Cerebras stock plunges after earnings as CEO says margin outlook was misunderstood

At a glance:

• Cerebras stock dropped nearly 20% despite delivering better-than-expected first-quarter earnings after the AI chipmaker's IPO
• The company forecast full-year gross margins of 38% to 41%, down from 47% in Q1, due to temporary equipment rental strategy
• CEO Andrew Feldman stated investors misunderstood the margin guidance, explaining the move makes capacity available while building data center infrastructure

Market reaction defies earnings beat

Cerebras Systems witnessed a dramatic market response that defied conventional logic on Wednesday, with shares plummeting almost 20% even after reporting better-than-expected first-quarter results. The stock's steep decline pushed it perilously close to the company's IPO price, highlighting investor anxiety over the AI chipmaker's financial trajectory. This volatility emerged shortly after Cerebras delivered its first earnings report as a publicly traded company, revealing robust top-line growth that initially appeared to validate Wall Street's appetite for AI infrastructure plays.

The disconnect between operational performance and market reception centered squarely on the company's forward-looking guidance, particularly its projections for gross margin compression. While Cerebras reported revenue of $193 million for the quarter—a staggering 94% year-over-year increase—and managed to narrow its net loss to $14 million from $23.9 million in the prior-year period, these positive metrics failed to offset concerns about future profitability.

The margin mystery and equipment rental strategy

At the heart of the market's unease lies Cerebras' revised gross margin outlook, which now projects a full-year range of 38% to 41%, significantly below the 47% achieved in the first quarter alone. CEO Andrew Feldman attempted to clarify the situation during his appearance on CNBC, asserting that investors had fundamentally misunderstood the company's financial roadmap. According to Feldman, the margin compression stems from a strategic decision to temporarily rent back computing systems from one of Cerebras' largest existing customers.

This rental arrangement emerged as part of Cerebras' broader infrastructure expansion strategy, designed to accelerate capacity deployment while the company constructs and commissions its own data center facilities. By making more computing power available to customers sooner, Cerebras aims to capitalize on the surging demand for AI training workloads. However, this tactical maneuver comes at a direct cost to near-term profitability, as the company essentially pays to re-rent its own hardware while simultaneously building out permanent infrastructure.

The first quarter's 47% gross margin represented a peak within this transition period, suggesting that the guided 38%-41% range reflects the ongoing tension between rapid scaling and margin optimization. Feldman's intervention indicates that management recognizes the market's confusion and is actively working to realign investor expectations with the company's long-term strategic narrative.

Growth fundamentals remain strong

Despite the margin headwinds, Cerebras' underlying business metrics paint a compelling picture of exponential growth in the AI infrastructure market. The $193 million in quarterly revenue marks a remarkable achievement, representing a 94% surge compared to the same period last year when the company was still private. This growth trajectory positions Cerebras as a significant player in the competitive landscape of AI chip manufacturers, which increasingly includes rivals like NVIDIA, AMD, and emerging startups vying for dominance in the training processor market.

The company's ability to narrow its net loss from $23.9 million to $14 million demonstrates improving operational efficiency even as it scales production and deployment capabilities. For investors evaluating Cerebras beyond the immediate margin concerns, these figures suggest a business model that successfully converts revenue growth into reduced burn rates, a critical metric for hardware companies in capital-intensive industries.

The rental strategy, while temporarily dilutive to margins, represents a calculated bet on market timing. In the rapidly evolving AI ecosystem, being first to market with unprecedented computational capacity can be more valuable than maintaining peak margins during the scaling phase. Cerebras appears positioned to leverage its wafer-scale integration technology to offer customers access to larger AI models than competitors can currently provide.

What this means for AI infrastructure investors

The market's reaction to Cerebras' earnings call reveals deeper concerns about the sustainability of AI infrastructure valuations as the sector matures from speculative growth to profitable operations. Investors who initially embraced AI chipmakers based on revenue multiples are increasingly scrutinizing unit economics, gross margins, and path-to-profitability metrics that traditionally matter less for software companies but are fundamental for hardware manufacturers.

For Cerebras specifically, the coming quarters will be crucial in demonstrating whether the temporary margin compression is indeed an isolated consequence of infrastructure transition or a harbinger of structural challenges in the AI chip market. The company's ability to maintain its revenue growth trajectory while gradually improving margins will largely determine whether Wednesday's sell-off represents a buying opportunity or the beginning of a longer-term re-rating of AI infrastructure stocks.

The rental arrangement with an unnamed largest customer also raises questions about customer concentration risk and contractual flexibility in an industry where leading-edge AI workloads demand specialized hardware. As Cerebras continues to build out its data center presence, the company will need to balance short-term revenue optimization with long-term customer relationships and technological differentiation.

Looking ahead: data center buildout and competitive positioning

Cerebras' current quarter performance places it at a pivotal moment in the AI infrastructure race, where the gap between theoretical performance and practical deployment continues to widen. The company's wafer-scale integration approach promises computational capabilities that traditional multi-chip configurations struggle to match, but realizing this potential requires massive capital investment in data center infrastructure and cooling systems.

The temporary rental strategy serves as both a bridge and a burden—accelerating time-to-market for new customers while creating short-term margin pressure that investors are struggling to quantify. As Cerebras completes its data center buildout, the company expects to transition from this rental model back to higher-margin direct sales, though the timeline for this normalization remains unclear.

Market observers will be watching closely to see whether Cerebras can maintain its revenue growth momentum while investors digest the implications of its margin guidance. The AI infrastructure space remains dynamic, with new competitors emerging and established players like NVIDIA continuously iterating on their own offerings. In this environment, Cerebras' ability to convert its technological advantages into sustainable competitive positioning will ultimately determine whether the current market skepticism proves warranted or presents an attractive entry point for long-term investors.