00:00 Speaker A
Edge AI is like on-device AI and cloud AI. And should people think of edge AI as a complement to the cloud or as an alternative to the cloud? Is it a replacement or an addition to the cloud?
00:16 Akash
Well, if you really look at what computing has looked like over the past few years, some of the computing is done on the phone, and the rest of the computing is done in the cloud. Edge AI is no exception. Some of your AI workloads run at the edge and some in the cloud. It depends on what you’re trying to do, the context and where the data is available. When you think about conversational agent AI talking to a device and getting information from it, glasses are a great example.
00:48 Akash
Well, if you’re wearing smart glasses like Meta Ray-Ban, you can look at something and easily ask, “What is this?” And the model needs to understand that question. You need to understand what the camera is looking at. You can then feed that data into your model. The model then decides whether that question can be answered at the edge itself or whether it needs to go to the cloud. And it’s almost irrelevant to the user. It doesn’t matter if it happens at the edge or in the cloud. This is actually what enables this unique use case using sensor information that is only available at the edge. And as we expand from consumer use cases like this to enterprise use cases to industrial automotive and robotics use cases, I think Qualcomm technology will be built into each of those devices. I think we’re at the very beginning of this kind of next wave of AI innovation. It will start at the edge and continue to the cloud.
01:52 Speaker A
Akash, do you think we are on the cusp of some kind of wearable revolution? I feel like there have been so many attempts to really popularize wearables, but they just haven’t caught on with the masses. Has that changed now?
02:12 Akash
Well, I think things are definitely changing. So let’s think about devices that people wear, that can see what they can see, hear what they can hear, and can talk to an agent AI. I think glasses are a very easy way for people to understand this concept. Ah, but actually think of it as something broader than glasses. It could be a watch, earphones with a camera, or a necklace.
02:47 Akash
But in any case you have a device with a camera that can have the same field of view as you have. I can hear the same thing as you. As you go through the day, you’ll learn about your daily happenings, and later in the day you’ll be able to actually answer questions. So one of the great use cases that I’ve heard recently is that one of our partners in India is implementing a payment use case where you can look at a QR code on your glasses and say, I’m going to pay 100 rupees for this code. The code is read and sent to your e-wallet to make the payment. This is very natural for users. We believe that by integrating these use cases, this revolution will begin.
03:41 Speaker A
Sometimes you just want a little more friction between yourself and yourself when making a big purchase. But I’m sure they can be incorporated. Hmm, I’m interested in competition in this field. Obviously, Qualcomm is already the leader in smartphone chips and device chips. Of course, as you know, other chip manufacturers are also getting into this. Nvidia is also getting into this space. Do you think Nvidia is a direct competitor in edge computing, or does it operate in a separate subsegment?
04:14 Akash
Well, when you think about edge computing, I think Qualcomm has a huge advantage. One of the big things about edge devices is that they are battery powered. This requires very low power and very high performance chips. That’s Qualcomm’s DNA. We did it over the phone. We’ve done similar things with a variety of other devices. Although we have competitors, we are the clear leader in this space and we expect to continue to lead in this space. I think Qualcomm’s technology portfolio is unique and one-of-a-kind, making it a perfect fit for their devices. So, um, I’m very hopeful and optimistic about what we can do going forward.
