ARTICLE AD BOX
![]()
Cerebras Systems CEO Andrew Feldman said investors misunderstood the company's profit outlook. This remark came after the artificial intelligence (AI) chipmaker's shares fell nearly 20% following its first earnings report as a publicly listed company.
The decline came after Cerebras forecast a lower gross margin for its core business this year. Speaking to CNBC, Feldman said the company's guidance had been misinterpreted and maintained that Cerebras remains ahead of the financial plan it shared when it went public. "It is misunderstood. You know, we laid out a plan at the start of ’26. We shared that plan as we went public a few months ago, and we’re beating that plan,” he said on CNBC's Squawk on the Street.
Why did Cerebras investors react to the company’s stock
For the first quarter, Cerebras reported a gross margin of 47% in its core business, excluding the impact of customer warrants and data centre pass-through revenue. The company expects that figure to range between 38% and 41% for the full year. According to Feldman, part of the outlook reflects the company's decision to rent back some equipment from one of its largest customers, which affects margins over time."I think it’s not going to be a straight line," he said.
Despite the market reaction, analysts at Mizuho and Wedbush raised their estimates for the company following its earnings call, CNBC reported.Investors are also keeping an eye on the company’s staggered lock-up schedule following its IPO. Cerebras’ prospectus said about 28 million Class A shares held by directors, officers and some non-employee shareholders were eligible for trading on the second trading day after the earnings announcement.Feldman said the staggered approach was intended to spread the release of shares rather than have them become available all at once after a fixed period following the IPO."Whether that’s a success or not, we’ll have to see," he told CNBC's Carl Quintanilla and Leslie Picker.Feldman also said Cerebras is not facing some of the supply constraints affecting competitors such as Nvidia, including shortages of high-bandwidth memory and advanced manufacturing processes from Taiwan Semiconductor Manufacturing Co.However, he noted that the company faces a different challenge as it works to expand its data centre footprint."We’re trying to move at the speed of AI, and data centres move with the speed of real estate," Feldman added.The comments come as AI infrastructure providers continue expanding capacity to meet growing demand while navigating permitting processes and site development timelines.

English (US) ·