A Conversation with MediaTek on the Latest AI Trends for Consumer Devices

How are the latest AI trends shaping consumer devices and enabling advanced AI capabilities across the technology landscape?

‍

Hosts Patrick Moorhead, Founder, CEO, and Chief Analyst at Moor Insights & Strategy, and Daniel Newman, CEO and Chief Analyst at Futurum, are joined by MediaTek's Adam King, Vice President and General Manager, Client Computing Business Unit, for a conversation on the latest AI trends for consumer devices. Find out where MediaTek is driving AI innovation in consumer devices, making advanced AI accessible at scale, and the vision to bring AI everywhere.

‍

Key Takeaways Include:

‍

🔹Silicon for the AI-First World: A look at MediaTek's vision to enable AI from the edge to the cloud, providing the powerful and power-efficient silicon required to meet the increasing demand for generative and edge AI applications.

‍

🔹Generative AI in Everyday Devices: A deep dive into the notable advancements of generative AI in products consumers use every day, including compelling examples in Chromebooks, tablets, and a growing number of smart home devices from smart speakers to TVs.

‍

🔹The Future of AI Applications: An exclusive peek into the development of emerging AI applications, including custom silicon for AR glasses and strategic collaborations on datacenter-focused solutions like the NVIDIA DGX Spark.

‍

🔹Building an Accessible AI Ecosystem: An exploration of how MediaTek is ensuring AI is accessible across various devices, market segments, and geographies, fulfilling its vision to bring AI everywhere.

‍

Learn more at MediaTek

‍

Watch the full video at sixfivemedia.com, and be sure to subscribe to our YouTube channel so you never miss an episode.

‍

Or listen to the audio here:

‍

Disclaimer: Six Five Media is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

‍

Patrick Moorhead: Welcome back to the Six Five podcast. Daniel and I are back and guess what? We are talking about AI. It is our favorite topic. Daniel, do we talk about anything else?

‍

Daniel Newman: It feels like more and more this is what we spend our time on, Pat. But it is the largest and most important trend that will probably take place in our lifetimes, the rest of my very long and yours. And it is one of those things that I think right now people want to hear about. So in my opinion, Pat, what else would we want to talk about?

‍

Patrick Moorhead: Exactly, and most of the conversation in totality has been about AI in the data center. But you and I both know, and history shows that over time any type of major compute phenomena goes farther and farther out on the edge for many reasons. We've seen it in many inflection points in smartphones and NPCs. There have been some fits and starts. Right. But that's just, it's natural in a market. Everything's not going to work and be an overnight sensation. We do have to build out the infrastructure and get developers ready and get the software ready to make that work. We also need to have the right chips in there. And I'd love to introduce our guest for MediaTek, Adam King, to talk about consumer AI. Adam, welcome to the Six Five.

‍

Adam King: Thanks Pat. Thanks, Daniel. It's a pleasure to be here.

‍

Patrick Moorhead: So welcome again to the show. Let's start at a high level here. I'd like to talk about your vision for AI at MediaTek and how you are positioning yourself as an AI enable enabler? Maybe beyond the obvious.

‍

Adam King: Sure. And first of all, I agree with what Daniel said earlier about AI being the biggest technology wave to hit us ever. And it's a great time to be at MediaTek because we think we're pretty well positioned to help the world along this AI journey. We actually have pretty deep roots in AI, but in the more traditional AI. We're now at the point where you can distinguish between old or traditional AI and the new generative AI, but certainly on consumer devices and on the client, which has historically been our focus and will continue to be going forward. Traditional AI, meaning things like image enhancement, voice recognition, object detection, those things we've been doing for a long time. We're actually on the seventh generation of our NPU shipping and products today. But, clearly what's transformational is the move to generative AI with LLMs and transformer models. And the question there is how much of that is going to be run on the edge versus in the cloud, by the way. We're also increasing our presence in cloud infrastructure, as I think you guys know. But let me talk about the edge side for a while, because even though we're an edge company, the reality is that we understand that generative AI and agentic AI is going to primarily live in the cloud. My 50 tops NPU on my MediaTek phone is great, but it can't compete with, you know, billions of dollars of GPUs and HBM in a data center. You're always going to get much bigger models, more performance, better quality from the cloud. But there are cases where it makes sense to run at least a portion of your workload on the edge, I think for all the reasons that we know, for reduced latency, for privacy, for reducing cloud costs. But it really depends quite a bit on what usage you're talking about and what device you're talking about.

‍

Daniel Newman: So clearly you're enabling AI. And Pat and I have been sort of at the, we'll call it the leading edge of the edge. We've been talking close to two years since we kind of rolled out that, you know, on device and the future is going to be this kind of mobile AI experience. Smaller models, more efficient. You know, Adam, I would say that some of it's happened, some of it's going to happen, maybe, some of it maybe won't happen. I mean, what's your take on how the market is evolving? I mean, we're seeing models hit every day. Some are bigger, some are smaller. You're seeing energy constraints, network constraints. I mean, you have to be really tracking this market evolution, both for how you continue this consumer devices side and of course, as you mentioned, the expansion of the MediaTek business in the cloud.

‍

Adam King: Sure, yeah. I mean, it's been really gratifying to see the really rapid progress on small language models and the efficiency and the quality just keeps going up every month. So it makes it more and more feasible to run good generative AI at the edge. But it really does depend on the use case as to where you want to have that workload. Let me give you maybe an extreme example of where it makes sense to run something at the edge. Let's talk about the smart home. You've probably got six or so security cameras at your house, most outside, maybe some inside. And when you think about those cameras running 4K video 24/7 and if you're a service provider to those cameras and you've got 10 million customers all uploading that amount of video it's just the cloud costs become enormous. It makes a lot of sense to have some amount of local AI to do the inference, the filtering, the object detection, to figure out when your dog escaped out the side door and to proactively alert you of that. Not to mention the privacy concerns. People are obviously going to be very wary about putting cameras, especially inside their homes, and they'll only do it if they know that stuff is not leaving the premises. So that's a great example of where you really need Edge AI. Other things, not so much. Like if I'm using my Copilot PC or my Gemini Chromebook and I want to summarize a 27 page legal document, do I care if it takes an extra three seconds to go back up to the cloud and back? And no, I probably don't. So the. There is a latency, but it doesn't really matter. I think what's really exciting is that Edge AI is going to improve, I think every existing device, but it's actually going to enable new categories that really just couldn't be successful without Edge AI. And that's where you get into this space of what Jensen Huang calls physical AI. Right, this progression of traditional AI, generative AI, agentic AI, up to physical AI, which is devices that live and interact in the real world beyond your desktop. Things like humanoid robots and autonomous vehicles and mobile AR glasses, where you're constantly having to sense and interpret what's going on around you. And you need it in real time if your AI glasses are going to warn you about the dog on the sidewalk a few steps ahead of you. A five second latency is a real problem. In that case, you want it to be immediate.

‍

Patrick Moorhead: Yeah, I could have used that low latency in one of my Roomba vacuums about five years ago. But I really liked your example. You didn't use the word Amazon, but I will. I have probably 50 Amazon based devices that at least based on certain teardowns I've seen, you're definitely inside, so you're definitely all over my home. I do have a question though, is maybe we can talk about even more consumer devices that you're in. So let's talk about maybe some of the devices you're in and specifically what you're doing in the devices, because they do vary.

‍

Adam King: Yeah, sure. We're in an incredible range of devices. I like to tell people that they probably have at least five MediaTek based products in their house that they use every day and may not be aware of the silicon provider, but we're in as unified smart home devices from certainly TVs, streaming devices, smart displays. We have a very large Wi-Fi or connectivity business. So there's a good chance we're in your home router. There's a chance where the Wi-Fi in your PC and we're in a lot of those compute devices as well. Android tablets. We are actually the number one silicon provider for Android tablets. We're the number one ARM provider, ARM Chromebook supplier. And we've had some really cool products come out recently. We're in phones in a very significant way, especially in China, but increasingly going global.

‍

Daniel Newman: Yeah, we've definitely seen the proliferation of MediaTek hear about it coming up a lot more, entering new spaces, creating as the premium mid tier, sort of the debate continues to grow. We've seen MediaTek make head inroads in both and it's been impressive to watch. Another place, you know, you sort of alluded to was the inroads in the smart in the home. And I've heard you kind of mentioned it a little bit. In the last answer, Pat talked about his vacuum cleaner running over his feet. But talk a little bit about, you know, kind of how the company, how's MediaTek and how, you know, more broadly, how is AI, you know, transforming the smart home. This to me, by the way, Adam, feels like something that was sort of promised for many years. You know, intelligent homes, touch panels that had everything coming out of ceilings and walls and TVs coming out of desks and tables and IoT AI, smart devices, applications and now AI and agents. Seems like it's the perfect chance to really democratize something that was really limited to the ultra uber homes with big dollars to, to make integrations. Are you seeing that trend now with AI?

‍

Adam King: Yeah, I really believe that AI is going to drive the next wave of growth and improve user experiences for the smart home. And you know, and let's give the smart home a little credit. Maybe we have to look at it at a bigger timescale, but compared to 10 years ago, your home is a lot smarter than it was. You've got, you know, you've got cameras, you've got devices, you've got the ability to control your lights, you've got, you know, the ability to wirelessly control or remotely control your thermostats. So, so the, the smart home has made tremendous progress, but there are still challenges with usability, interoperability. You know, one of the challenges is that people don't just upgrade their smart home overnight. You kind of do it piece by piece, right. You're like, oh well, you know, maybe I need a new fire TV stick or maybe I'll, I'll put in this wireless sensor in my bathroom or, or I'll install this, you know, new remote controlled curtain motor. But you have to do that, you know, you want that to seamlessly work with whatever system that you have. And so there are still challenges with setup, interoperability, too much time spent just, you know, setting up and managing these things versus making them seamless. And AI has the power to transform all that. There are a couple of other technologies that are helping with that as well that I know you guys are aware of with thread radios that allow mesh networks. So everything is connected seamlessly so that you can solve some of the interoperability challenges across these devices. You put all that together and I think that the smart home is going to get another boost here.

‍

Patrick Moorhead: Yeah. So we talked a lot about the current market, smartphones, tablets, smart homes, smart TVs, and that you have a wide array of solutions. And let's talk about the future here. What are some investments that you're making in future new target segments out there that you can talk about?

‍

Adam King: Sure, yeah. We're getting, I sort of break up the world into two things, which is how do we make existing product categories better with AI and then what entirely new product categories will there be in the world and that we can participate in? It's just to give the existing devices one, one more sort of plug. I do think they will essentially all be enhanced by AI in one way or another. The amount of cloud versus edge workload is going to vary depending on the device and the usage. But we are increasingly putting AI capabilities, we're putting NPUs, we're putting our software stacks into just about every product that we make. I wanted to say one more thing too on the smart home, Daniel, to finish answering your question, which is I do think that there's long been this idea of should people have a hub in their smart homes? Right, for local storage, for local control, and certainly the, the diyers do that. But the world largely went away from having a hub because everything just ran and connected in the cloud. You didn't really need a hub. But with AI there's maybe an interesting shift back to a hub model where it now makes sense to have some amount of local AI to be able to connect all of your devices. And it solves that upgrade problem because if I connect all my existing cameras and sensors and light bulbs to my smart home hub. Essentially, I've made them all smarter and I didn't have to go out and buy a new one of each of those devices. So I think we're seeing the big smart home providers experiment with bringing more capabilities back to this idea of a hub. And the target form factor seems to be the smart display. So it's a device with a screen that a user can interact with. But then Pat, to ask, to answer your question about future product categories. We’re certainly investing in vehicles, autonomous vehicles, and then just cockpits for non autonomous vehicles as well. So that's a big new bat that we kicked off a couple of years ago. We are investing more in cloud infrastructure. We are doing more and more premium compute devices. So you've seen what we did with Nvidia with the GGX Spark. We're making existing devices more premium as well. So we have like the Lenovo Chromebook 14's best Chromebook on the market, 50 tops of on device AI. We are now in Samsung's flagship Android tablets. You know, these are the most premium Android tablets on the market. You're seeing us in the Tab 11 Ultra and the Tab 10 before it. So we're both, you know, we're delivering higher levels of performance for existing categories as well as these, these new markets. We're also increasingly playing in the wearable space. We announced actually a couple years ago that we were partnering with Meta for their future augmented reality solutions. And I can't say much more about it other than that's going along nicely and we're studying the space to see if there are other opportunities. We're really starting to see AI glasses take off in a significant way. I know Anshul did a piece a year or so ago on the Ray Ban Metas. Really great review. There's more of that coming. I'm looking forward to his updates.

‍

Daniel Newman: Oh yeah, it's only. There's a whole lot more to come, Adam, and clearly as you kind of go down, I feel like this was the and conversation because it's like MediaTek is doing this and we're doing this and we're doing this and it seems that the company is really doing not only a very good job, but a very important job right now, which is diversifying itself to be able to handle a pretty significant pivot you've got sensing in terms of the homes you've got handsets and devices. But we know that having a small system is going to be important as new potential form factors. Enter whether that's some of the stuff we've seen with glasses and foldables. But also, of course, we have a trend with things like these pins and earpieces and things that we don't quite know yet. But it's good that you're there because it's going to be important that you stay flexible. And then, of course, you talked about TVs, devices in the homes and of course, in the cloud. So really good to get a background from you. And I appreciate you taking so much time with us to, to kind of take us through everything, give us the rundown. And I wanted you to say, you know, as, as, as an analyst looking at the market, you opened my eyes to a few new things, which is really important and I appreciate that. So thanks so much for taking some time to join us here today.

‍

Adam King: Well, thank you, Daniel, it's an honor to be here. I appreciate you giving me the time.

‍

Daniel Newman: Thank you everybody, for being part of this episode of The Six Five. This is a Six Five Webcast and we appreciated the chance to talk to MediaTek’s Adam King. Great conversation. Hit subscribe. Join us for all of our other content. Check out the show notes to learn more about MediaTek and what we discussed here on this Podcast. But for Patrick Moorhead and myself, it's time to say goodbye. We will see you all later.

‍