Six weeks ago, Cerebras was Silicon Valley's golden child. The AI chip maker had just debuted on the Nasdaq with a $5 billion valuation, its 'wafer-scale' chips touted as the next big thing in artificial intelligence. Today, the stock sits below its IPO price, down more than 50% from its peak. Ouch.
Cake or Wreckage?
This isn't your typical post-IPO stumble. The slide has been brutal. Fast, too. After opening at $35, shares surged to $48 on sheer hype. Now they're scraping $23. Early investors who bought in at the IPO are underwater. The ones who bought after the pop? Drowning.
The warning signs were there. Cerebras was burning cash like a teenager with a credit card. Last quarter's earnings showed revenue growth, sure, but also a net loss of $200 million. That's not sustainable. Unless your name is Amazon, investors eventually stop buying the story.
The AI Hype Machine Idles
Let's not pretend this is just a Cerebras problem. The entire AI chip sector is cooling off. Nvidia's stock is off 10% this month despite crushing earnings. AMD is flat. The market is realizing that building data centers costs billions and the payoff isn't immediate. Cerebras just happens to be the smallest ship in the storm.
But here's the thing: Cerebras's technology is genuinely impressive. Their chips are the size of a dinner plate—no, really—and they pack more processing power than anything Nvidia sells. The problem is customers. Big tech companies like Google and Amazon are building their own chips. Startups can't afford Cerebras's $2 million price tag. Who's left? National labs and a handful of rich universities. That's not a mass market.
“The AI chip space is like the Wild West. Everyone has a six-shooter but nobody has a saloon to drink in.” — anonymous semiconductor analyst
Translation: Cool tech. Tough business.
Early Investors Take the Hit
The biggest losers in this bloodbath are the venture capitalists who poured money into Cerebras before the IPO. Firms like Benchmark and Sequoia loaded up at valuations north of $4 billion. They needed the public market to cheer. Instead, they're watching their paper gains evaporate. Some are reportedly looking for exit strategies—selling stakes to secondary buyers at a discount.
There's an irony here. Cerebras spent years marketing itself as an antidote to Nvidia's dominance. 'Our chips are faster, more efficient, better for AI training,' they said. Turns out, in the stock market, being better isn't the same as being profitable.
What's Next?
CEO Andrew Feldman is trying to keep a brave face. In a memo to employees this week, he wrote, 'We're building a long-term company. Short-term market noise doesn't change our strategy.' Noble, but investors don't care about strategy. They care about share prices. And the price is screaming 'sell.'
Some analysts think Cerebras could be a takeover target. A bigger player—maybe Intel or AMD—might scoop up the technology at a discount. Others predict a continued slide. The next earnings report is due in August. If Cerebras misses expectations, the stock could drop another 20%. If it beats, maybe a dead cat bounce.
Either way, the champagne has gone flat. Cerebras's party started with a bang and ended with a whimper. For early investors, the hangover is real.
The question now: Is this a buying opportunity or a falling knife? If I had to guess, I'd say the blade is still dropping. But I've been wrong before. Just ask the people who bought Cerebras at $48.
