Cerebras & Marvell: Reshaping AI's Future

Cerebras Systems is challenging the AI chip market with its unconventional strategy of creating enormous processors designed for accelerated AI computation.

This approach targets applications demanding rapid responses, such as autonomous agents and coding assistants. In its second quarter of fiscal year 2026, Cerebras experienced substantial revenue growth, with core revenue doubling and cloud revenue nearly quadrupling year-over-year. The company also reported $25.4 billion in remaining performance obligations, ensuring future revenue visibility.

Management projects core revenue to more than triple in 2027, following an estimated $880 million to $890 million this year. If Cerebras sustains its historical growth rate of doubling revenue annually, it could reach approximately $28 billion by 2031. Strategic partnerships with entities like OpenAI, AMD, and Amazon's AWS further bolster its potential.

Marvell Technology is taking a different approach, positioning itself as a critical "pick-and-shovel" supplier for AI infrastructure by providing custom chips and high-speed connectivity solutions for data centers.

This strategy aligns with cloud providers' increasing demand for specialized hardware to enhance efficiency and reduce reliance on single suppliers. Marvell's involvement with major hyperscalers like Alphabet, Microsoft, and Meta Platforms, alongside its expertise in optical networking, addresses the growing challenge of data movement within expanding AI clusters.

Marvell's ambitious goal is to increase its revenue tenfold by fiscal year 2031, aiming for $70 billion to $90 billion from $8.2 billion in fiscal year 2026. This growth is contingent on scaling its operations and maintaining relationships with key customers, despite risks of program delays and competition.