FB reportedly working on custom server chips

Tech giant Facebook is reportedly developing a chip that powers machine learning (ML) for tasks such as recommending content to users.

According to The Information, Facebook is developing a suite of custom chips for its data centres.

One of the processors is reportedly designed to power ML tasks, including the algorithm that handles Facebook's content recommendations, while another would assist in transcoding videos to improve the quality of livestreamed videos.

In addition to helping reduce its dependence on outside chipmakers, the move to custom silicon could help Facebook lower the carbon footprint of its data centres, reports Engadget.

The new chips would work alongside the third-party processors the company currently uses in its servers and reportedly aren't meant to completely replace what it already has.

"Facebook is always exploring ways to drive greater levels of compute performance and power efficiency with our silicon partners and through our own internal efforts," a spokesperson for the company told Engadget.

"We have nothing new to share on our future plans at this time," the spokesperson added.

The report also mentioned that the company has already dabbled in semi-custom chips.

In 2019, for instance, it announced it was working on an application-specific integrated circuit (ASIC) for video transcoding and inference work.