Cerebras posted its first quarterly results as a public company with revenue above Wall Street estimates and a smaller-than-expected adjusted loss, but investors focused on sharply lower margin guidance and heavy spending to expand AI infrastructure, sending shares lower.

Cerebras reported its first quarterly results as a public company on Tuesday, and the numbers showed both the strength of demand for its AI systems and the cost of scaling them.

The wafer-scale chip maker said first-quarter revenue rose to $193.4 million, above the $181.2 million FactSet consensus cited by multiple outlets. It also reported an adjusted loss of $2.48 million, narrower than analysts had expected.

Investors did not reward the beat. Shares fell after the report, with coverage putting the decline at roughly 8% to 15% depending on whether the snapshot came from after-hours, premarket or early trading.

First post-IPO quarter

The report was Cerebras' first since its May IPO, making it a key test of whether the company can turn strong AI demand into a durable public-market story. The company sells wafer-scale AI processors and related cloud services, and investors have been watching whether it can scale quickly enough to meet demand without crushing profitability.

Cerebras also guided second-quarter revenue to about $194 million, above the $177.7 million consensus cited by analysts. That outlook suggested the top line could keep growing even after a strong first quarter.

But the market focused on the parts of the forecast that pointed the other way. Cerebras said first-quarter core gross margin was 47%, then guided second-quarter core gross margin to 36% to 38%. It also projected full-year core operating margins of negative 28% to negative 32%.

Margin pressure and expansion costs

That guidance underscored how much the company is spending to build out AI infrastructure. The margin outlook suggested that revenue growth is arriving alongside heavy capacity investment, a tradeoff that worried investors who were already cautious about a newly public AI hardware company.

Wall Street coverage also pointed to concern about customer concentration and the economics of Cerebras' buildout. Barron's reported that the company said backlog reached $25 billion, with $4 billion expected to become revenue over the next two years, while other coverage highlighted the pressure from the company's infrastructure commitments.

Chief executive Andrew Feldman tried to steer attention back to demand. He said Cerebras was seeing significant momentum with customers including OpenAI and Amazon Web Services, and he said the company's wafer-scale technology delivers the fastest AI in the world.

MarketWatch reported that Cerebras is now serving OpenAI's GPT-5.4. The same broad theme ran through the post-earnings coverage: demand appears strong, but investors want proof that the company can convert that demand into profitable growth.

What investors are watching next

Cerebras' next test is execution. Investors will be watching how quickly the company adds data-center capacity, whether the mix of hardware versus cloud and services revenue changes, and whether margin pressure eases as more of the backlog turns into revenue.

The company went public in May, so this first earnings release sets an early baseline for how the market will judge it: as a fast-growing AI infrastructure company with major customer momentum, or as a business whose expansion requirements keep profits under pressure.

Revision note

Initial automated publication.