-
Qualcomm is taking a quite different approach from Nvidia, Intel and AMD on AI
-
It is mainly building AI into its device chipsets for a multitude of gadgets
-
It does, however, build the Cloud AI 100 chip for AI workloads on the edge cloud
Where is Qualcomm in the artificial intelligence (AI) silicon scramble?
While Nvidia is making headway in the AI-cloud infrastructure nexus, with AMD and Intel following in its wake. Qualcomm told Silverlinings that it is mainly currently centered on providing AI for the multitude of gadgets it already produces Snapdragon and other chipsets for.
“Our primary focus is around on-device AI, specifically across smartphones, laptops, vehicles, XR devices and IoT devices,” a Qualcomm spokesperson told us in an email.
Although Qualcomm noted that it does produce the Qualcomm Cloud AI 100 chip, which is designed for AI inference acceleration on an edge cloud in a data center. The AI 100 addresses “unique requirements in the cloud, including power efficiency, scale, process node advancements, and signal processing,” the spokesperson told us.This facilitates the ability of data centers to run inference on the edge cloud faster and more efficiently, the spokesperson said.
“I think Qualcomm’s position is nascent in ‘cloud’ infrastructure,” neXt Curve executive analyst Leonard Lee told us on LinkedIn. “It would also take quite an investment for them to build the channels and the organization to begin competing with Nvidia, AMD, and Intel in the hyperscaler server and AI supercomputing market,” he said.
Lee noted that is clear from neXt Curve research that AI workloads continue to proliferate across edge environments and on endpoint devices for industrial customers, enterprises and consumers.
“Qualcomm is much better positioned here and has been advancing what you might call edge infrastructure AI computing with its Cloud AI 100 accelerator cards that have proven quite performant for inference workloads,” Lee said.
So AI in devices and at the cloud edge looks to be Qualcomm’s play in this market, at least for now.
Want to discuss AI workloads, automation and data center physical infrastructure challenges with us? Meet us in Sonoma, Calif., from Dec. 6-7 for our Cloud Executive Summit. You won't be sorry.