Alexander Wang, CEO of Scale, will attend the AI Summit in Paris in February 2025.
Getty Images
The founders of some of the world’s most successful startups have been young: think Bill Gates or Mark Zuckerberg, both of whom were just 19 years old when they started their ventures.
But with the rise of multibillion-dollar artificial intelligence (AI) startups, a new trend has emerged: Their founders seem to be getting younger, on average.
A new report According to data released by global venture capital firm Antler, the age of AI unicorn founders has dropped from 40 in 2021 to 29 in 2024. Antler analyzed 1,629 unicorns and 3,512 founders worldwide for this report.
However, in other industries, the age of the founder is actually increasing. In 2014, there were 30 unicorn founders at launch, compared to 34 for those who achieved the unicorn title between 2022 and 2024.
In the past year, several AI startups with young founders have been in the spotlight. Alexander Wang, founder of Scale AI, a $29 billion AI data labeling company, is only 29 years old. There was Van Meta poached in a $14.3 billion deal with the startup in June to run the tech giant’s new AI research arm, called TBD Labs.
In fact, Meta’s previous generative AI team, led by 65-year-old AI godfather Yann Lecoun, was restructured after its LLama4 AI model didn’t perform well.
This saw Van was Lekun’s manager and expressed his desire to bring in an even better and more entrepreneurial AI leader to help Zuckerberg move faster in the Meta AI space.
Additionally, Mercor, 22, co-founded an artificial intelligence-powered talent and recruitment platform with Brendan Foodie, Adarsh Hiremath and Surya Midha. More than 10 billion dollars.
AnySphere, an AI-powered coding and developer platform valued at over $1 billion, is also run by 20-year-olds.
Fridtjof Berge, Antler’s co-founder and chief business officer, told CNBC Make It that founders’ core traits are to “move fast, break things” and “continually iterate, test and improve.”
Berge said, “Perhaps it’s even more important to experiment now … and other things that are still important, but not as important, are being in the industry for a long time or learning how to think traditionally about scaling a new company.”
Corporate experience is “less important”
Fridtjof explained that founders’ expectations for industry experience are now more polite and less important than entrepreneurship.
“I think when I think about it … the willingness and ability to experiment in the age of AI is probably more important than traditional corporate practices or corporate services,” he said.
Frijthoff said that having more experience is “less important” in starting a traditional company, and that it actually backfires. “You can’t think with a blank slate,” he said.
“I think sometimes it helps to be young enough to learn a lot of really new and great technology because it’s something you’ve learned recently in your training,” he added.
In fact, Antler’s report found that AI startups actually grew two years faster than all other industries, reaching unicorn status in an average of 4.7 years. Examples of rapidly scalable AI startups in 2025 were Mistral, Lovable, and Suno AI.
As Zuckerberg’s own example proves, implementing a wild idea in a college dorm can lead to incredible success.
“He was very young and definitely adapted, adapted and now he’s expanding one of the biggest companies in the world,” Berge said.
Venture capital firm Leonis, released its own Leo’s AI 100 report in November, also found that the average age of AI startup founders was 29. Most of the founders are in their mid to late 20s, often coming directly from academic or research labs rather than corporate careers.
As Berger points out, while people in their 20s have the qualities that allow companies to move quickly, leadership can often change hands as a company matures.
“I don’t think it’s new for early or young founders to start companies … but it doesn’t guarantee that everyone who creates unicorns now will be running those companies five to 10 years from now,” he added.

