Google is in talks with Marvell to build new AI chips for inference, according to a report published Tuesday, April 20, 2026. The effort is aimed at speeding up AI results and, by extension, taking on Nvidia in one of the most important parts of the semiconductor market.
The push matters because inference is where AI systems do the work of answering questions, generating text and serving results to users, the part of the market that determines how fast and efficiently those tools can operate at scale. Google is said to be exploring the partnership with Marvell as it looks for a way to sharpen that edge, and the move puts Mrvl stock in view as investors weigh whether the chipmaker could land a larger role in a fast-growing corner of AI hardware.
Marvell has spent years building its business around custom silicon and data-center networking, but a possible arrangement with Google would give it a more visible place in the race to power AI services. Google’s interest also shows how far the company is willing to go to control more of the technology stack behind its AI products rather than rely entirely on outside suppliers.
What is not yet clear is how far the talks have advanced or whether they will lead to a deal. For now, the key fact is that Google is looking beyond its current setup, and that alone is a reminder that the fight over AI chips is moving deeper into the infrastructure that determines who can answer faster and at lower cost.
