Cerebras Systems (NASDAQ: CBRS) saw its stock fall as much as 17% in after-hours trading on Wednesday, August 13, even after the AI chipmaker surpassed Wall Street expectations across most financial metrics and raised its full-year revenue outlook.

The market reaction underscores a broader question for AI infrastructure investors: whether leading names in the sector can continue to sustain premium valuations. Cerebras, which went public in May at $185 per share, had seen its stock run up sharply since its IPO, and the post-earnings pullback reflects the heightened scrutiny applied to newly public AI companies whose market capitalizations have scaled rapidly on forward expectations rather than proven profitability.

Q2 Financial Results

Cerebras reported core revenue of $210 million for the second quarter, more than double the same period a year ago and comfortably ahead of the $191 million analysts had projected, according to CNBC figures cited by Quartz. On an adjusted basis, the company lost 5 cents per share, significantly narrower than the anticipated loss of 17 cents per share. The adjusted operating loss came in at $34 million, well below the $63 million estimate reported by Barron's.

GAAP total revenue reached $180.1 million, representing a 74% year-over-year increase.

Wafer-Scale Engine Revenue Declines

Cerebras is best known for its Wafer-Scale Engine, a design that preserves an entire 300-millimeter silicon wafer as a single massive processor tailored for AI and scientific computing workloads. That technology earned the company its reputation as an "Nvidia rival," but the hardware segment faced headwinds in the second quarter.

Physical hardware revenue dropped 23% year-over-year to $54.1 million, as reported by Yahoo Finance and Binance News.

CEO Andrew Feldman attributed the decline to delivery timing rather than weakening demand. "Hardware is going to be lumpy," he said, noting that the company occasionally cannot book sales because customers lack the data-center space required to house Cerebras' large-scale systems. That constraint reflects a well-documented bottleneck across the AI infrastructure industry, where power availability and facility build-outs have struggled to keep pace with accelerator deployments.

Cloud Revenue Surges

GAAP cloud and other services revenue climbed to $126 million, up 281% from a year earlier. Core cloud revenue rose 287% to $127.7 million, according to Quartz. The figures indicate that Cerebras, which positions itself as a chip challenger to Nvidia, now generates its largest revenue stream from renting compute capacity rather than selling hardware outright. The shift mirrors a broader industry trend in which AI chip designers are building managed-service businesses around their silicon, offering customers access to compute without the capital expenditure and deployment lead times associated with purchasing systems outright.

Cerebras also swung to a GAAP net loss of $450.5 million, or approximately $2.89 per share, compared with a $309.5 million profit in the same quarter of the prior year. The company attributed the reversal primarily to $386.6 million in stock-based compensation linked to its May initial public offering.

Core gross margin improved by nearly 940 basis points from the second quarter of 2025, reaching 41%, which the company credited to premium pricing on fast inference services.

Raised Guidance

Management raised its full-year 2026 core revenue guidance to a range of $880 million to $890 million, up from the prior range of $855 million to $865 million. The revised outlook also exceeded the $867.6 million consensus estimate.

For the third quarter, Cerebras guided to core revenue of $214 million to $216 million, ahead of analyst consensus near $212 million per Koyfin, with gross margin projected between 38% and 40%.

CFO Bob Komin stated that the company aims to more than triple revenue in 2027 as production scales and component costs decline.

CEO Feldman said in a statement that "speed changes what AI can do" and "opens entirely new markets." He characterized gross margins as "in a good spot, and growing, because fast inference is priced at a premium."

Post-IPO Balance Sheet

The May IPO left Cerebras with substantial financial resources. The company raised $6.4 billion in gross proceeds and ended June with $8.6 billion in cash, cash equivalents, restricted cash, and short-term investments. It also reported $25.4 billion in remaining performance obligations, a figure that provides multi-year revenue visibility as those contracted commitments convert to recognized revenue over time.

Shares closed Wednesday's regular session at $262.06, approximately 42% above the $185 IPO price, before the after-hours decline trimmed a portion of those gains.

Cerebras is also expanding its industry partnerships, including a collaboration with AMD and an arrangement enabling OpenAI to run models such as GPT 5.6-Sol on its systems. The OpenAI arrangement in particular positions Cerebras as a compute provider for large-scale inference workloads, placing it more directly in competition with cloud platforms that currently dominate AI model serving.