Cerebras is an AI chipmaker with processors as big as dinner plates.
It just went public in one of the hottest IPOs of the year. Shares priced at $185, opened at $350, and hit $385 intraday. Wall Street cheered as the AI hype machine spun at full throttle.
On the surface, Cerebras (Nasdaq: CBRS) looks like the next big thing.
But behind the scenes, the picture doesn’t look so rosy.
Today, I’ll explain why you should avoid the stock — and what to buy instead.
The Sizzle
Cerebras isn’t just another GPU wannabe.
Its Wafer-Scale Engine chips are massive — the size of dinner plates — and pack the power of dozens of Nvidia H100s onto a single piece of silicon. That means no shuffling data between separate memory and processing chips.
The result? Reports of up to 15x faster performance.
That edge helped fuel a blockbuster debut. In fact, it was the biggest IPO of the year, with shares popping nearly 100% on opening day.
Furthermore, the semiconductor sector is on fire. It now makes up over 15% of the S&P 500.
Nvidia still dominates with ~85% market share. But Cerebras is carving out a differentiated niche. In the AI arms race of today, this is exciting stuff.
But here’s where the picture gets cloudy…
The Risky Reality
Last year, a whopping 85% of Cerebras’ revenues came from a single customer: G42, a UAE state-backed AI firm.
That’s why investors (as well as CFIUS, the US agency that ensures foreign investments don’t pose a risk to our national security) flagged the geopolitical risk.
Cerebras said it would fix the issue, and soon dropped G42 exposure to 24%. But if you dig in, you’ll see that 62% of its revenue now comes from Mohamed bin Zayed University of AI — another UAE-linked entity.
That means 86% of its revenues are still tied to a single sovereign wealth-fund customer!
Furthermore, the company’s finances aren’t screaming “inevitable winner.” Its revenue doubled to ~$510 million in 2025, but in today’s AI world, that’s table stakes to earn a sky-high valuation. At its current level, it’s trading at a massive price-to-sales multiple — well above its peers on modest (and concentrated) revenue.
Then there’s the company’s OpenAI deal — the big hope. Up to 750 megawatts of compute, which could potentially add up to $7 billion to $10+ billion in annual revenue at full capacity.
That could be transformative… if it fully materializes. The thing is, this deal has an exclusivity clause that limits sales to competitors. And meanwhile, it all hinges on trusting OpenAI (and Sam Altman’s track record) to pay up.
That’s not a bet I’d make with public shares at these levels.
$10k into $4.5 Million
But now let’s look at the investors who got in early — when the company was still a private startup.
For example, venture-capital firms Benchmark and Foundation Capital got in at around 85 cents per share. At the IPO price of $185, that’s already an enormous profit. And if they sold anywhere near the $385 peak, they could be pocketing 450x their money.
At that level, a $10k investment turns into $4.5 million.
Here’s a chart, courtesy of The Information and PitchBook, that show the investors who got in early, and the share price they paid:
These investors earned life-changing wealth by getting in before the stock ever traded publicly.
Meanwhile, by the time the rest of us could buy shares on the Nasdaq, they were trading at $385. Now they’re trading for closer to $240.
So if you’d bought at the open, you’d now have lost around 38% of your money.
The Smarter Path Forward
Cerebras highlights a timeless truth in tech investing:
The biggest wins come from backing innovative companies before Wall Street discovers them. IPOs are often the exit party for early believers, not the entry point for new ones.
We’re not saying avoid AI chips or high-growth tech. Just be strategic about when and how you get in. Public markets right now are pricing in perfection for these names — despite the fact that the sector appears to be in a bubble.
At Crowdability, we’re currently digging into several private companies that could be the “next Cerebras” — AI-related startups with strong tech differentiation, a wide range of customers, and ample room to deliver big returns. We’ll share more as our research progresses. Stay tuned.
In the meantime, if you’re excited by the AI chip story but wary of chasing CBRS at today’s levels, consider the private markets.
That’s where the real asymmetry lives — before the hype sets the price.
Happy investing.
Founder
Crowdability.com
