Hello and welcome to News Center. I am Arsh Mitchell Kumar. Now, Qualcomm Snapdragon Summit 2026 put agentic AI center stage while pushing the boundaries of mobile performance, power efficiency, and on-device intelligence. But what does the summit mean for the market and the investors at a time when Nvidia dominates the AI market? And what advantages does this create for companies like Qualcomm? To answer some of those very specific questions, I'm now joined by Akash Palkhiwala, the CFO and the CEO of Qualcomm Inc. Mr. Palkhiwala, thank you so much for joining us. Firstly, what is the biggest message, the biggest takeaway that you're sending out to investors, to the market through the announcements that you've made at the summit?
This summit for us is really about kind of highlighting the technologies that Qualcomm is going to bring to consumer devices over the next several years. And the central theme of the conference is around agentic AI. I think you're seeing this big transition in AI where we started off really thinking of it more like a search use case where you'd ask a question and you'd get back an answer from ChatGPT or any other model that you would use. But we're seeing this significant transition happening where rather than use it in a question and answer mode, I think we're going to end up using agents and very capable agents who can do significant tasks for you. And these are agents that are going to be inherent on the devices that the consumers use. And these agents will have context and a lot of information around kind of location of the consumer, the data that the consumer has on the phone, other sensors that are in the device, and be able to leverage all those things to do full tasks for the consumer. And it's something that'll be spread across smartphones, PC, wearable devices, automotive, industrial, and really data center as well. So, very excited about what we're bringing to the table and we see these technologies being deployed in hundreds of millions of devices over the next couple years through Qualcomm Snapdragon platform.
Well, when we speak of AI agents, I can't help but ask — a lot of conversation recently has revolved around safety and managing these AI agents, managing the risks associated with these AI agents. How are you looking at that debate?
I think this is one of those areas where the conversation itself to me is a very significant positive because it's obvious that AI has a lot of use cases that are very positive for all of us. And those things will get deployed one way or the other. Of course, there are risks that come with any new technology and AI is just one example, the current example on this conversation. And the fact that there is a debate on both sides will really ensure that we end up doing this in a responsible way. The end point in my mind is pretty decided that there'll be broad deployment of AI across devices, across the cloud and this debate will ensure that we do it in the most responsible possible way. So, I think of this as a positive, as something that's encouraging and over time we'll find the right way to deploy it.
Well, speaking of risks, do you see these risks as perhaps a headwind because we see a lot of tech titans suggest that perhaps it's time to slow down, but this comes at a time when Qualcomm also is initiating its data center push?
Yeah, I think the way we see it is this conversation is relevant to new models and new technologies coming in. But there is a lot of catch-up that use cases have to do to what is already available today. And if you think about enterprises, enterprises are really in the first innings of deploying something that would get leveraged using the AI model capabilities. And even if you second for a moment imagine that further model improvements are delayed, the existing models are going to give us a very long runway in terms of use cases and improvements that those models can deliver. And so I think we're going to see the adoption continue at a rapid pace. It's going to be on the device side on consumer devices, but it's also going to be in enterprises. As an example, I'll give you — at Qualcomm, we use approximately 4 trillion tokens per week today to enable use cases through AI. This is to write software, to design chips, to serve and support our customers. And we think there's a long runway of use cases that can be expanded on. So, I'm very optimistic that there is not going to be an impediment that this conversation puts on the adoption. And I think we're excited to work with the ecosystem to drive these new use cases.
When you talk of this data center opportunity, this is of course a key thrust area for Qualcomm. Again, you were giving us what the size of that opportunity looks like from a Qualcomm lens.
I mean, you look at the TAM of the market. It's a lot around $1 trillion, right? And the aspiration that we outlined at our investor day is to be 5% of that, which would be $50 billion. So, tremendous opportunity. This market is expanding in a big way. The customers are looking for alternate suppliers, but they want people who have technology expertise, who have scale, and those things fall right in Qualcomm's sweet spot. So, we'll be leveraging those things and really deploying our data center solutions across connectivity, custom chips, CPU, and AI accelerator. And it's just the beginning, I think, of our journey in this massive, massive market.
I'll just labor on that point a little further. Nvidia dominates the AI accelerator market. Tell us what is Qualcomm's specific advantage? How does it seek to compete with a giant the size of Nvidia?
Yeah, so I think you think about the AI accelerator market today, as you said, it's largely dominated by GPUs. But one of the changes that is happening in the architecture of data centers is the need for disaggregated computing. So, rather than use a GPU for everything, you're going to use a solution that is optimized for prefill, for decode within inference. Depending on the workload, you might want a solution that serves a different purpose versus a general purpose GPU. And so, we're seeing this big trend towards this disaggregation, and the evidence of that comes through Nvidia's acquisition of Groq or the success that Cerebras is having. And that's the transformation of the market that's happening. From Qualcomm's perspective, what we're bringing to the table is this expertise that we've developed on the edge devices of delivering high performance at very low power. Power is a problem statement in data centers, and this kind of expertise that we have on the edge, when we bring it to the cloud, it's something that's very interesting to players, to the customers. We have also developed this technology called high bandwidth compute, which allows us to deliver a solution that has very high memory bandwidth, which also is another problem statement for data center customers. And so, the proof of whether Qualcomm is building the right things is in the traction that we're getting with customers and the very close engagement that we have. And based on all of these things, we've set a target of about $5 billion in revenue next year versus $300 million this year. So, just a rapid ramp in one year because of the traction we're seeing.
