Apple joins UALink group tasked with taking on Nvidia’s AI hardware dominance

Apple has officially gained a board seat on the Ultra Accelerator Link Consortium, a group of more than 65 members developing next generation AI accelerator architecture. Apple’s involvement will allow the company to influence the new standard and push for its adoption. That basically means that Apple will be involved in creating and promoting a standard for connecting a bunch of GPUs together in data centers powering AI tasks.

“UALink shows great promise in addressing connectivity challenges and creating new opportunities for expanding AI capabilities and demands,” said Becky Loop, Director of Platform Architecture at Apple. “Apple has a long history of pioneering and collaborating on innovations that drive our industry forward, and we’re excited to join the UALink Board of Directors.”

What exactly is UALink? Simply put, UALink is an industry effort to compete with Nvidia’s dominance in AI hardware. Specifically, the standard aims to compete with Nvidia’s proprietary NVLink technology. This piece from Tom’s Hardware last fall explains:

UALink seeks to become the industry’s open standard for scale-up connections of many AI accelerators, and to become a competitor with Nvidia’s proprietary NVLink. NVLink, Nvidia’s solution for GPU-to-GPU communication in servers or pods of servers, uses Infiniband — another effectively Nvidia-owned communication technology — for higher-level scaling. Infiniband is being challenged by newcomer Ultra Ethernet, which is another major consortium of tech giants creating an open standard to counter Nvidia’s dominance.

Backed by AMD and Intel, the UALink Consortium includes promoter members including AWS, Google, Meta, and Microsoft. Now Apple is involved as well.

“The initial 1.0 version 200Gbps UALink specification enables the connection of up to 1K accelerators within an AI pod, and is based on the IEEE P802.3dj PHY Layer,” according to the UALink Consortium. “The specification will be available to Contributor Members in 2024, and will be released to the public during the first quarter of 2025.”