2 Sources
[1]
No-Nvidias networking club seeks open GPU interconnect tech
Ultra Accelerator Link consortium promises 200 gigabits per second per lane spec will debut in Q1 2025 The Ultra Accelerator Link Consortium - an alliance of enterprise tech vendors that pointedly excludes Nvidia because it wants a shared standard for accelerator-to-accelerator links - has opened
[2]
UALink Consortium officially incorporates -- NVLink competitor headed by AMD and Intel opens doors to contributor members
The Ultra Accelerator Link (UALink) Consortium has officially incorporated, which means it's now a legal entity. The Consortium seeks to create a new standard for high-speed, low-latency communication across servers in AI data centers, and has board members from AMD, Intel, Meta, Hewlett-Packard
Share
Copy Link
The Ultra Accelerator Link (UALink) Consortium, comprising major tech companies, has officially incorporated to develop an open standard for high-speed AI accelerator interconnects, aiming to compete with Nvidia's proprietary NVLink technology.

The Ultra Accelerator Link (UALink) Consortium, a coalition of enterprise tech giants, has officially incorporated as a legal entity. This marks a significant step in their mission to establish an open industry standard for AI accelerator interconnects
1
2
. The consortium, which notably excludes Nvidia, includes industry leaders such as AMD, AWS, Broadcom, Cisco, Google, HPE, Intel, Meta, and Microsoft.The formation of UALink is seen as a direct challenge to Nvidia's stronghold in the AI networking market. Nvidia's proprietary technologies, including InfiniBand and NVLink, have contributed to its annual networking business revenue of over $14 billion
1
. The consortium aims to create an open alternative to these proprietary products, potentially disrupting Nvidia's market dominance.UALink has announced plans to release its version 1.0 specification in the first quarter of 2025. The consortium claims that this specification will enable:
1
2
The development of an open standard for AI interconnects could have far-reaching implications for the industry:
1
.UALink is actively seeking new contributor members to support its mission. Willie Nelson, president of the UALink Consortium, has encouraged interested companies to join
2
. The consortium's board already includes representatives from major tech companies, and Samsung has expressed interest in joining as a contributor member2
.Related Stories
Alongside UALink, another industry initiative called Ultra Ethernet is challenging Nvidia's dominance in data center networking. AMD recently announced the industry's first Ultra Ethernet-ready 400GbE card, with the Ultra Ethernet standard also set for release in Q1 2025
2
.Despite these developments, Nvidia remains confident in its position. CEO Jensen Huang dismissed UALink as a threat, stating, "By the time the first gen of UALink comes out, we will be at NVLink seven or eight"
1
. However, Nvidia has shown some openness to industry standards, as evidenced by its Spectrum X Ethernet technology being used in Elon Musk's xAI 100,000-GPU AI training cluster1
.As the AI industry continues to evolve rapidly, the incorporation of UALink and the development of open standards like Ultra Ethernet signal a potential shift in the landscape of AI interconnect technologies. The coming years will likely see increased competition and innovation in this critical area of AI infrastructure.
Summarized by
Navi
[1]
09 Apr 2025•Technology

15 Jan 2025•Technology

13 Mar 2026•Technology

1
Technology

2
Technology

3
Policy and Regulation
