Forbes AI 50 2026: OpenAI and Anthropic Dominate, But New Players Are Rising

I check Forbes’ AI 50 list every year. This year’s edition shows a striking change—OpenAI and Anthropic together account for nearly half of the total funding across the entire list.

$242.6 billion.

That’s a scary number. The AI 50 companies combined have raised just over $500 billion total; these two alone take half.

But more interesting: there are 20 new faces on the list. What does this mean?

The barriers to AI entrepreneurship are rising, but the window of opportunity hasn’t fully closed. The game has just changed.

I looked at what these new companies focus on and found a pattern: vertical AI applications are on the rise. Legal AI, medical imaging, architectural design—these companies aren’t trying to ‘build a general-purpose large model,’ they’re ‘using large models to solve specific problems.’

This aligns with my assessment.

I wrote before that large model entrepreneurship has entered ‘deep waters’—the window for foundation models is basically closed, but the application layer is just getting started. This year’s AI 50 list confirms that trend.

Another notable phenomenon: hardware companies are increasing.

Several AI chip and edge computing device makers made the list. This reflects market focus on compute costs—when inference costs become a key variable in business models, companies offering more efficient compute solutions become valuable.

I also noticed Chinese companies’ share rose from 6 to 9 this year. While US companies still dominate, China’s AI ecosystem remains globally competitive.

But one data point surprised me.

The fastest-growing companies by valuation aren’t foundation model builders—they’re ‘AI infrastructure’ providers: data labeling, model fine-tuning toolchains, deployment platforms. What does this tell us?

As AI applications proliferate, the ‘picks and shovels’ business proves more stable than ‘gold mining.’ We saw this pattern in the internet era; now it’s playing out again in AI.

Finally, a personal observation.

Looking at this list, my biggest takeaway—the AI industry is transitioning from ‘hype phase’ to ‘value delivery phase.’ Most listed companies have clear monetization paths and revenue data, not just ‘we have a cool AI model.’

That’s good for the industry. After the bubble deflates, the real winners emerge.

For entrepreneurs entering AI: don’t focus on foundation models. Find a specific problem that large models can solve, and go deep. The new faces on this list mostly took that path.