Forbes AI 50 List: OpenAI and Anthropic Took 80% of the Money

Forbes dropped another list—AI 50 this time.

Honestly, when I saw that number, I was stunned—$242.6 billion. That’s the combined funding of OpenAI and Anthropic, accounting for 80% of the $305.6 billion total raised by this year’s AI 50 companies.

80%. Two companies took eight-tenths. That’s interesting.

Let’s look at the numbers first. This is Forbes’ eighth AI 50 list—50 companies total, 20 new faces. Selection criteria include technical capability, commercial potential, market performance. But honestly, seeing that funding data, suddenly “technical capability” feels a bit… pale in front of capital.

My sense is this list reflects two trends. One: AI’s Matthew effect is intensifying—the strong get stronger. OpenAI and Anthropic have formed a “duopoly,” making it increasingly hard for new players to break in. Two: AI investment门槛 is rising—the era of getting funding with just a demo is over.

Last week I chatted with a friend doing AI startup. He said investors ask increasingly “hardcore” questions: What’s your technical moat? Where’s your data coming from? Can your business model work? Three years ago, you didn’t need to answer these—just tell a good story.

It’s ironic. AI should be the most tech-focused industry, but now it’s the most capital-focused battlefield. OpenAI has Microsoft backing, Anthropic has Amazon and Google investing—combined resources probably exceed all of Silicon Valley’s VCs. New companies competing with them isn’t technically impossible, it’s resource-wise impossible.

That said, 20 new companies on the list shows AI still has innovation. I noticed one called Reflection, valued at $8B, doing AI safety products. Worth watching—as AI grows more powerful, ensuring it’s safe and controllable becomes a huge market itself.

One detail: Forbes notes many listed companies do “AI infrastructure”—compute scheduling, data management, model monitoring. Shows AI shifting from “model competition” to “ecosystem building”—not just whose model is stronger, but whose supporting infrastructure is more complete.

Honestly, my attitude toward these lists has changed. Used to study them carefully, thinking listed companies are good. Now it’s more like reading a “capital barometer”—who got the most money has the most say. Sometimes technical quality is secondary.

Final question: Do you think AI’s “duopoly” helps or hurts innovation? Does it concentrate resources for efficiency, or suppress new players’ growth?