AI

Meta is reportedly testing in-house chips for AI training

Meta is reportedly testing an internal chip for training AI systems, part of a strategy to reduce dependence on hardware makers such as Nvidia.

According to ReutersMeta’s chip, which is designed to handle AI-specific workloads, was manufactured in collaboration with Taiwan-based company TSMC. The company controls a “small bet” of the chip and plans to scale up production if the test is successful.

Meta has used previously adjusted AI chips, but only to perform models – not train. As Reuters notes, several of the company’s chip design efforts have been canceled or otherwise reduced after they have not met internal expectations.

Meta expects to spend $ 65 billion in capital expenditure this year, a large part of which will go to Nvidia GPUs. If the company succeeds in even lowering a fraction of those costs by switching to internal chips, it would be a big victory for the social media giant.

Source link

See also  Wall Street analysts explain how AMD’s own stock will pay for OpenAI’s billions in chip purchases 
Back to top button