Facebook’s parent company Meta Platforms plans to deploy its own custom-designed artificial intelligence chips, codenamed Artemis, in its data centers this year, according to an internal document seen by Reuters. The move could reduce Meta’s dependence on Nvidia’s market-dominating H100 chips and control the rising costs of running AI workloads.
Meta has invested billions of dollars to boost its computing capacity for the power-hungry generative AI products it integrates into services like Facebook, Instagram and WhatsApp. This includes purchasing specialized chips and reconfiguring data centers to house them.
According to Dylan Patel, founder of silicon research group SemiAnalysis, the successful deployment of Meta’s own chip could potentially save the company hundreds of millions of dollars in annual energy costs and billions in chip purchasing costs.
Still dependent on Nvidia for the time being
Despite this move towards self-reliance, Meta will continue to use Nvidia’s H100 GPUs in its data centers for the foreseeable future. CEO Mark Zuckerberg has stated that the company plans to have approximately 350,000 H100 processors in use by the end of this year.
The deployment of the proprietary chip marks a positive turn for Meta’s internal AI silicon project, following a decision in 2022 to abandon the first iteration of the chip in favor of Nvidia’s GPUs.
The new chip, Artemis, like its predecessor, is designed for AI inference, which uses algorithms to make ranking judgments and generate responses to user queries.
“We see our internally developed accelerators as highly complementary to commercially available GPUs in delivering the optimal mix of performance and efficiency on meta-specific workloads,” said a Meta spokesperson.
While Meta’s move to reduce its reliance on Nvidia’s processors may mark the first crack in Nvidia’s AI wall, it’s clear that Nvidia’s GPUs will continue to play an important role in Meta’s AI infrastructure for the time being.