Meta Platforms is moving to manufacture its custom artificial intelligence chip, code-named "Iris," starting in September. This internal project aims to double the company's total computing power to 14 gigawatts by next year.
The chip's testing phase was completed in just six weeks with no major issues, marking a significant step for Meta's in-house silicon effort, which has been in development for over five years. Meta designed Iris for its specific needs and is collaborating with Broadcom for design and Taiwan Semiconductor Manufacturing Co for manufacturing. This strategy is intended to lower computing costs and decrease dependence on major chip suppliers like Nvidia and AMD.
"You can't become an AI titan if you are dependent on another company for chips," stated Mike Gualtieri, a principal analyst at Forrester. "The hyperscalers... all plan chips because it will be the only way to compete on price for model usage."
Meta currently plans to deploy seven gigawatts of computing infrastructure this year. The company has secured long-term supply agreements with firms like Samsung Electronics and Sandisk to support this expansion amid a global memory chip shortage, which has driven prices up significantly.
The tech giant expects to spend as much as $145 billion on AI infrastructure this year, contributing to a massive industry-wide investment projected to exceed $700 billion.