On this episode of The Six Five – CXO, hosts Daniel Newman and Patrick Moorhead sit down with Micron’s Chris Moore, Vice President of Marketing, Mobile Business Unit, and Ziad Asghar, Vice President of Product Management for Qualcomm.
Their discussion covers:
- The history of their partnership and collaboration
- Why SoC, memory, and storage play a pivotal role in the future of next gen mobile experiences
- The recent launch of Qualcomm’s flagship chipset
- How they are enabling innovations for the future of mobile, including smartphone form factors
- What’s ahead for the dynamic duo
Be sure to subscribe to The Six Five Webcast, so you never miss an episode.
You can watch the full video here:
You can listen to the conversation here:
Disclaimer: The Six Five CXO 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: Hi, this is Pat Moorhead and The Six Five is live in the Micron booth at Mobile World Congress in Barcelona, Spain. I hope you can feel all the excitement around. Maybe not hear too much of the excitement, so you can listen to us. But Dan, is this a big year or what?
Daniel Newman: It is a big year. And that’s been thematic in the conversations that we’ve had, Pat. In the briefings, in the advisory sessions, and also on all the videos that The Six Five has done here.
The executive interactions, customer interactions. Everybody seems to feel pretty optimistic. The energy here at the show. I think people are doing business, which is why we’re all here.
Patrick Moorhead: Yes. As we’ve always said, Daniel, on our shows, there’s a triad of things that have to come together to make devices work. And that’s compute, connectivity, and memory and storage. Memory and storage doesn’t get enough of the conversation.
But quite frankly, if you don’t have those three moving at the same time, the technology does not move forward. And it’s just the coincidence that we have here, folks from Micron and Qualcomm. Ziad. Chris. How are you doing?
Ziad Asghar: Very good.
Chris Moore: Very well.
Patrick Moorhead: Thanks for coming on the show.
Ziad Asghar: Thank you.
Patrick Moorhead: Ziad, you’re a veteran of The Six Five. Chris, I think this is your first time. It’s great to have you on the show.
Chris Moore: Thanks for having me.
Ziad Asghar: Thank you for having us.
Daniel Newman: Pat, you set it up pretty nicely. You’re absolutely right. If you take the memory and storage out of the equation, then all of that compute doesn’t do a whole lot, does it?
Patrick Moorhead: Nope.
Daniel Newman: Having Micron and Qualcomm here, I think maybe a great place to start is just the overview of the partnership. Chris, why don’t you jump in first? I’d love to hear a little bit from you guys about what is the partnership that’s going on between Micron and Qualcomm.
Chris Moore: Sure. I think you said it great in the beginning. In order to have any compute work, it has to have memory and it has to have storage. That’s one of the fundamental building blocks of any compute system. We work very closely together on the forefront of new technologies, new interfaces.
Really happy with your launch of the Snapdragon 8 Gen 2. It was coupled with our launch of DDR5X, which is the fastest LPDDR in the market today. We did that last November.
Ziad Asghar: That’s right.
Chris Moore: That was very exciting. We work very closely together integrating our products to make sure that our mutual customers can take them and run them seamlessly, so that their customers, you, get to use the devices.
Patrick Moorhead: Appreciate that. Ziad, any comments on the partnership to start?
Ziad Asghar: I think Micron has been a great partner for us as we build this amazing platform, but there’s so much that goes around it. All the critical pieces, of course, serves memory. We have had a longstanding partnership with Micron. Specifically, in the Snapdragon 8 Gen 2.
We were able to bring in the new LPDDR5X. Much more capable. Able to really enable all the amazing things we are packing into our products. But it requires that memory and storage solution and partners like Micron allow us to make it happen.
Patrick Moorhead: So let’s do the double-click on the value of memory to the SoC, to the user experience. I know we’re going to talk about the future in a little bit. We’ll talk about the future in the future, but how about the here and now? You talked about LPDDR5X. What’s the benefit? What’s the improvement to the end user experience?
Ziad Asghar: Smartphone is this amazing platform. Every user still wants more. We do surveys regularly, and they want the device to have better a camera.
Patrick Moorhead: By the way, thank goodness.
Daniel Newman: Right?
Ziad Asghar: Thank goodness for that, indeed. Indeed. They want better cameras. They want to do more gaming. They want new use cases. They want to have a virtual assistant. What does that mean inside the device? It means now at the same time, you might have a camera running along with the AI processing that’s happening on that camera. All of that requires you to have basically large neural models that need to sit in the memory, be able to board that to the SoC.
And then, the other thing that’s really happening. What people don’t realize is in the past many times these use cases were such that you ran one or the other use case. Now, you run them concurrently. You have the camera running, the AI running. You might be doing a telephony call with that. You have your Microsoft Teams session going on the side. At the same time, you’re sending an email.
That means actually the emphasis and the stress on the SoC, but also on the surrounding pieces including memory and storage is pretty massive. And that’s where we really need that additional bandwidth. Partners like Micron, of course, come into play over there. We work very closely with them. With 8 Gen 2, we were able to actually show the first LP 5X working with Micron.
Patrick Moorhead: Love it.
Daniel Newman: The whole process though, Ziad, of making the selection of a partner like Micron. You have choices. Just like every smartphone device maker has a choice whether or not to use the Snapdragon 8 Gen 2.
Ziad Asghar: Right.
Daniel Newman: Talk a little bit about the whole thinking behind that selection process. What sort of made you say, “Hey. This is our flagship device. We know the interdependence that storage and memory have on making this thing go. Micron is the right partner for us.”
Ziad Asghar: We of course have a platform. Like I said, we have multiple customers. And of course, we were talking more from a Snapdragon as in the smartphone-like application perspective. But we have a lot of emphasis today on automotive, on PC, on XR, on auto as well. We of course enable multiple partners across all of those different businesses.
Of course, we are choosing the people who are able to show our products in the best light. With the usage of memory, you want of course a partner where you’re able to get best-in-class power consumption, best-in-class experience. We actually set up labs, where we have all our partners come in and essentially take our device through the paces, such that we are able to deliver something to a customer where they can launch extremely quickly to the market.
Time to market is absolutely critical. So this early engagement that we have with partners like Micron allow us to be able to ship these parts quickly and to be able to ship multiple partners at the same time. Of course, each partner has different reasons to go with a particular memory vendor or not.
Patrick Moorhead: Chris, it always amazes me how when we have a major launch and the device containing SoC, storage, and memory all shows up on the first day and it magically works. Now, I’m an insider or was an insider in the industry and see how it works. But I’m curious though.
How does Micron get ahead of the curve in defining the right features? It’s easy to say, “Optimize PPA.” That’s what we say in the entire industry, but it’s a lot harder than that. Especially, when you’re leading like Micron is. What’s your process? How do you know how to get ahead of the curve? I’m sure you’re working with companies like Qualcomm, but you also have an independent process as well.
Chris Moore: Absolutely. Absolutely. Leadership is a culture and Micron has been a leader in memory and storage for 40 years. Over 40 years now. And that culture drives us, for example … As you were just speaking, Ziad, about low power. We were first to market on the 1-Beta technology, and we’re working closely with Qualcomm on doing that. By shrinking our nodes faster than the rest of the industry, we can give the end users that low power that you’re looking for, which of course, translates to longer battery life.
Patrick Moorhead: Well, the crazy part is there’s low power and low battery life, but these new smartphones have an inordinate amount of performance can crank out too.
Chris Moore: That’s right. That’s right.
Patrick Moorhead: It is amazing when you’re developing, again, that special triad that I talked about in the beginning, what developers can do with that. In a way, the two of you are setting this canvas for developers, where you have to make these decisions three to five years in advance and have a thesis of what they can do. Without even actually knowing what will come out on the other end.
Chris Moore: That’s exactly right. I’m an engineer by trade. When I think about what I do every day on the hardware side, it’s really about getting out of the way. We’ve developed these platforms together to allow the application to be there. For example, we talked about LPDDR5X, which just launched.
One thing we all know these days is your camera takes these night pictures. As Ziad was talking about, to take a night picture, you have the image processor going, you have AI going. All of these different parts of the SoC are turned on at the same time. Well, guess what? Your bottleneck is actually the LPDRAM. When we came out with that LPDRAM 5X spec, you had no idea it was going to be used for image processing at night.
But you can actually see that. How many times have you sat there and you’re trying to take that picture? It takes three, four seconds. Well, you move to the LPDDR5X with our newest products together, it gets done in a couple of seconds. Things like that. Again, the culture of innovation is driving the platforms for the next generations, like you were saying, before we even know it’s there.
Daniel Newman: With AI, for instance, you mentioned that. Ziad, I know you spend a lot of time focused on that.
Ziad Asghar: Right.
Daniel Newman: That’s sort of a use case of the future. We of course have metaverse use cases. We have so many things. How do you collaborate to make sure that you’re able to be on the front edge to be able to support those?
Because like I said, you kind of alluded to it. You don’t always know, but we know. At this point, we know that AI is going to change everything. Stable Diffusion. We were playing with this at your booth. Seeing on-device, at the edge, generative AI. I was drawing … What was it? Unicorns in a wheat field.
Chris Moore: I wonder why?
Daniel Newman: You were doing a leprechaun. What was it? A leprechaun?
Ziad Asghar: A leprechaun in a field of roses.
Patrick Moorhead: A leprechaun in a field of roses.
Daniel Newman: In a field of roses.
Patrick Moorhead: I was trying to stump this billion parameter model, but sure enough, it cranks it out.
Ziad Asghar: That’s right.
Daniel Newman: It did a pretty good job. It looked remarkably like me. Overall, how do you get ahead of this stuff? Because I feel like in some ways we’re just trying to keep up.
Ziad Asghar: I think what we of course do is there’s a lot of modeling and everything that goes behind it. We know what our camera spec is going to look like. We are know what our artificial intelligence capability is. But what we don’t know … Yes. AI is going to be big. What we don’t normally know is, “How big of a model do I need to make it work with?”
Patrick Moorhead: Exactly.
Ziad Asghar: Usually, the question is, “How much bandwidth do we need?” As much as we can have many times. But of course, we need to make sure that we have very good reasons on, “What are the concurrent use cases that I am planning to do inside the SoC?” Again, you mentioned DDR as being one of the critical parts in the chain. There are multiple parts in the chain.
Daniel Newman: Right.
Ziad Asghar: Many a time, your SoC, and you know this well, is going to be the weakest link that you have in that chain. We of course need to make sure that the accelerators, the memory, the internal plumbing inside the chip is capable of running at that level. And that’s why, in this particular case, we think the large language models, for example, or billion plus parameter models are here to stay.
We’ll actually have more of these use cases and you’ll see us talk about it a lot more in the coming days. But it requires us to be able to store these models into the DDR, be able to move it into the SoC and back. And that requires the fastest and most capable memory bandwidth and memory solutions possible.
Patrick Moorhead: Chris, you must be loving this. Sure. There’s a billion parameters model on the device today, but we interviewed Qualcomm CEO, Cristiano Amon, where he was speculating about 10X or 20X the size of these models. That’s got to get you thinking as well.
Chris Moore: It’s a lot of fun. And it gets me thinking as an end user too, not just as a developer. Because where we were talking actually just before we came on camera, what is clear that is happening in the next years is all the compute power that’s happening in the cloud is coming to your hand.
And so, people are not going to want to wait for your data to go to the cloud to translate and come back to be able to read the menu. Today, you can do that in your hand. Pretty soon, you’re going to be able to translate your voice in real time. All of these things that you’re driving with your great processors do require faster memory. Bringing it back around to our partnership.
We actually work at pre-silicone levels. Our engineers are working with Qualcomm engineers to validate the processors, to validate the interconnectivities. Making sure that these bandwidths are achievable before we even tape out. You’re talking three, five years in advance. We’re talking about what’s coming next and making sure that all of that is capable.
Again, for me, as an end user and watching my children grow and thinking about this digital companion that they’re holding. They’re going to be learning from this thing through their life. It’s going to be learning from them through its life and how the world is really going to expand. Digital and human interaction is becoming more and more and more powerful.
Patrick Moorhead: We talked a little bit about cutting edge use cases. Software, in this example. Stable Diffusion, generative AI, and then the future. But also, we’re seeing kind of a sea shift in form factors.
Ziad Asghar: Absolutely.
Patrick Moorhead: And if there are anything that challenges PPA, it’s a form factor. How do I cool this thing? Is it going to have a much smaller battery? How do you get ahead? How are the two of you addressing unique smartphone form factors?
Ziad Asghar: I can start?
Chris Moore: Sure.
Ziad Asghar: Basically, the idea is we of course work together. You know that the volume inside the smartphone is pristine landscape.
Patrick Moorhead: Yes.
Ziad Asghar: Each and every millimeter cube is essentially looked at and miniaturized as much as possible. We work very closely with partners like Micron. Especially, when they’re talking about POP-like solutions to be able to make it work inside a smartphone. But you can imagine as we go into augmented reality, for example, these challenges become much more complex.
Not only do we have to now make it fit in a very small form factor like glasses, but we’ve got to fit in the battery. And that requires even greater collaboration between us and partners like Micron to come up with more compact solutions. What it also means is that the billion parameter model we’re talking about … Today, it’s running on a smartphone. It still has a decent size battery.
Patrick Moorhead: Sure.
Ziad Asghar: Now, you can imagine, in the future, we’ll have it run on our augmented reality glasses. Because you want to augment your view, in the case of Dan … In your case, with unicorns, for example. My point is we’ll have that ability to be able to do all of that on a small form factor device, but it requires again much better collaboration with all of our partners.
Chris Moore: For Micron, we’re really excited. Everybody this year is talking about foldable phones. I remember the first one in 2018. I’ve been a big fan of them. You mentioned earlier that you want to be able to have all of those apps open at the same time. Pretty soon, you’re going to be able to cut and paste and view across the phone.
That’s going to drive a lot more capacity requirements. You need a lot more DRAM to be able to do that to act like your traditional PC. But then, going even further into form factors. One thing I’m really excited about is automotive. We don’t even think of that as a mobile form factor, but is there any better mobile form factor?
Patrick Moorhead: It’s definitely mobile.
Chris Moore: Driving 100K miles an hour on the outbound or whatever.
Patrick Moorhead: That’s right.
Chris Moore: Again, coming down to future usage models where your phone and your car are truly syncing. I’m sure I’m not the only one here who does a lot of meetings in his car.
Daniel Newman: Hope you do it safely.
Patrick Moorhead: Well, I always hope pull off to the side of the road. Always.
Daniel Newman: Safely. Yes.
Chris Moore: On the side of the road. Absolutely. And I’m definitely not letting my car drive itself while I’m doing that, but actually integrating the form with the car. The form factors are going to be changed. Going to be moving across. I’m going to be able to take my favorite networking and actually run it on the car in real time.
Daniel Newman: Yes.
Chris Moore: So I can see it here, while I’m on the side of the road. These things that we’re working on, the interoperability between all these devices. It’s really exciting what’s coming down the pipe.
Daniel Newman: I love it. The automotive space is really interesting. We’ve seen it get a lot better, but there still seems like such a big opportunity. What did I say? About 20% of the vehicle BOM by 2030 will be just semiconductors. I actually can see that getting higher when you actually add in all the possibilities that chip makers can provide.
We’re coming to the end here. I kind of want to do that, “What else,” question, but I’m going to get a little more specific than just, “What else?” I’m going to ask you guys this. Pat, you and I have been in a lot of meetings. Some CEO meetings this week. I remember Chuck Robbins at Cisco.
Patrick Moorhead: Right.
Daniel Newman: The sustainability topic kept coming up. Chip makers have a big role to play, because power is such a hard problem to solve without sacrificing performance. Is there anything through the partnerships that are created between companies like Qualcomm and Micron that could really lean on practical, measurable, lowering of power consumption while still giving those great experiences?
Ziad Asghar: We work a lot on this. Of course. It comes through also from what our consumers and customers are asking for. They actually want the device to be able to do more, like I said, but they want to be able to do it for longer. What that means is we are very focused on always reducing power consumption … Especially, from a Qualcomm perspective, we’ve taken our milliwatt and microwatt thought process, but we’ve taken that into automotive. We’ve taken that into PCs.
And that actually offers a unique advantage for us to really do a lot more on the sustainability side. Because now, those same devices are really skimping on power and using far lesser power than what they used to do in the past. I think that opportunity in general is how we …. Of course, when there’s a need from a customer, it’s even easier for us to be able to do that.
Actually, the customer requirement and what we need for sustainability are actually aligned. Of course, DDR being a component that’s very critical for working off all of these different businesses means that as we get better power consumption on memory and others, it becomes a total solution power reduction, which is very key for everything that we do.
Chris Moore: That’s absolutely important to us. In fact, for all the folks out there that aren’t memory geeks. In the mobile device, we have LPDDR, which is low-power DDR. That’s very specifically developed for mobile phones. But you’re now seeing that in the client space, as you know. These devices are getting thinner and lighter and you want longer battery life.
They’re using these low-power devices for sustainability, as well as just as you’re saying. For the user. We offer the full gambit of products, but power is going to become more and more important even in the data centers. You’re spending all this money to cool the data center. Well, if you keep following Moore’s Law and you use these low-power specifications, you’re going to spend less money there. So the total cost of ownership, the total cost for us as consumers is going to come down.
Patrick Moorhead: That’s a good way to wrap up the show. Chris and Ziad, thank you so much for coming on. You’re literally combined superpowers of the triad. Again, compute, connectivity, storage and memory. Wait a second. Is that that a quad? And I’m calling it a triad. What’s going on here? Did I gyp Micron here? I hope not.
Chris Moore: Don’t forget input, output, and how you tie that to the whole thing.
Patrick Moorhead: There we go. I know. I know. Well, I’ll hear about that later from somebody else. I’m sure. But anyways, thank you so much for coming on the show.
Ziad Asghar: Thank you for having us.
Chris Moore: Thank you. Appreciate it.
Daniel Newman: Absolutely. Appreciate it very much.
Ziad Asghar: Always nice to talk to you.
Daniel Newman: All right, everybody. Thank you so much for tuning in to this episode of The Six Five here at the Micron booth. We had Qualcomm. We had Micron. We were talking all about the way storage and memory and processing come together to drive great experiences.
But for this episode, Pat, we’ve got to say goodbye. We hope everybody hits that subscribe button, joins us for all of our episodes. We’ll see you later.