Meta to Mass-Produce In-House AI Chips From September

Meta Platforms will begin mass production of its self-designed artificial intelligence chips in September, a move aimed at reducing its reliance on Nvidia and Advanced Micro Devices and lowering surging AI infrastructure costs.

Reuters reported on July 9, citing an internal company memo, that Meta plans to start mass-producing the chips in September.

Meta in March unveiled four in-house AI chips, including the MTIA 400, which was introduced under the code name "Iris." The MTIA 400 is part of the fourth-generation Meta Training and Inference Accelerator project. Broadcom designed the chip, and Taiwan Semiconductor Manufacturing Co. will produce it.

The chip push comes as competition to secure AI semiconductors intensifies. Meta's internal memo said obtaining the latest graphics processing units had been difficult and time-consuming for large companies.

Meta is also ramping up AI infrastructure investment. It plans to build 7 gigawatts of computing infrastructure this year. The company added 1 gigawatt in the first half and expects to add another 5.5 gigawatts by year-end. Next year, it plans to add 7 more gigawatts, bringing total capacity to 14 gigawatts.

The company's capital expenditure forecast for this year stands at $145 billion, with most of that set to go toward AI infrastructure. That amounts to about one-fifth of the $700 billion in projected capital spending by big tech companies this year.

Meta is also working to secure its supply chain. The company said it signed long-term supply contracts for Samsung Electronics memory chips, SanDisk flash storage and Sumitomo Electric's optical fiber equipment business.

Meta has also begun monetizing software. It released its AI coding model, "Muse Spark 1.1," to developers through a paid API. The launch marks the first time Meta has generated revenue from its own AI model, diverging from its open-source Llama strategy. Pricing is set at $1.25 per 1 million input tokens and $4.25 per 1 million output tokens. That is higher than OpenAI's lower-cost models but below Anthropic's premium offerings.

Meta shares, which had been down earlier in the session, rebounded after the release of Muse Spark 1.1. The stock closed up 4.7% from the previous day at $631.5.

Lee Song-ryeol, Hankyung.com reporter yisr0203@hankyung.com