- Apple joined the board of directors of the Ultra Accelerator Link consortium
- Bonding is a key technology that binds GPUs together, much like neuron synapses
- UALink emerges as the biggest rival to Nvidia’s proprietary NVLink
In June 2024, we reported how a number of big tech names had banded together to form the Ultra Accelerator Link (UALink) Promoter Group, a strategic initiative aimed at reducing Nvidia’s dominance in the accelerator market. AI.
Directly competing with Nvidia’s proprietary NVLink technology, UALink seeks to develop a new industry standard for high-throughput, low-latency communication for scalable AI systems in data centers. It already has support from Intel, AMD, Google, Microsoft, Meta, HPE, Cisco and Broadcom, but Apple has now also joined the ULink board.
“UALink shows great promise in addressing connectivity challenges and creating new opportunities to expand 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 advance our industry, and we are thrilled to join the ULink Board of Directors. »
ACDC Project
By joining the consortium, it seems likely that Apple plans to use UALink technology in “Project ACDC” (Apple Chips in the Data Center), also known as “Baltra”.
This initiative, in collaboration with TSMC and Broadcom, aims to develop proprietary AI chips for Apple’s data centers, strengthening the capabilities of its new Apple Intelligence.
Apple isn’t the only new company to join the consortium; other supporters include Alibaba Cloud Computing and Synopsys.
Scheduled for the first quarter of 2025, the UALink 1.0 specification will enable scalable connection of up to 200 Gbps per lane for up to 1,024 accelerators within an AI module.
“We are pleased to welcome Alibaba, Apple and Synopsys to the UALink Consortium Board of Directors,” said Kurtis Bowman, Chairman of the UALink Consortium Board of Directors. “Since our incorporation, the Consortium now has more than 65 members in total, spanning cloud, silicon and IP providers, software vendors, systems OEMs and others. Continued support for the Consortium will help accelerate adoption of this key industry standard, defining next-generation interconnection for AI workloads.
