In the race to power AI, Google and Marvell potentially explore new chip designs, eyeing to reshape competition with NVIDIA and redefine the future of cloud computing.
Alphabet’s Google is reportedly in discussions with Marvell Technology to co-develop two new semiconductor chips designed to improve the efficiency of artificial intelligence (AI) workloads, according to The Information.
The talks, cited by two people familiar with the matter, suggest Google is seeking to expand its hardware portfolio beyond existing processors.
As per Reuters, one of the proposed chips is a memory processing unit intended to operate alongside Google’s tensor processing unit (TPU). The other is a next-generation TPU, purpose-built for running AI models. The collaboration would mark a significant step in Google’s efforts to strengthen its position in the AI hardware market, where NVIDIA’s graphics processing units (GPUs) currently dominate.
Google has been working to position TPUs as a credible alternative to GPUs, with sales of the processors contributing to growth in its cloud division. Demonstrating returns from AI investments has become a priority for the company as it seeks to reassure investors of the commercial viability of its technology.
Neither Google nor Marvell responded immediately to requests for comment. Reuters noted it could not independently verify the report. According to The Information, the companies aim to finalise the design of the memory processing unit by next year, before moving into test production.
If confirmed, the partnership would highlight the increasing competition among chipmakers to supply the infrastructure underpinning AI development, a sector that has rapidly become central to the strategies of major technology firms.
