Meta tests its first AI training chip
Meta is testing its first in-house chip for training artificial intelligence systems to reduce the company's dependence on external suppliers such as Nvidia in the future.
Meta is testing its first in-house chip for training artificial intelligence systems to reduce the company's dependence on external suppliers such as Nvidia in the future.
If the testing is successful, Meta plans to scale up production for large-scale use, the sources told Reuters, part of the company’s long-term plan to reduce its huge infrastructure costs.
Meta forecasts total spending for 2025 of up to $119 billion, including up to $65 billion in capital expenditures, largely driven by spending on AI infrastructure.
One source said the new Meta learning chip is a dedicated accelerator, meaning it’s designed to perform only AI-specific tasks. This could make it more power-efficient than integrated graphics processing units (GPUs) typically used for AI workloads.
Meta is working with Taiwanese chipmaker TSMC. The test deployment began after the company completed the first tape-out stage, a major sign of success in chip development that involves sending an initial design to the chipmaker's factory, another source said.
A typical test costs tens of millions of dollars and takes about three to six months, and there is no guarantee that the test will be successful. If it fails, Meta will be forced to diagnose the problem and repeat the write phase.
Meta executives said they want to start using their own chips by 2026 for training, or the laborious process of feeding an AI system a bunch of data to “teach” it how to work. The goal for the training chip is to start with recommender systems and later use it for generative AI products like the Meta AI chatbot.
China plans to encourage the use of open-source RISC-V chips across the country
China plans to encourage the use of open-source RISC-V chips across the country
A shortage of graphics chips forced OpenAI to suspend the deployment of the new GPT-4.5 model
A shortage of graphics chips forced OpenAI to suspend the deployment of the new GPT-4.5 model
Trump announced that Taiwanese microchip giant TSMC will invest $100 billion in manufacturing in the U.S. What does this mean for the tech industry?
Trump announced that Taiwanese microchip giant TSMC will invest $100 billion in manufacturing in the U.S. What does this mean for the tech industry?
Read the country's main IT news in our Telegram
Read the country's main IT news in our Telegram
