Cerebras' core revenue more than doubled year over year in the second quarter, and management raised its full-year outlook.
Even after a 52% decline, the stock trades at about 145 times next year's expected earnings.
Cerebras expects about $5.6 billion of its $25.4 billion backlog to convert to revenue by the end of June 2028.
Cerebras Systems (NASDAQ:CBRS) has been cut in half. Shares of the artificial intelligence (AI) computing specialist peaked at $386.34 on their first day of trading in May, a day after the company priced its initial public offering (IPO) at $185 and raised $6.4 billion. As of this writing, the stock trades around $184 -- about 52% below the high, and just under that IPO price.
A decline like that usually follows a stumble: a guidance cut, a lost customer, a bad quarter. Cerebras hasn't had one. Shares dropped double digits anyway the day after its mid-August update, arguably the company's strongest yet and one in which management raised its outlook for the year.
Missed AI’s "Act 1"? Act 2 Could Be 15x Bigger. Most investors think they missed the AI boat because they didn't buy Nvidia in 2005. But according to our analysts, we’re only at the end of "Act 1"—the R&D phase. "Act 2" is the global rollout. Continue »
Yet even at half price, the stock costs about 145 times the earnings analysts expect the company to generate next year.
What did the fall reprice -- the business, or just the price tag?
Image source: Getty Images.
It wasn't the business. In its mid-August update, Cerebras reported core revenue (a measure that strips out pass-through data center revenue and adds back a noncash deduction tied to customer warrants) of $209.9 million, up 103% year over year, with core gross margin improving to 41% and core operating margin to negative 16%. In other words, the company still loses money, but far less of it on every dollar of sales. And management raised its full-year outlook to a range of $880 million to $890 million in core revenue, up from a prior range of $855 million to $865 million.
The cloud business is driving the growth, and the growth is accelerating. Cloud and other services revenue, as reported under generally accepted accounting principles (GAAP), was $33 million in the year-ago quarter, about $83 million in this year's first quarter, and $126 million in the second, up 281% year over year. Renting out computing is increasingly the story -- the cloud side supplied 70% of the company's reported revenue in the latest quarter, up from about a third a year ago.
Whatever cut the stock in half, it wasn't the fundamentals.
What changed is what buyers will pay for that growth. Analysts expect Cerebras to earn about $1.25 per share next year.
At around $184, the stock costs about 145 times those expected earnings. And measured against the same estimate, May's $386.34 peak worked out to about 310 times next year's expected earnings.
That, I'd argue, is the honest reading of the sell-off. Investors didn't mark down Cerebras' business. They marked down what they were willing to pay for it. Even now, buyers are paying for profits that are still mostly forecast.
For today's price to make sense, earnings have to grow into it. Suppose the stock eventually traded at 30 times earnings -- still a premium price for most businesses. At about $184 per share, Cerebras would need to earn about $6 per share annually, nearly five times what analysts expect for next year.
The bull case is that much of the growth needed to get there is already under contract. Cerebras ended June with $25.4 billion in remaining performance obligations (work customers have committed to pay for but the company hasn't yet delivered), much of it tied to a single agreement under which OpenAI committed to purchase 750 megawatts of AI computing capacity. Cerebras expects about $5.6 billion of that balance, or 22% of it, to become revenue over the 24 months ending June 30, 2028.
"We have made rapid progress in key areas required to deliver exceptional growth ... and plan to more than triple revenue in 2027," chief financial officer Bob Komin said in the company's second-quarter earnings release.
Sure, revenue could triple next year. The contracts support it. But revenue isn't earnings.
After all, Cerebras ran a core operating loss last quarter even with sales doubling, and most of the backlog converts after mid-2028. The profits the price depends on are still years of execution away, and the timing can shift.
So, what did the halving reprice? Mostly the price tag. The business is in better shape than it was in the spring, and the backlog gives Cerebras' growth plans unusual visibility.
But at about 145 times next year's expected earnings, the stock still assumes years of rapid growth and a smooth swing to profitability, and it leaves little room for the timing to slip.
Ultimately, cut in half isn't the same as cheap. I'd stay on the sidelines at this price. If Cerebras starts turning that backlog into profits, and not just revenue, I'd take another look.
Before you buy stock in Cerebras Systems, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Cerebras Systems wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $420,109!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,303,689!*
Now, it’s worth noting Stock Advisor’s total average return is 938% — a market-crushing outperformance compared to 211% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
See the 10 stocks »
*Stock Advisor returns as of September 16, 2026.
Daniel Sparks and his clients do not have positions in any of the stocks mentioned. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.