Cerebras (CBRS) Q2 2026 Earnings Release

Cerebras Systems has revised its full-year revenue guidance upward following its second quarterly earnings report since its May initial public offering. Despite this positive adjustment, the company’s stock experienced a notable decline of approximately 14% in extended trading.

In the second quarter, Cerebras reported revenue of $210 million and a loss per share of $2.89.

The company announced on Wednesday that it anticipates core revenue for the current quarter to be in the range of $214 million to $216 million. This projection narrowly surpasses analyst expectations of $212.6 million in revenue, although it remains unclear if these analyst estimates specifically referred to core revenue. The reported $210 million in second-quarter sales represents its core revenue figure, distinct from the GAAP revenue of $180.1 million, which excludes “pass-through revenue.”

Cerebras recorded a net loss of $450.5 million for the quarter. This stands in contrast to a profit of $309.5 million, or $1.91 per share, in the same period a year ago. A significant portion of this loss, $386.6 million, is attributed to stock-compensation expenses.

The updated full-year outlook now projects core revenue between $880 million and $890 million, an increase from the previous guidance of $855 million to $865 million.

Cerebras CEO Andrew Feldman expressed strong optimism regarding the artificial intelligence market, describing demand as “through the roof” and highlighting that clients are willing to pay a premium for its specialized inference chips. Cerebras is strategically positioning itself to compete with industry leader Nvidia in specific AI workloads, particularly those requiring ultra-low latency for interactive applications, a capability the company terms “fast inference.” To address investor concerns regarding profitability, Cerebras anticipates its core gross margin to expand to between 38% and 40% in the current quarter.

“Gross margins are in a good spot, and growing, because fast inference is priced at a premium,” Feldman stated in an interview, underscoring the company’s ability to enhance the AI processing power of its systems.

Cerebras made its public debut on the Nasdaq in May, leveraging the surge in investor interest in semiconductor companies focused on AI model acceleration. The company priced its IPO at $185 per share, raising $6.4 billion. While the stock reached its peak shortly after going public, it has since seen a correction. Nevertheless, on Wednesday, it closed at $262.06, representing a 42% increase from its IPO price.

The chip manufacturer boasts $25.4 billion in remaining performance obligations, a figure it interprets as a strong indicator of “extraordinary future demand.” Feldman is confident that as Cerebras scales, it will benefit from economies of scale, projecting revenue to triple in the upcoming fiscal year. “We will manufacture more efficiently. We’ll get better pricing on componentry. We’ll amortize our manufacturing organization over more units,” Feldman explained, projecting a positive growth trajectory.

In recent weeks, Cerebras has solidified strategic partnerships, including one with Advanced Micro Devices, a key competitor to Nvidia, with products slated for production later this year. Furthermore, the company announced that OpenAI will utilize its chips to power its latest model, GPT 5.6-Sol. Cerebras also offers cloud-based access to its chips, a segment that generated $126 million in revenue during the June quarter. The company’s advanced wafer-scale engine architecture, designed for massive parallel processing, positions it to handle the escalating computational demands of cutting-edge AI models, particularly in areas like natural language processing and complex simulation, where low-latency inference is paramount for real-time user experiences and rapid scientific discovery. The unique design of Cerebras’s Wafer-Scale Engine, which integrates millions of AI-optimized cores on a single silicon wafer, aims to overcome the limitations of traditional multi-chip architectures, promising significant improvements in performance, power efficiency, and reduced communication overhead for large-scale AI deployments. This focus on specialized hardware for inference, a critical phase in the AI lifecycle where trained models are used to generate predictions or classifications, is becoming increasingly vital as the adoption of AI solutions across various industries accelerates. Cerebras’s strategy to address both on-premise deployments and cloud-based services allows for flexibility and caters to a broader market spectrum. The company’s ability to secure such high-profile partnerships and its strong backlog of future business underscore its potential to capture a significant share of the rapidly evolving AI hardware market, especially as companies seek to optimize inference costs and performance for their AI-driven applications.