Cerebras Systems saw its shares fall about 14% before the opening bell on Wednesday after management signaled that annual profitability metrics would not match the company’s first-quarter performance in its first set of public financial results following a high-profile initial public offering.
The chip maker projected adjusted gross margins of 38% to 41% for 2026, compared with the 47% it recorded in the first quarter. That outlook is substantially below the mid-70% range associated with some competitors and the mid-50% range reported by other rivals, though it did exceed analysts’ consensus estimate of 29.58%.
If the stock’s decline held into the market open, Cerebras was poised to trade at its lowest level since listing more than a month earlier and to erase in excess of $6 billion of market value.
Analysts have pointed to specific operational factors that could weigh on gross margins. One concern is the company’s production of relatively larger-sized chips, which can increase manufacturing costs per unit. Another is a short-term arrangement in which Cerebras is renting back its own systems from an existing client to satisfy immediate customer demand while the company expands data center capacity.
The shares are down more than 27% from their price at market debut, reflecting a cooling of investor enthusiasm for artificial intelligence-related equities and growing unease about the substantial capital outlays required to build the physical infrastructure for AI workloads.
Despite the market reaction, some Wall Street analysts maintained positive longer-term views. Morgan Stanley raised its price target on the stock to $273 from $250. TD Cowen emphasized the significance of deals with major cloud and AI customers for the company’s long-term trajectory.
Cerebras has secured a multi-year arrangement with OpenAI valued at $20 billion, and on a post-earnings call Chief Executive Officer Andrew Feldman said that OpenAI’s GPT 5.4 is currently running on Cerebras chips. Under that agreement, the ChatGPT developer is set to deploy 750 megawatts of the company’s semiconductors.
Feldman also stated that Amazon Web Services will begin deploying the company’s chips in its data centers in the near term, with associated revenue flows expected within the next year.
Context and implications - The guidance illustrates the margin challenges that can accompany rapid scaling of specialized hardware and the temporary measures companies sometimes take to bridge capacity gaps. The outlook has immediate bearing on investor sentiment in semiconductor and cloud infrastructure markets, and it highlights the trade-offs between short-term revenue generation and margin preservation as companies expand data center footprint and customer deployments.