Forbes AI 50 2026: The Cold Truth Behind OpenAI and Anthropic’s 80% Funding Share

Honestly, when I saw this number, I was stunned.

$305.6 billion. That’s the total funding for all companies on this year’s Forbes AI 50 list. Then I looked further — OpenAI and Anthropic together: $242.6 billion. 80%. These two combined have more than the remaining 48 companies combined.

This number exceeded my expectations by a wide margin.

Let’s talk about the data itself.

Forbes released this 8th annual AI 50 list on April 21st, covering the world’s most promising private AI companies. OpenAI’s revenue allegedly exceeded $25 billion, and Anthropic’s run rate allegedly exceeded $30 billion — that word “run rate” is interesting. It means annualized figures, not actual quarterly revenue.

Both companies’ code products dominate the programming market. OpenAI has ChatGPT, Anthropic has Claude. It sounds like a worn-out story, but the real question is — why are these two companies able to capture 80% of the pie?

My first question: How deep is their moat?

Many people say it’s about data, compute, and talent — all three cards held by big tech. But the reality is far more complicated than this narrative suggests.

OpenAI’s advantage is first-mover and ecosystem lock-in. Millions of developers have built applications on OpenAI’s API, and switching platforms would be prohibitively expensive. This isn’t a technical moat — it’s an ecosystem moat.

Anthropic took a different path. Their Claude has earned quite a reputation for programming capability. Many independent developers around me say Claude’s code quality is more stable than GPT-4. But this is temporary — GPT-5 is coming soon, so who knows what happens next.

Second question: What does this 80% figure actually tell us?

Capital is voting with its feet. They believe these two will win. Everyone else is just running in the background.

But here’s an interesting paradox — if everyone believes these two will win, who’s actually doing the rest of the AI startups? The answer: either companies being acquired by these two, or companies doing vertical applications in niches.

This isn’t a healthy ecosystem. Historically, when capital concentration in an industry exceeds a certain threshold, innovation usually slows — because real innovation tends to come from the fringes, not the center.

Third question: Where are the Chinese players?

I didn’t count exactly how many Chinese companies are on this list. But something interesting is that AI investment approaches in China and the US are increasingly diverging. In the US, foundation model companies are sucking up capital; in China, application layers and vertical scenarios are more active.

This isn’t necessarily bad. It shows the Chinese market is seriously thinking about how to deploy AI, not just drawing pie charts.

Let me say something that might ruffle some feathers.

I’m not buying the narrative that “the AI industry structure is set.” Historically, there’s a reshuffle every few years. IBM missed the internet, Google missed social networking, and Facebook almost missed mobile. Saying OpenAI and Anthropic will win forever is like saying Google would win forever in 2015 — might be right, but not necessarily.

The more practical question: What does this mean for other AI startups? The answer: either get acquired by one of these two, or become extreme in some vertical field, otherwise it’s hard to survive independently.

This isn’t a special phenomenon in the AI industry — it’s the norm in the business world. The AI industry is just going through this phase early.

The interesting part is that capital’s judgment and technology’s judgment are often misaligned. Capital thinking someone will win doesn’t mean their tech path is right. These two things often move independently.

The bigger significance of this list might not be telling you who will win, but showing you — where is the water level in the AI industry right now, where’s the money, where’s the opportunity. As for winners and losers, that’s a different story.