Technology and Science09 July, 20262 min read
Meta Plans To Begin Iris AI Chip Manufacturing In September To Double Computing Capacity
This article is an automated aggregation of coverage from news outlets. Our editorials offer in-depth analysis curated by the NewsCord team.
Meta will start manufacturing its Iris AI chip in September. Iris is an in-house design under MTIA, with Broadcom designing and TSMC fabricating.
Iris Chip Production Starts
Meta plans to begin manufacturing its in-house artificial intelligence chip, code-named Iris, in September, according to an internal memo reviewed by Reuters. The chip is part of Meta’s four-generation Meta Training and Inference Accelerators (MTIA) program, and it is intended to support AI training and inference for systems behind Facebook and Instagram. Reuters reported that testing took just six weeks and faced no major problems, and Meta plans to launch a new chip roughly every six months through 2027.
Meta’s effort is designed to complement the GPUs it buys from Nvidia and AMD, while reducing reliance on third-party silicon vendors. The memo also ties the Iris timeline to a broader infrastructure buildout, with Meta aiming to double its computing capacity to 14 gigawatts by 2027.
Partners and Supply Deals
Meta is working with Broadcom on the Iris chip’s design and with Taiwan Semiconductor Manufacturing Company (TSMC) on production, as Reuters reported. The memo reviewed by Reuters says Meta is aiming to deploy seven gigawatts of computing infrastructure this year and double that to 14 gigawatts by 2027. To support the expansion, Meta has secured long-term, multi-year supply agreements with Samsung Electronics for memory chips, Sandisk for flash storage, and Sumitomo Electric for fiber-optic equipment, according to Reuters.
Meta’s internal memo also says adopting the latest GPUs at Meta’s scale “has been a heavy lift, and it has cost us time,” and the company is using Iris to address that operational friction. Meta declined to comment on the plan, while the Reuters report described the chip as part of Meta’s push toward more in-house silicon for AI workloads.
Cost, Independence, and Scale
Meta’s Iris program is framed as a way to lower massive computing costs and gain more independence from chip suppliers such as Nvidia and Advanced Micro Devices, according to the Reuters-reviewed memo described by CNBC. Meta expects to spend as much as $145 billion on AI infrastructure this year, and Reuters reported that this spending is part of a broader push to expand data-center compute. The Reuters report said Meta plans to introduce a new chip roughly every six months through 2027, compared with a typical annual-or-longer release cycle for AI chips.
Meta’s custom silicon strategy is also presented as a response to demand strain in chip production resources, with the Reuters report describing manufacturing and packaging capacity as outpaced by demand. In the same Reuters coverage, the memo’s production milestone is linked to Meta’s goal of scaling AI infrastructure while continuing to buy GPUs from Nvidia and AMD for AI workloads.