One of the most respected AI investors just gave founders a warning that doesn't sound like marketing. Elad Gil told the "No Priors" podcast that most AI startups are only going to hit peak value for about 12 months. After that, the business "crashes out."
For investors holding AI stocks or looking at private rounds, that's a pricing signal.
The Logic Behind The Window
Most AI startups exist because OpenAI, Anthropic, and Google haven't yet decided to eat their category. When those companies ship a feature, the startup's moat usually disappears overnight.
Gil's point: founders often see this coming six months before the market does. But they don't sell. They wait for one more round, one more growth milestone, one more partnership. By the time they're ready to exit, the buyer isn't there anymore.
That's the crash. It's not a demand problem. It's a timing problem.
The Playbook From Past Bubbles
Gil pointed to three names: Lotus, AOL, and Mark Cuban's Broadcast.com. All three sold at or near peak value. All three knew the wave was about to break. And all three made their founders billionaires because they didn't wait.
Think of an AI startup like a surfer. The ride is great. But the surfer who stays on the wave one second too long gets crushed under it. The exit isn't when the wave is biggest. It's right before it closes out.
What This Means For Public Markets
Gil is talking about private companies, but the same logic shows up in public AI stocks. The ones with real defensibility - chip makers, infrastructure plays, companies with proprietary data - are a different trade than the application-layer stocks that live or die on whether OpenAI adds a feature.
Separating the two is the whole game right now. It's also why some AI stocks are up 10x in two years and others are flat or down.
Worth Noting
The next six months are packed with big AI product launches and IPOs. Cerebras filed last week. Factory just hit a $1.5 billion valuation. The money is still flowing into the category. But a veteran is on record saying the easy part is about to end.
One more ride, or one more sale. That's the question founders are quietly asking each other.
Source: TechCrunch