Meta Platforms Inc, which recently reached major deals with top chipmakers, still has ambitions for its own custom silicon, including plans to develop processors that can train future artificial intelligence models, the company’s chief financial officer said.
"Some of our workloads really are very customised to us,” Meta CFO Susan Li said Wednesday at a technology conference sponsored by Morgan Stanley. "The sort of ranking and recommendations workloads have been where we have started, and that’s the place where we have rolled out custom silicon at the most scale. But we expect and are hopeful that we are going to expand that over time, including eventually to training AI models.”
While not a cloud computing provider, Meta is one of the biggest operators of data centres used to train and run AI models. In the past few weeks, it has struck massive agreements with market leader Nvidia Corp and rival Advanced Micro Devices Inc for the chips and equipment that power AI workloads. The social media parent has also continued work on its in-house AI processors.
Li said the company is buying different types of chips to handle different types of tasks. "Based on what we know today and our current needs, what do we think is the best chip to use for each of these use cases?” Li said. "Custom silicon is a big part of that. – Bloomberg
