Meta plans AI chip production from September as it expands computing capacity: Report

The chip, codenamed “Iris,” is part of Meta’s multi-generation Meta Training and Inference Accelerator (MTIA) programme.Meta Platforms plans to begin manufacturing a new in-house artificial intelligence (AI) chip from September as part of its strategy to significantly expand computing capacity, according to an internal memo reviewed by Reuters.The chip, codenamed “Iris,” is part of Meta’s multi-generation Meta Training and Inference Accelerator (MTIA) programme, which aims to develop custom AI silicon for powering AI features across Facebook, Instagram and other products.Meta is designing the chip in collaboration with Broadcom, while Taiwan Semiconductor Manufacturing Co. (TSMC) will manufacture it.

The company aims to reduce its dependence on third-party suppliers such as Nvidia and Advanced Micro Devices (AMD) while lowering the costs of expanding its AI infrastructure.According to the memo, testing of the Iris chip was completed in just six weeks without any major issues, marking a notable milestone for Meta’s in-house chip development efforts, which have faced challenges over the past several years.The chip is intended to complement, rather than replace, the large volumes of graphics processing units (GPUs) that Meta continues to purchase from Nvidia and AMD for AI workloads.The internal memo noted that deploying the latest GPUs across Meta’s infrastructure has been “a heavy lift” and has “cost us time.”Following the report, Meta shares initially declined before recovering after the company announced developer access to a new AI coding model that competes with offerings from OpenAI and Anthropic.

The stock was trading about 4.6 per cent higher in late afternoon trading.“You can’t become an AI titan if you are dependent on another company for chips,” Mike Gualtieri, vice president and principal analyst at research firm Forrester, told Reuters. “The hyperscalers and even SpaceX all plan chips because it will be the only way to compete on price for model usage.”Meta had unveiled Iris under its technical MTIA branding in March alongside three other AI processors.

It plans to introduce a new AI chip roughly every six months through 2027, a significantly faster cadence than the industry norm of annual releases.AI infrastructure spending to reach $145 billionThe social media giant expects to spend as much as $145 billion on AI infrastructure this year, accounting for a substantial share of Big Tech’s projected investment in AI.To support the expansion of its AI infrastructure, Meta has also secured long-term supply agreements with several component manufacturers, including Samsung Electronics for memory chips, SanDisk for flash storage and Sumitomo Electric for fibre-optic equipment.