Meta has begun testing its first in-house chip for training artificial intelligence (AI), created to reduce reliance on external suppliers like Nvidia, according to Reuters on Tuesday.
The Facebook parent company has begun a small deployment of the chip and plans to ramp up production for wide-scale use if the test goes well. The tech giant also hopes that it would help bring down significant infrastructure costs as it places expensive bets on AI to drive growth.
Meta, which also owns Instagram and WhatsApp, has forecast total 2025 expenses of $114 billion to $119 billion, including up to $65 billion in capital expenditure largely driven by spending on AI infrastructure.
According to reports, Meta’s new training chip is a dedicated accelerator, meaning it is designed to handle only AI-specific tasks. This can make it more power-efficient than the integrated graphics processing units (GPUs) generally used for AI workloads. Meta is working with Taiwan-based chip manufacturer TSMC to produce this chip which recently announced a $100 investment in the U.S.
This chip is the latest in the Meta Training and Inference Accelerator (MTIA) series. While the program initially had a wobbly start, last year an MTIA chip was used by Meta to perform inference, or the process involved in running an AI system as users interact with it, for the recommendation systems that determine which content shows up on Facebook and Instagram news feeds. Meta executives have said that they want to use their own chips for training by 2026.
“We’re working on how we would do training for recommender systems and then eventually how do we think about training and inference for gen AI,” Meta’s Chief Product Officer Chris Cox said at the Morgan Stanley Technology, Media & Telecom 2025 conference last week.
As of now, most tech companies continue to rely on Nvidia for AI chips. According to Nvidia’s latest earning report, Nvidia surged to the position of the second most valuable companies. However, tech companies are looking for other alternatives as well.
Source: Reuters
Bd-pratidin English/ Afia
