Meta's New AI Chips to Enter Production in September
Meta plans to start manufacturing its latest AI-specific chips in September to reduce reliance on GPU suppliers like Nvidia, according to an internal memo.
Meta is on track to begin production of its latest AI-specific chips in September as it seeks to lower GPU costs amid a severe component shortage, Reuters reported, citing an internal memo. At least one chip passed its testing phase in about six weeks, the memo said.
The company is working with Broadcom on chip design and will use Taiwan's TSMC for manufacturing. Meta is also buying RAM from Samsung, storage from Sandisk, and fiber optic equipment from Sumitomo Electric.
In March, Meta detailed four new chips developed under its Meta Training and Inference Accelerator (MTIA) program. Some are already in deployment, while others will be introduced this year or next. The company is taking a modular approach to chip design, anticipating that needs will evolve as AI advances rapidly by the time chips are in production.
"Each MTIA generation builds on the last, using modular chiplets, incorporating the latest AI workload insights and hardware technologies, and deploying on a shorter cadence," the company wrote at the time.
The chips are expected to help Meta save on buying GPUs from Nvidia and AMD, though the company still plans to spend heavily with those suppliers. Meta intends to use MTIA chips for training models for its ranking and recommendation algorithms, broader AI workloads, and inference for its applications.
Meta has been producing its own AI chips since 2023. The company has been spending heavily to secure computing capacity for its AI efforts. In April, Meta said capital expenditures would be between $125 billion and $145 billion this year, much of it directed at AI.
The company has been striking data center and power deals worldwide, spending tens of billions to secure computing capacity for its new Muse Spark series of AI models. It plans to deploy 7 gigawatts of compute this year and double that next year, according to the memo.
Last year, Meta signed a deal with ARM to secure compute for its recommendation systems, a multi-billion deal with AMD for its Instinct GPUs, and a multi-billion deal with Amazon to use its homegrown CPUs for AI needs.
Meta is not alone in trying to reduce capital flowing to Nvidia. OpenAI last month unveiled an inference processor built with Broadcom, and Anthropic is reportedly considering developing its own chips with Samsung. Amazon and Google both develop their own chips for AI training and inference, and a host of startups are building in the space to meet surging demand. Meta declined comment.
