AI is currently driving the tech industry, from both the demand end and supply. Everyone is trying super hard to get it right and lead the industry. Nvidia is one of the tech giants that are sitting on top with its AI chips and technology. Keeping that thought in mind, industry leaders including Intel, Google, Microsoft, and Meta are forming the Ultra Accelerator Link (UALink) Promoter Group. Why is that? Why are they collaborating? Let’s find out.
The group comprises AMD, Hewlett Packard Enterprise, Broadcom, and Cisco, though notably not Arm. So the main reason behind this group is that the members want to standardise the components linking AI accelerator chips in data centres. They want to create an industry-wide standard to enhance the connectivity of AI accelerators—specialised chips designed to expedite AI model processing.
Also read: Nvidia unveils ‘world’s most powerful’ AI chip: All details here
So this group will create a consortium called the UALink Consortium, in Q3. This will then take charge of the development of UALink spec going forward. We might get to see the UALink 1.0 around the same time and the UALink 1.1, which will have higher bandwidth, will come in Q4 2024.
Also read: Nvidia’s Project GR00T will help develop AI-powered robots: Here’s how
If you take a look at the members list, you will note that Nvidia is missing. And why is that important? Currently, Nvidia is leading when it comes to specialised chips for AI. Its market share has also grown substantially.
So this can be a joint effort to deal with Nvidia’s market presence. And the reason why Nvidia is not a part of this is probably because it offers its own proprietary interconnect tech for linking GPUs within a data centre server. So it is not interested in adding to a rival tech.
It will take some time for them to finally offer something, until then let’s wait and see.
