Meta Platforms plans to start manufacturing an artificial intelligence chip from September as part of its plan to boost overall computing power to 14 gigawatts next year, according to an internal memo reviewed by Reuters.
The chip, code-named 'Iris', is part of a four-generation family of Meta Training and Inference Accelerators (MTIA) designed entirely in-house. Testing completed in only six weeks with no major issues found — a sign of positive momentum for an internal silicon effort that had struggled since its launch over half a decade ago.
'You can't become an AI titan if you are dependent on another company for chips,' said Mike Gualtieri, a vice president and principal analyst at Forrester. 'The hyperscalers and even SpaceX all plan chips because it will be the only way to compete on price for model usage.'
Meta is working with Broadcom on chip design and Taiwan Semiconductor Manufacturing Co (TSMC) for fabrication. The company aims to release a new chip approximately every six months through 2027 — a faster cadence than the industry norm of a year or more.
Seven gigawatts in 2026
Meta plans to deploy 7 gigawatts of computing infrastructure this year alone. It added 1 gigawatt in the first half of 2026 and forecasts adding another 5.5 gigawatts by year-end. One gigawatt of energy is enough to power approximately 800,000 homes. The company expects to double capacity again next year to 14 gigawatts in 2027.
Meta expects to spend as much as $145 billion on AI infrastructure this year, a significant portion of Big Tech's projected $700+ billion outlay on the technology. To support expansion, Meta has secured long-term supply agreements with Samsung Electronics for memory chips, Sandisk for flash storage, and Sumitomo Electric for fiber-optic equipment.
The chip is designed to augment — not replace — the large quantities of GPUs Meta purchases from Nvidia and AMD. However, the memo acknowledged that adopting the latest GPUs 'has been a heavy lift, and it has cost us time.'
Shares rose 4.6% on the day after the company also announced developer access to an AI coding model, directly competing with OpenAI and Anthropic.
