Cerebras shares drop sharply after-hours despite raised full-year outlook, indicating weaker demand for wafer-scale hardware

Cerebras Systems’ stock fell roughly 15% after hours following an update that mixed confidence with caution. The company was widely framed by investors as a potential alternative to Nvidia in the race to build faster, more efficient AI computing systems, but the market reacted negatively to signs of cooling near-term demand for its hardware business.

According to the report circulating on market platforms, Cerebras said it had raised its full-year outlook. That kind of guidance increase typically indicates management expects revenue, margins, or other performance metrics to come in ahead of prior expectations. But the immediate market reaction suggested investors were still looking through the company’s longer-range tone.

The post tied the decline to quarterly operating results, pointing specifically to a decline in hardware revenue in the second quarter. The reported figure showed hardware revenue down about 23% year over year, a drop large enough to overwhelm the positive message from the raised full-year outlook.

The same report also described Cerebras as a competitor with “wafer tech” characteristics, a reference to its approach of building large-scale AI processing capabilities on silicon rather than relying only on traditional multi-chip systems. In these systems, the hardware roadmap and customer uptake are often the primary variables that determine short-term revenue trajectories, even when the company expects improvement later in the year.

Investors’ focus on the hardware line suggests the market is still judging Cerebras on how quickly orders and deployments convert into revenue. Even when a company takes steps to lift full-year guidance, a pronounced contraction in a quarter can raise questions about customer procurement timing, product mix, or the pace at which new systems move from evaluation to paid deployments.

The report also framed the stock move as part of a broader “fewer buyers” narrative, implying that demand indicates were less favorable than traders had expected. Without additional figures in the circulated post, it is unclear whether the concern centered on the number of customers, the average deal size, inventory digestion at partners, or a more general pullback in enterprise and cloud spending on AI infrastructure.

For context, the market for AI accelerators and training infrastructure remains highly competitive. Nvidia’s ecosystem is tightly connected to hardware, software, and developer tools, which can make switching costs and procurement cycles complex for customers. Cerebras’ pitch, as reflected in the “wafer tech” framing, is aimed at boosting performance and efficiency by rethinking how compute is packaged, but that strategy still depends on sustained buy-in that shows up in revenue each quarter.

Cerebras did not provide details in the circulated report about the specific drivers behind the second-quarter hardware revenue decline, such as customer-by-customer timing, backlog changes, or whether the quarter reflected product transitions. That means it is difficult to determine whether the weakness is cyclical, tied to a temporary step in deployments, or indicative of a more structural slowdown.

The next set of disclosures to watch would likely include more detailed commentary on hardware demand, customer rollout schedules, and how the company plans to convert its wafer-scale approach into consistently growing revenue. Traders will also look for whether the raised full-year outlook is supported by improved quarterly trends in subsequent results, rather than primarily by assumptions for later periods.