So Cristiano, you run Qualcomm and your technology powers many Android phones, Meta smart glasses, a bunch of coming wearables, and a growing number of cars. And I want to get into all of that, but I want to start with something else, which is you also have a very interesting view of the US and China. Qualcomm has a very strong Snapdragon brand in China. That's something I've actually learned a lot being here at the summit that you guys are having — is just the rabid fan base for Snapdragon in China. And I think that gives you a unique window into the two markets. And I want to start with how people in China are feeling about AI relative to the United States. The sentiment, when you look at the sentiment and how it polls on AI around the world, it's very different based on geographies. And maybe could you contrast the two a little bit?
If I do a compare and contrast, the broad adoption of AI into the devices and in consumers, I see that happening in China very, very fast. In the United States, I see the development of frontier models and enterprise adoption of AI. Those kind of dominate the conversation. That's the difference in conversation between both.
And when you talk to people in China about AI, does it feel like they're imagining maybe a different future than people in America? The sentiment in America, when you look at the polling, it's incredibly negative.
Look, every new technology will bring uncertainty, right? So, you know, it's interesting — if you go back in time and you see different technology transitions, some of the concerns obviously... What happened in the United States, given my perspective, is AI has been an important productivity tool for a number of sectors. For example, a company like Qualcomm — I was even asked today in a different conversation, said, 'Well, you guys are doing a lot. You're transforming the company, you're doing a lot of things in parallel. How can you allocate capital to do all of those things in parallel?' And if you actually look, we have been relatively expanding the company with kind of a flat scale of operating expenses. A great part of this is productivity gains using AI in the development process. So you have clearly those aspects of AI for productivity gains, and there are obviously concerns about is that going to impact jobs, or are companies, you know, going to leverage their own AI and they're not going to need as much manpower. So, and that conversation is probably universal. I don't think that's unique to the United States.
You do have this conversation about data centers. Data centers is massive. You look at those massive deployments, you need more and more compute. It's a wall, I think, of the demand for compute. And you obviously have concerns about is this going to increase my electricity bill. So those conversations, I think, are natural. The latter I don't see in China. I see more of it's like a feature, it's an evolution. The two countries are looking at this a little bit differently. I think, for example, China has been really focused on physical AI and everything on the edge. There's a bigger focus on that in China right now. But I believe in general, because of the uncertainty associated with AI, you're going to have a lot of questions. And to be frank, we're in the early days of AI — early, very early. I like to do this metaphor: it's almost like in the early days of the internet. If I had to say, 'Okay, we already know what's going to happen. AltaVista is going to win the search, and MapQuest is going to win the map, and Orkut's going to win the social media' — no. I think we still have a lot of things that need to happen on AI, and some of those questions will be answered.
This episode is brought to you by Mercury, AI-native banking that's loved by more than 300,000 entrepreneurs, including me. Visit mercury.com to learn more. Mercury is a fintech, not a bank. Check the show notes for details. This episode is also brought to you by Jira by Atlassian, where teams and agents get the context, coordination, and control to move work forward. Try it free at jira.com. That's jira.com. Thanks also to Granola, the AI notepad for people in back-to-back meetings. It works everywhere you do and lets you focus on what matters. Try it at granola.ai/sources and use the code sources for 3 months off.
Right now in the US, a lot of people are talking about this idea of pacing, pacing the frontier of AI development. You're seeing things like the Hugging Face breach, the models becoming more capable, cyber risks increasing. And I'm curious where you stand on this, because you, the company, and you — you've been in Washington. You're obviously a voice in the policy space in the US on this. When you're seeing this debate, what do you make of it right now?
So, I'll provide my opinion, and I will — two comments, I think. First, I'll give you my opinion for free. So, it may not be very good. I'm just going to give it for free. Second, it's just an opinion. This is a changing landscape. It's hard to be certain about this. But I think the answer is never on the bookends. The answer is somewhere in the middle. Right. So, first, it's a brand new technology and it has a lot of potential. There is a development roadmap and continuing to invest in the development roadmap. It's going to be important. And the second part of it, it's also important to have guardrails. For example, yes, there's a lot of comments about what are the risks associated with AI, but let me talk about some of the more practical risks. You know, cybersecurity — it is a very reasonable risk because everything is digital. I think the reason AI exists is because we as humans digitize everything. You know, you have all this digital information, every area of computing power, but the surface area that runs on software is very, very big. So if you have models that can find vulnerabilities in things and they can hack things, and you don't need to be, you know, a cyber warfare specialized team — anybody can get a model and do it. It's a problem. You need to have control. You need to have very serious companies. You need to have the right regulation. You have to have the hard guardrails. It's no different than other major technology like that. So, I think that's really the answer. It's not in extremes. It's not like stop it or go full speed. It's do it responsibly. I think there's nothing wrong with that. And have guardrails. In our little world, we do chips. We do the chips. We do chips to do the...
Chips are not little, but that's okay.
It is little compared to some of the other big companies. And we do all of this math. We are building in the hardware hooks and controls and knobs for you to store your information securely, for you to decide whether you want your personal graph to be stored or not, and to be able to have controls. I think that's an important thing. All of those companies need to do it with responsibility.
You traveled to China with President Trump in May, and when you're in those rooms, I'm curious what do you hear from other world leaders about how they're grappling with AI and the risks.
Look, there is obviously a lot of semiconductor... The industry that I happen to be in became a geopolitical industry, I think, right now. So you have to be part of those conversations. AI is important in every part of this conversation. I think there's a common understanding about how to do those things responsibly. And then the other part, more of the conversations that I'm part of, there's a lot of strong, healthy commercial and partnerships kind of relationship between American enterprises and Chinese enterprises. And those relationships — what I have seen, and hopefully I will see next week — those relationships actually create a very strong point of stability and common understanding, I think, between the two countries. And if that's the role that we can continue to play, that's the role we're going to play.
Let's talk about the smartphone. So, it's really interesting. I mean, the smartphone market is obviously massive. It's one of the most consequential, if not the most consequential, computing markets to date. At the same time, it's not growing like it once was. I think I saw there's a stat that it's expected to shrink maybe about 14% in terms of overall units shipped this year, maybe go down a little more next year. And I'm curious to hear from you, how much of that is, you know, the incredible spike in memory costs and component costs for these devices, and just sticker shock that consumers have — and, you know, the phones are good now, and, you know, maybe you don't want the next one, you don't need the next one if the price goes up, you can hang on to the one you've got — and how much of it is something bigger.
Okay, there's a lot in that question. Yes, there's a lot. So let's go break that down a little bit. So first I'm going to give you some of the facts. So, interesting thing, and maybe people don't realize yet: the smartphone market has yet to recover to what the market was before the pandemic. The market is still smaller now — it's artificially smaller, but still smaller than what it used to be in 2019. Right? So the market has been somewhat still recovering from what the market used to be, and there's a number of factors for that. I think one of it is, of course, it's a great thing — everybody in the world has a smartphone, so it became a mature market. It became kind of a replacement rate, and it's totally normal. Look, I'm probably going to show my age, but I started working on 1G. So I've been through every cycle, and at the end we're almost to six.
At the end of the cycle, right, you do get a portion that, you know, the market gets to a state of maturity, becomes more of a replacement rate market. Having said that, the smartphone is a big market. It's 1.2 billion phones sold every year. Even if you don't want any new feature, you use your phone a lot — after two, three, four, five years, you're going to buy a new one. However, within that base, there are a couple of things that happened which have been very telling. The premium and the high tier, as a percentage of the total market, continue to expand, and it's been like clockwork year over year. Even in a compressed market, the premium tier continued to expand. And the reason is because generation after generation, people are doing more with their phone. I'm sure you still have a laptop. The laptop didn't get replaced by the phone. But there are certain things that you say, 'Oh, I need my laptop for this. This long email I'm going to write, I'm just going to wait until I get to the laptop, I'm going to do it.' But now with AI, what we're seeing is more and more workloads shift to the phone. So people want better phones. People want phones that are more capable. So independent of this transition to an AI phone, we're starting to see the premium tier increasing. Okay. The next layer: the market is now artificially suppressed because of what happened with memory. Memory prices went up 5x to 6x. So what happened is the premium tier, the high tier, has been more resilient. Yes, it's more expensive, but consumers see the value, they want to buy the phone. We saw the mid-tier, low tier has been way more sensitive. So the market is down 20% this year because of the high memory prices. It's not a demand problem. There's now pent-up demand in the marketplace. The silver lining on this: it's happening at an interesting time when AI smartphones are starting to come, and it could create opportunity for people who want to buy now an AI smartphone. We have seen this market in cycles. So for the past few years, as you said, not a lot of excitement about the phone market. 'Oh, the phone market does not grow.' We have, because of this driver of premium expansion, we have been growing — you know, in the Android market, double-digit compounded annual growth rate just on more of this premium expansion, share gains, and more content because people want a better phone. But my prediction is nobody wants to talk about the phone market, everybody's data center or their markets, until the AI super cycle. And then when that happens, and 6G happens, I think there will be a lot of conversations about phones. It's important to say phones are not going to go anywhere, and it is the largest development platform ever created by mankind. And humans are not going to travel to a data center, ring the bell, and say, 'Give me some AI.' You're going to use that through the mobile devices, and they're going to have a part in this picture.
I want to know what you mean by AI smartphone. But first, you all showed Muse on the screen during the keynote, which is taking, you know, the Silicon Valley world, the early adopter world by storm. And then there's also products like Instinct and others, and Grokbot and all these things. And I want to stress test what you're saying about workloads moving to the phone, because, I mean, these products are browser virtual machines in the cloud, right? Like Muse is running a browser in a cloud somewhere. It's not on your phone doing most of the token usage, beaming you back the answer, right? So if that's the way that these, at least right now, these early agentic personal products are working, does that not make the phone less crucial?
I have to be careful not to disclose names, but some AI companies right now, some of the companies at the forefront of AI, are telling Qualcomm, 'Qualcomm, by 2028 I need to make sure that I have the ability to run at least a 100 billion parameter model into a phone, and I need to have this thing running kind of all the time.'
So you're saying there's AI companies that are talking about...
Well, AI companies are starting to do phones. Look, it's interesting if I just give you an example of the China phones that we just talked about. The DAB phone is a bytedance phone, you know, made by Nubia.
There's more coming. But I think there's an understanding — there's an understanding that the phone is pretty important for agents. There's going to be the battle of the agents. Which agent you want to do what? That's — I said before, this is early days on the AI, and there's a lot of clarity now that there are things you have to do on the phone. There are things that you have to do on the phone, also supporting other devices. I think it's futile to have a conversation about cloud and edge, because what happens is more and more things you're going to do on the cloud. You're going to find more interesting and sophisticated things that you're going to have to do only on the cloud, and you're going to do on the cloud. And there are going to be things that you're going to have to do on device because you have to do it on the device. There's a lot of discussions about cloud and edge about one competing or replacing the other. That's incorrect. Let me give an example now. Let's say that we pull up our phones and we're going to say, 'Okay, I have 300 apps organized in folders. Let's go one by one and say what is running on the phone and what's running on the device.' It's almost like it doesn't matter. But the moment I put a phone that has a lot of processing power — I've been doing that today, I know from one chip to the other there's more processing power — if I put it in airplane mode, it becomes useless. We put it in our pocket, we don't use it. And I think that's going to be like that with AI. There's going to be things that are going to happen on the phone because they have to happen on the phone. The things going to happen on the cloud. And I'll finish by saying this comment I made about friction. That's important. If you pull up your phone and it takes time for the agent to do certain things for you, then you're going to do it yourself. But certain things you're going to just get the agent to do it. And the last part is I talked about in my keynote: the phone is very unique to you. So you may have this little app in the corner that does something, maybe open your garage door or check something, and that app is on your phone, and you need the agent to go over there in the phone and operate the app. So you're going to have a combination. You need to think about those things as one integrated computing system, the way the smartphone is today.
So you don't ascribe to a potential world where Muse and these personal agents had abstracted away the computing off the device, and most workloads or token usage is happening in these cloud VMs?
No, and there's two reasons. Okay. I'm going to first give you a reason, which is there's the coexistence of the old and the new by definition, and the coexistence of the old and the new requires those systems to be, you know, leveraging those both worlds. Remember — you probably... I'm going to show my age, but I remember when everybody talked about the webOS of Palm and all of the applications were going to be cloud-based, and the phone was just going to be a shell. It didn't happen that way, right? Actually, you saw app stores developing. Okay, you see now, especially with open-weight models, you see now the growth of on-prem. It's very interesting if you look at just — if I got this right — if you look at the Nvidia last earnings, when they talked about, I think it was in the forward guide, I don't remember, I may get the numbers wrong, but they talked about 40% of the revenue was AI companies and hyperscalers, the other 40% on-prem, neocloud, sovereign AI. The moment you believe in on-prem, you accept hybrid AI. Certain things you're going to do on-prem, certain things in the cloud. Then the question is, what is the prem? Where does it end? Does it end in your PC? Does it end in your phone? And I think that's the way this is going to go.
Okay, I see that. I still am struggling to understand what an AI smartphone is. You said that a couple times. You said it this morning. I mean, you could say these phones are already AI phones. Siri is kind of good now. Google Gemini is very good. You know, you can have all the chatbots on there. Like, what is an AI smartphone? What is the difference?
So, the AI smartphone vision became very clear when OpenClaw was launched, and the ability for you to tell an agent, or the ability for you to text to an agent and say, 'Here's what I want you to do for me.' And that's why I kind of provided the example. You're just going to say — you get a question, you're going to say 'Yes' — and then the phone, the agent is going to do things for you. It's going to run on that computer. It's going to deal with the planning. It's going to deal with the smart routing about things they're going to be doing locally. It's going to go to your memory, personal memory graph, and your context. Your context only exists in the phone, right? Context does not exist in the cloud. It exists in the phone. And the sensor data comes from the phone. So it's going to look into those things. It's going to make it relevant to you, and it's going to take action for you. That is very different than the things that you can open your browser and do it yourself. So it's way beyond a browser. I think that's what the AI smartphone is. Now, it's important to explain: we're just at the beginning of this, early days of AI, and I love making parallels with the smartphone transition. I'll go there. But early days of AI, you have some of the phone companies talking about this is the AI. So you have the set of features from the smartphone maker, set of features of the OS. So maybe those are the ones that you're going to use, but that's not an AI smartphone. Later you saw with orchestrators and agents — this is a 3P issue. It's not a first-party issue. And there's going to be different agents, different skills, different things. They're going to be integrated across different companies. We use the example of some services like payments. The application becomes horizontal and it becomes part of everything. The boundaries of OS and apps no longer matter, and those experiences happen across multiple devices. That is the fundamental difference.
Early days of the phone. I remember that when Android launched and iOS, the iPhone launched. So you have an incumbent — at the time it was like Nokia — and you also have like BlackBerry, and they look at the features that came with the smartphone. Let's pick the iPhone as an example. You have iTunes. You have the little music app that you can move the drums and the guitar and things around and play music. You have those, I don't know, 15 apps that come with the OS. Not a lot of things in the app store. So some of the existing companies, BlackBerry and Nokia, said, 'Okay, we can do those features too.' But that wasn't the smartphone. The smartphone was defined by the third-party applications. Nobody knew at that time what was going to be invented and what people were going to do. I feel the same way about the AI smartphone. Once you put the harness in, you put the capability, you will see how those things are going to develop, and I think that's going to change the phone experience.
One of the biggest pains I've had running Sources has been managing my books. I'm not an accountant, and all of the software out there is unintuitive, clunky, and detached from where my transactions actually live — until now. Mercury recently launched Mercury Books. It's AI-powered accounting that works with your Mercury banking and credit card transactions, plus external cards and payroll systems as well. You can choose between cash and accrual accounting and quickly generate reports for your cash flow, P&L, and balance sheet. Like everything Mercury does, the design of Mercury Books is super approachable and clean. I finally don't feel overwhelmed when I'm trying to understand the big picture of my business. I love that AI does most of the heavy lifting, including for things like auto-categorizing transactions, and that Mercury's Command AI agent can handle tasks end to end. You don't need to be paying for separate accounting software anymore. And you can even invite your accountant or CPA to work alongside your Mercury Books with you. This doesn't cost extra, and Mercury will even help you find a human accountant if you need one. Visit mercury.com/books to learn more and apply online in minutes. Mercury is a fintech company, not an FDIC-insured bank. Banking services provided through Choice Financial Group and Column N.A., members FDIC.
AI is only as useful as the context it has. But when that context is scattered across tools, threads, and DMs, your team and your AI agents are flying blind. That's the problem Jira by Atlassian solves. What's the goal tied to your project? What got decided last week in Slack DMs? Atlassian's Teamwork Graph pulls all of the valuable pieces together from Jira, Confluence, GitHub, Slack, and more, so nothing falls through the cracks. You get 44% more accurate results with 48% less token usage. With Jira, you can easily share your work context with the AI agents you already love, like Claude, Cursor, and GitHub Copilot. Assign them work directly or connect your tools through MCP. All of this lets you spend less time digging through endless links and messages, chasing down what got decided and by who, and spend more time actually shipping. Learn more at jira.com. That's jira.com.
I spend a lot of time context switching between meetings, often with no time to process one before the next starts. Thankfully, Granola runs in the background the whole time. It's an easy-to-use AI notepad for meetings that works everywhere, even on phone calls. I use Granola to recall what was said in meetings and create helpful summaries. I use it every day to stay on top of what I need to get done with my team. It connects to my email and suggests follow-ups for me to quickly review and send, saving me valuable time. Granola isn't just a core part of my workflow — it's basically my second brain. Try Granola at granola.ai/sources and use the promo code sources for 3 months off.
Framer is the AI website builder that powers the Sources podcast website at podcast.sources.net. Framer brings AI agents into the same canvas where your website is designed, managed, and published, so you can move faster without giving up your taste or control. I use Framer to make the Sources podcast website the destination for everywhere you can find the show. Plus, you can also see recent issues of my newsletter. Start building with agents for free today at framer.com/sources for 30% off a Framer Pro annual plan. framer.com/sources. Rules and restrictions may apply.
Turning to wearables, and I'm old enough to remember when we talked about wearables, we were just talking about VR and then AR a little bit, but now it means a lot of things. We're talking about watches, pins, whatever you're working on with OpenAI — you can't talk about all these things, right? All these devices that are coming. I'm curious to hear the state — before we go into the AI wearables — the state of VR and AR, because Qualcomm's been a player in that industry for a long time. And it feels like that industry has struggled to gain real commercial traction, though Meta is starting to have it with the glasses, and Google's coming with their glasses with Warby Parker, etc. I think Apple will come, everyone's coming now. What do you make of the wearable market as it exists today, and is VR something that you and Qualcomm still believe in, or are we kind of past that?
No. This is a good question, and because there's a lot of things built into that, I will... I'll start with VR. It's like a console-type market. It's like a gaming market. So, you should think about that.
That wasn't always the case though. People thought it might be like mobile, right?
I know. People talk about this kind of the future, the metaverse, but — and I will give you very specifics on those things. But if you look, the technology developed to the point that you actually have a pretty good experience, but it's more like a gaming console experience. A lot of the applications was about gaming, not even social. It's about gaming because it's not a device that is convenient for you to walk around, you know, every day. You're just going to put this thing on, you're going to get fully immersed. Then you got to the middle of the road, which is mixed reality, right? And which you have the combination of augmented reality and see-through. But then you get to smart glasses. And the thing was not about the visualization technology. Think of that as a platform. The inflection point came because of GenAI. Let's just pick of those three categories. The one that I believe is the one that is going to be very successful — it's already successful today. I think today the market is in the 50 to 60 million units shipped. I think we have about 40 different designs across different companies. But I think that's going to grow dramatically. And let me explain to you what happened and why. First, I remember the first smart glasses. The intent was to extend the camera functionality: take a photo or record a movie, post it on Instagram as a story. But then GenAI happened, and then you started to do different things: 'What is this? Translate this. Take me there,' etc. 'Who's this person?' And you started to see exactly this — the agentic or new AI workloads getting traction. This became a more useful device. And what is interesting about that category, if you believe what we've been saying for a couple of years — AI is the new UI — then this is prime real estate close to your eyes, to your ears, in your hand. Very natural. Humans know how to wear glasses. There are some rules, like — that's the problems we need to solve. The rules is you need to have about 35 grams or less. I'm giving you the ideal. You need to have 6, 8, 10 hours battery life, and it needs to be a fashion device that is going to get traction. I actually believe in all of this category of wearables. All of it. The one that I am the biggest bullish on, and that will be the biggest category, is glasses — because of see what I see, read what I read, and other things. And some interesting use cases, like for example our research team did something super simple: it's a small model in the glass, you load up your company directory, and then you just walk around and say, 'Oh, who's this person?' 'Oh, that's Alex, here is the employee number.' So there's so many different applications, and I think that's the one that develops. Just to finish the answer: watches will play a role. Pins, pendants, jewelry will play a role. So, we have a number of different designs with companies. We see everything now. People are experimenting with earbuds with cameras, jewelry — people can wear jewelry, pendants, badges, the thing we announced with Microsoft. So, we'll see. It's early, but it's promising. Of all of those, I think the biggest one is smart glasses.
