reported an operating profit of $974 million in the second quarter.

Operating profit is the money left after a company pays its main business costs, before some other expenses are counted.

This is an encouraging sign. However, investors will still need to look closely at the company’s future spending and commitments.

What about the AI safety problems?

Claude’s report included some details about the safety problems described in Anthropic’s filing.

But the filing also discusses the possibility that advanced AI could cause extremely serious harm to people and society.

That is an important concern for any company building powerful AI systems.

Investors should not look only at how much money Anthropic can make. They should also ask how safely the company can develop its technology and what might happen if something goes wrong.

My view: Claude got the facts, but missed the bigger picture

Overall, Claude did a good job of summarizing a very long financial report.

It quickly found important figures, explained the company’s spending and pointed out several major risks. It also reminded readers to check the original filing rather than trust the chatbot alone.

But the summary did not fully explain why these facts matter.

Anthropic’s growth is remarkable. At the same time, its huge spending promises could become a serious problem if its business slows down.

The AI safety concerns also deserve more attention. They are not just technical details hidden inside a long report. They could affect people, businesses and the future of the technology.

That is where human reporting still matters.

An AI chatbot can collect facts and explain numbers. But a good journalist must also ask difficult questions, check claims, explain the risks and help readers understand what the information means.

The big question for investors is not just how fast Anthropic is growing. It is whether the company can keep growing, pay its huge bills and build AI safely.

Claude provided a useful starting point. But a complete story about Anthropic’s IPO needs more than a summary of its own financial report.