Cerebras CEO says margin forecast was 'misunderstood' as stock plummets after earnings

Andrew Feldman, CEO of Cerebras Systems, expressed on Wednesday that investors may have misinterpreted the company's margin guidance, following a significant 17% drop in shares after the chipmaker's earnings report, its first since going public. Following the earnings call, analysts from Mizuho and Wedbush adjusted their estimates upward, despite the company's forecast of a tighter gross margin in its primary operations. The reported margin for the first quarter stood at 47%, with projections for the full year ranging between 38% and 41%. "It is misunderstood," Feldman stated during an appearance on CNBC's Squawk on the Street. He emphasized that the company had outlined a strategic plan for 2026, which was shared during its public debut a few months prior, and claimed that they are making progress against that plan. Additionally, Feldman noted that Cerebras will need to lease back some equipment from one of its major clients. "It's not going to be a linear path," he remarked, acknowledging the challenges ahead. Investors are also facing the complexities of staggered lock-up restrictions for Cerebras insiders, which involves approximately 28 million Class A shares that can be traded starting the second day after the earnings announcement, as detailed in the company’s prospectus. The staggered schedule aims to ease the trading transition that typically follows an initial public offering, Feldman explained, adding, "Whether that's a success or not, we'll have to see." Meanwhile, competitors like Nvidia are grappling with supply shortages in high-bandwidth memory and advanced processes from Taiwan Semiconductor Manufacturing Company. However, Feldman pointed out that Cerebras is not dependent on these resources. On the other hand, the company is under pressure to expand its data center capabilities amid rising public opposition and lengthy permitting processes, which pose significant hurdles. "We're aiming to operate at the pace of AI, but the development of data centers operates at the pace of real estate," he concluded.