In an effort to lower GPU costs amid an unprecedented component shortage, Meta is on track to start making the latest AI-specific versions of the chip in September, Reuters was mentionedciting internal memo.
At least one chip made it through its testing phase in about six weeks, the note said. Meta is working with Broadcom to design the chips, but will use Taiwan Semiconductor Manufacturing Company (TSMC) to manufacture them. It also buys RAM from Samsung, storage from Sandisk and fiber optic equipment from Sumitomo Electric, according to the report.
After detailed the four new chips, developed under the Meta Training and Inference Accelerator (MTIA) program, in March, some of which are currently in development or will be this year or next. The company is taking a modular approach to designing these chips, anticipating that their needs will change as artificial intelligence rapidly evolves by the time the chips are in production.
“Each MTIA generation builds on the last, using modular chiplets, incorporating the latest AI workload insights and hardware technologies, and developing at a slower pace,” the company wrote at the time.
The chips are expected to help the company save on buying GPUs from chipmakers such as Nvidia and AMD, though it still expects to spend big with those providers as well, Reuters reports. Meta plans to use the MTIA chips to train models for its ranking and recommendation algorithms, broader AI workloads, and inference targeting its applications. The social networking company was has been producing its own AI chips since 2023.
Meta has spent heavily to secure enough computing capacity to power its various AI efforts. The company said so in April is waiting capital expenditures of between $125 billion and $145 billion this year, much of which is earmarked for its artificial intelligence efforts.
The company has struck deals for data centers and power around the world, spending tens of billions to secure computing capacity to train and develop its new Muse Spark series of AI models. It plans to deploy 7 gigawatts of computing this year and double that next, according to Reuters, which cited the memo.
It also signed a deal with ARM last year to provide computing for its recommender systems, in addition to a multibillion-dollar deal with AMD on its Instinct GPUs and a multibillion-dollar deal with Amazon to use the cloud giant’s domestic CPUs for AI-related needs.
Meta isn’t the only company trying to stem the tide of funds going to Nvidia. OpenAI last month unveiled an inference processor it’s building with Broadcom, and Anthropic is said to be considering developing its own chips with Samsung. Amazon and Google are both developing their own chips for AI training and inference, and there are a number of startups forming in the space to meet the growing demand.
Meta declined to comment.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
