Apple has joined a consortium creating next-gen technology to link together chips in AI data centers.
The consortium, the Ultra Accelerator Link Consortium, is developing a standard called UALink, which connects the AI accelerator chips found within a growing number of server farms. As of Tuesday, Apple is a member of the consortium’s board, along with Alibaba and semiconductor company Synopsys.
In a statement, Apple director of platform architecture Becky Loop said that UALink “shows great promise” in addressing connectivity challenges and creating new opportunities for expanding AI capabilities and demands.
“Apple has a long history of pioneering and collaborating on innovations that drive our industry forward,” Loop added, “and we’re excited to join the UALink board of directors.”
UALink aims to connect chips ranging from GPUs to custom-designed solutions to speed up the training, fine-tuning, and running of AI models. Based on open standards including AMD’s Infinity Fabric, the first UALink products are expected to launch in the next couple of years.
Intel, AMD, Google, AWS, Microsoft, and Meta are among the UALink consortium’s members. Nvidia, which is by far the largest producer of AI accelerators, is not. That’s perhaps because Nvidia offers its own proprietary interconnect tech, NVLink, for linking chips within a data center cluster.
Apple’s participation in UALink comes as the company increases its investments in infrastructure to support Apple Intelligence, its suite of AI product features. According to The Wall Street Journal, Apple is reportedly developing a new server chip to improve the efficiency of its AI data centers.
Some of Apple Intelligence’s capabilities have been met with mixed reviews. (My colleague Sarah Perez called them “boring and practical.”) Last week, Apple said it would update one of those capabilities, AI-summarized news alerts, after users reported seeing inaccurate headlines, including that tennis star Rafael Nadal had come out as gay.
