TikTok owner ByteDance will design its own server SoCs


ByteDance, owner of social media app TikTok, has started hiring specialists to develop its data center system-on-chips (SoC). ByteDance will join several Chinese tech companies that have recently deployed their own data center SoCs designed to run on specific workloads.

ByteDance has posted 31 chip-related openings on its website so far, reports TMTPost. Thirty-one chip engineers will barely be able to build an advanced processor, but what they could do is create a set of design requirements for contract chip developers, like Alchip. Then, once a contract chip developer creates a suitable design, a contract semiconductor manufacturer (like TSMC) will mass-produce it for ByteDance.

For now, ByteDance only indicates that it plans to develop its chips without giving details. Typically, chip design cycles take years, so we won’t learn about the company’s SoCs until a few years later.

What exactly ByteDance plans to develop is a big question. CNBC (opens in a new tab) A company spokesperson said the social media giant’s datacenter-grade system-on-chip will support a variety of business areas, including video platforms, information and entertainment applications. Meanwhile, multiple China-based sources (such as) say ByteDance is eager to design AI chips.

ByteDance’s core business is its multimedia application TikTok, so it would make sense for the company to develop a SoC similar to Google’s Argos video (trans)coding (VCU) to transcode video streams uploaded by TikTok users. Meanwhile, to further reduce the power consumption, traffic and storage capacities used by TikTok data centers, ByteDance could introduce brand new codecs or AI algorithms. Also, adding AI enhancements could improve image quality.

ByteDance will be another China-based cloud giant developing its chips. For example, Alibaba has its own AI inference chip and versatile processor. Meanwhile, Baidu has its own Kunlun AI processor. Tencent also has its processors for various cloud workloads.


Comments are closed.