Editor’s note: Qualcomm provided travel, lodging and other accommodations associated with Snapdragon Summit.
HAWAII—With the launch of its ambitious Oryon CPU for PCs and plan to embed the 12-core powerhouse across its business segments, Qualcomm is changing. Whereas previous iterations of the annual Snapdragon Summit have put 5G and Wi-Fi front and center, the event this year is focused on leveraging longstanding cross-domain expertise to capitalize on the rise of generative AI. That’s not to say Qualcomm’s connectivity pedigree is losing relevance; quite the opposite really given the need for robust, reliable network access in increasingly distributed cloud environments, not to mention the latency gains and associated applications opened up by on-device AI.
But back to the point around the company’s evolution—CEO Cristiano Amon put it succinctly in a tweet plugging Fast Company coverage of the Snapdragon X Elite launch. He described “our evolution from a communications company to a connected processing company.”
This is also a bit of an evolution of corporate messaging which has for a few years focused on Qualcomm’s role at the “connected intelligent edge.” The concept there is around the correct belief that all devices will at some point be connected to the cloud and will require high performance, power efficient computing—all things Qualcomm does very well. And that applies to smartphones, to PCs, to vehicles, and to any type of device that fits under the umbrella of the internet of things.
Enter AI, another broad technology category that Qualcomm has been working on for a decade-plus. If you believe AI will become table stakes for consumer and enterprise devices of all sorts, then you need to start parsing where AI workloads are run. Obviously data centers can support whatever it is you’re trying to do with AI but that gets expensive and is fundamentally inefficient. In that context, any amount of AI that can be run on a device, that’s probably the way to go. And as Qualcomm has driven home at its showcase in Maui, their compute and mobile platforms can run a hell of a lot of AI on a device; think 10 billion parameter models on smartphones with its Snapdragon 8 Gen 3 mobile platform and triple that with its Snapdragon X Elite PC platform.
Patrick Moorhead, an excellent analyst and hell of a nice guy, broke it down also on Twitter, where I apparently source a good bit of my coverage. He hit on the latency point and also drew in privacy. He wrote: “It’s easy to justify on-device gen AI for PCs and phones…We have apps today to improve the experience by reducing latency…the industry has tried and failed to deliver predictable [and] flexible experiences streamed from a data center. This is why it’s such a niche market. Hence the Windows, iOS and Android app store existence. Foundational models up to 10 [billion] parameters will run faster on-device versus streamed from the cloud. This is gen 1.” Then, on privacy, “New gen AI apps will record every second of your life. Your phone calls, your chats, your video calls, what you look at on the internet. Think people will want to upload all that to the cloud? Think again.”
Tying this all together, and giving a nod to the branding exercise the company is going through with its focus around differentiation and awareness of its Snapdragon portfolio, Qualcomm SVP Kedar Kondap said, “We are at the dawn of a new age…We are bringing powerful generative AI to the device to create innovative new experiences. We are fortifying Snapdragon’s position as the platform for the next generation of AI and computing.”
