Prakash Sangam:
Hello, everyone; Welcome to another episode of Tantra’s Mantra where we go behind and beyond the tech news headlines. I am your host, Prakash Sangam, founder and principal at Tantra Analyst.
This is a special edition of the podcast today on which we are discussing the updates from Microsoft’s blockbuster Windows on ARM AIPC event. It’s quite a mouthful. It was held as a precursor to Microsoft’s developer conference. It’s called Microsoft Build, just starting from the 21st. The announcements and reveals from the event were mind-blowing and exceeded anybody’s expectations, really.
It wouldn’t be wrong to say that this will be starting a new era in PCs. Obviously, Qualcomm, X Elite and X Plus were the heroes of the event, and on-device AI was a major theme. Along with the event, today we will also discuss Qualcomm’s AIPC business strategy, positioning, competitive landscape, app developer ecosystem, and more. And I’m sure we will bust a few myths along the way.
To do all that, we have a special guest with us, and that is Nitin Kumar, Qualcomm’s VP of Product Management for Snapdragon Compute Chipsets. Nitin, welcome to the show.
Nitin Kumar:
Hello, Prakash. Thank you. Thank you for having me over. Very excited.
Prakash Sangam:
Indeed, should be. Before we get to the questions, I want to remind our listeners to check out our previous episode in which we had the analysis of X Elite and X Plus platforms, AIPC market impact, and so on. Please do check that out.
So let’s get started. Nitin, congratulations. As I mentioned, it was a blockbuster event for you guys. There were a lot of rumors and leaks on AIPCs and X Elite and X Plus. I think it’s safe to say that what came out of the event exceeded all of those expectations. Around 20 devices from 7 PC OEMs, which is almost everybody in the PC world today across both enterprise and consumer segments.
That’s quite an achievement. Again, congratulations.
Nitin Kumar:
Thank you, Prakash. And as you said it, yes, we could not be more excited about this moment in the industry where we are defining, call it the rebirth of the PC, the rebirth of a PC reborn on how this will be an inflection point on what is expected out of your PC. And we are super excited to be a part of this. We have our Snapdragon X Elite and X Plus powering all these devices.
We have across wide OEM, across all the PC OEMs that you can think they showcase their platforms powered by Snapdragon X architecture. And as I said, this would be a turning point in how you look at your PC, what you expect out of the PC. This is the pivotal point in what defines a true enhanced PC experience. And this is, as we are calling it, like a PC reborn.
Prakash Sangam:
Indeed, indeed, I agree with that. So before we get to all the details, since your first time coming on to the show, why don’t you get started with a little bit of your background? Could you explain your roles and responsibilities at Qualcomm? And you’ve been around for a long time, right? Almost 20 years at the company?
Nitin Kumar:
Yes, Prakash, yeah. I’ve been at Qualcomm almost 19 years, coming up to my 20th very soon. So definitely very excited, very proud to be a Qualcomm employee. I’ve had a long career here, part of different teams, different roles. My prior role before this was Snapdragon Mobile Chipset Product Manager, where I was handling mobile chipsets and mobile roadmap. My recent role for about two and a half years, I’ve been managing Snapdragon Compute Chipset Roadmap Product Portfolio for the business.
And I’m super excited about this. This is a phenomenal opportunity for the company.We are a very technology driven company. And if you look at it, we see that the PC industry didn’t really see a lot of innovation happening over many years. We’ve had the bleeding edge technology, new use cases, new experiences on mobile that Qualcomm has delivered. And our goal is to drive new experiences on the PC industry. And I’m very glad to be a part of this and glad that I have this opportunity at Qualcomm to execute on this roadmap and in our vision.
Prakash Sangam:
Yeah, perfect timing for you to be at the perfect place right now, right? So yeah, going back to the event. I mean, the event actually answered a lot of questions. Us analysts were asking you guys and others for a long time, right? So that was pretty good. And then we were asking about traction, who, when, how, and so on. So a lot of those were answered.
And I’m sure with so many OEMs launching in one go, in a very short period of time, how are you managing that? I mean, there’s a lot of support, a lot of OEMs, a lot of models and so on. How are you planning to support all of that? I’m sure, you know, with your mobile background, where this is nothing new, but it’s new for the PC unit, right? So how are you guys planning to address all of this and make sure all the launches happen smoothly?
Nitin Kumar:
Absolutely, Prakash, and very rightfully so. Yes, there are multiple of these devices that were launched across a variety of our customer base. We’re proud of each and every single one of them. And Qualcomm as a company and pen culture thrives in a challenge like this, where we have a A-class engineering team, our customer support team, our design team, our solutions team, and we all work together to ensure that every single device is carrying the best of the experience when we launch.
And we have a very large customer support team that ensures that all of these devices meet up to the mark. We’re helping our customers in designing their platforms and helping them optimize their platforms and launching their platform. And as you also mentioned, this is no different of a challenge where we do it on a mobile platform and we’re doing it with our PC customers now.
We thrive on our mission on delivering the best user experiences.
When it comes to Snapdragon, our goal is to always deliver premium experiences and nothing drives more to our engineering team and business team where we are bringing that technology, that enhancement to the hands of every single of our customer with Snapdragon powering that device. So it works like a charm on our side. And kudos to our A-Class engineering team at Qualcomm.
Prakash Sangam:
Very well. So from the event, it was pretty apparent that, you know, how Gen AI and especially on-device AI is going to be important to both Microsoft and the OEMs. Of course, you guys have been talking about that for a long time. So what was the key decision point for the OEMs to select X Series platforms for their AI PCs?
Nitin Kumar:
Absolutely, Prakash. And I mean, if you look at AI as a technology in general, we have been actually integrating, and I was a part of the mobile team, as I mentioned, we’ve been integrating dedicated NPUs on our smartphone silicon for probably eight, seven, eight years with an AI dedicated hardware research going at least more than a decade at Qualcomm.
So we’ve been on this AI journey and AI as a fundamental technology, a pivotal technology to drive new experiences. We’ve been on this journey on our smartphone roadmap for many years. And even on our PC roadmap as well, if you look at our previous generation product that we announced at Snapdragon Summit in 2021, almost like two and a half, three years back, that product itself had 15 tops of dedicated NPU AI engine.
And that product, of course, was defined like at least two years before that. So that was somewhere in maybe 2019 when the definition of that product started to take shape, right? So we’ve been on this AI journey for long as we truly believe in this as a differentiator and a disruptive technology. On our current X Elite and X Plus platforms, the two platforms that we had announced earlier, and you saw the devices announced based on these platforms, both have a dedicated 45 tops NPU capability.
And that drives the leading edge NPU technology from an on-device AI capability perspective that basically acts as the fundamental engine that drives new experiences when it comes to Windows operating system and third-party applications on top to leverage the capability and enhance the experience for the end user.
And when we look at our customers, the customers understand our technology, our differentiation, the value that we are bringing on their platforms and the value that they’re able to take that in the hands of the consumer, which is what all we are striving for.
Prakash Sangam:
Yeah, and I mean, X Elite and X Plus have the highest NPU tops in the market right now, right? And although they are a little bit tiered, X Elite is higher and both are, I think, were in premiere, but X Elite is a little bit higher and they both having same NPU tops, I think that says something, right? So, you know, you need, for AI PCs and the namesake itself, you need lots of AI processing in these devices.
Nitin Kumar:
Absolutely, because you really said that, yes, X Elite and X Plus are two-tiered approach and they differ predominantly in our CPU core count on both the platforms, but as a common driver, as the fundamental portion of the technology, the AI engine in both the platforms has a dedicated 45 tops of MPU, which is leading edge in the industry.
Prakash Sangam:
Cool. So, thinking about using this AI power, right, and the devices. Microsoft talked about Copilot, you know, moving at least part of the workload for Copilot to the device. I think that makes sense and is a clear business case. What other applications and use cases do you see using these on-device AI capabilities? Let’s say from, say, next six months to one year in the short-term period and say up to three years. There are not many applications that utilize on-device AI right now, right? What do you see on the horizon, short and long term?
Nitin Kumar:
Prakash, first, just starting with yes, as Microsoft said that in terms of like Copilot Plus, and just to reiterate that PC is powered by Snapdragon, X Elite, and X Plus are the only devices that can deliver Microsoft Copilot Plus experiences. That’s all because of the advanced leading 45 tops, dedicated NPU capability that we have, and you learn that from Microsoft on that with the Copilot Plus, and we’re the only PC, only platforms that can power that experience.
Of course, on top of that as well, when you look at it from a third-party application perspective, there are several of these applications across multiple domains that take the benefit of the 45 tops dedicated NPU capability to showcase differentiation. Then there are many such applications available across different domains. Let me maybe just tell you a few of those. I think we’ve shown some of these demos earlier at the Snapdragon Tech Summit last year.
We had, in terms of creativity suite of applications, we had Blackmagic design with DaVinci Resolve, extremely popular application for video editing application. And one of the key algorithms like Magic Mask is completely ported on our NPU architecture and that significantly reduces the time from a design perspective, opens up new possibilities as one example from a creativity suite application. I’ll name a few more for our listeners.
I very encourage them to go check them on their own when they get their hands on the device. And like Luminar Neo is another one. We have DJ Pro, I think we showed that as well. In terms of generative AI, we showed GIMP and GIMP plus stable diffusion. And maybe let me elaborate on that. GIMP, a very popular photo editing tool. And we’ve always worked with GIMP and we have a plugin available as part of GIMP photo software, where you can generate a stable diffusion image through a prompt as part of the GIMP software application. That runs completely locally on device on our 45 tops NPU. And you can generate an image in under seven seconds, just about seven seconds. As a comparison point, we ran some tests on our own with the latest competitive platform from Intel. That takes, I think, more than 22 seconds. And it also doesn’t even run fully on NPU. It also leverages. They have to divide the workload across CPU, GPU, and even then they take three times more.
Use cases like this are just going to be all over. There’s also Visual Studio CodeGen that we can generate code on the fly. We showed another demo with Audacity as well in terms of music generation, where you can generate music on the device by using a prompt. And maybe let me touch on one more class of application, because the benefit is just going to be so much.
I can keep talking and talking in this, Prakash, but when you look at it on what a true value we are providing with this on-device AI, and you also put in the enterprise angle on that, I think if you look at one, there is a latency aspect of it. Running the model locally on-device with that 45 tops NPU will give you a faster response versus taking the data onto the Cloud, running the model over there and then getting the data back. Running it locally will reduce the latency, will be a much better, faster experience.
The second aspect, of course, from a total cost of ownership perspective or a TCO perspective, running the model on the Cloud is going to be more expensive versus just running it locally on a resource that is already available at your hand. So that’s a cost angle as well. And then third is that there’s a security angle as well. A lot of enterprises would not want their data or their employees’ data to leave the device and go to the Cloud.
They’ll rather have the data set on the device and have the model or run the inference locally on the device itself and still improve productivity for their user, add more uses, enhance the experience without compromising the security. And it will be very personalized as well because your personal data resides on the device, contextual information resides on the device, and you can leverage that for a better experience. As the saying goes, like what happens in Vegas stays in Vegas. I think in this case, what happens on the edge stays on the edge.
Prakash Sangam:
Good catchphrase that you could use, right, for advertising purposes. I agree with all that you said. Actually, I have written a couple of articles on it, on why you need hybrid. One aspect is that most of the applications they talked about, they are either ported or developed for on-device itself. I think there is this huge scope for many applications, many Gen AI applications specifically, for example, where it is not just on-device AI, but kind of dynamically moves the workload between cloud and the device, depending on the need, depending on if you’re taking the creation as an example, creativity.
The first version is created under the cloud, but later iterations happening on the device more efficiently, much quickly and so on. So do you see any progress towards that, wherein there is a framework that is defined? So any third-party app developer coming in, can basically develop an application that dynamically shifts the workload between the cloud and the device, right? Do you see any of that sort of happening soon?
Nitin Kumar:
Absolutely, Prakash, 100% aligned on how we look at it, and we look at the market and the developer ecosystem exactly in the same way as well. There is a lot of benefit of running the models locally on device, and we talked about many such benefits, but there could be a hybrid model where a portion of the model or some models run locally and some may still need cloud depending on the size. And as the capability of these platforms and developers come on board, there is more of a need for the developer ecosystem to port their models and run them locally on the platform because of all the advantages that we talked about.
And with that in mind, Qualcomm has a lot of tools set. And again, this goes to that we have been working on AI and on device AI for more than a decade on our mobile, a large number of tools that are available. We have publicly announced our Qualcomm AI Stack, which is essentially think of it as a set of tools targeted for developer community, for a developer ecosystem. We leverage those tools and ensure the ease of development and deployment of their models locally for on-device AI capability.
So we provide all that as part of our tools to make them easy in their journey to bring their models and make them available for on-device. As part of that tool set, we also provide them flexibility in regards to heterogeneous AI, where the developer will have the choice to run the model locally on the NPU or CPU or GPU, and it gives them all the flexibility and helps them with their journey in porting the algorithm to run that locally. We have that Qualcomm AI stack already available, and I know there’s tomorrow’s Microsoft build on that. Expect another key announcement in that regard to be announced tomorrow, so stay tuned.
Prakash Sangam:
Okay. And that will also kind of allow developers to move between cloud workload versus on-device and dividing between them and dynamically moving over the workloads? Of course, it will depend on developer workload and how they want to break the model. You know, it can manage, it can marshal the app developers. So, how do you see the different parties coming together to come up with a common way of doing things which will be easy to scale?
Nitin Kumar:
I think the way I look at it, Prakash, and the industry is looking at it that way as well, there’s a common vision across the players in terms of providing the best experience. I think that enables the industry, that enables on how we are looking at the market. We have a very strong partnership. We’re very proud of our partnership with Microsoft and all our customers. And together, when we discuss with them, we’re all in the market to drive, hey, how do I bring the better technology, better experience out there?
And as long as that mission is common, we are working towards the goal of enhancing the user value of what they would expect out of their PC. It works seamlessly with our partners and Microsoft as the OS provider. And we tune a combined platform solution to drive that.
Prakash Sangam:
Okay, very well. So, and I think the good thing for you guys and everybody in the PC industry is that this is starting off now, right? So app compatibility issues and other things will not be an issue because they’re starting off and then they can start natively with all the tools that you have. However, that is not the case with many of the existing application. I mean, I used Windows on ARM PCs for more than a year, I think two years ago, I used Surface Laptop, Samsung, Galaxy Book based on ARM, all were on 8CX, 5G PC that Lenovo had. I mean, all were great, but the biggest issue by far was app compatibility, right?
And when I talk to IT managers, CIOs, in their mind, that is still the case, especially when you’re looking at the enterprise market, if they are buying thousands of AI PCs for their employees, they have to be absolutely sure that all their applications will run on these new laptops that they are buying. So I’m sure there have been a lot of improvements with Windows 11 and a lot of things started working. And Microsoft has been saying that they’ll keep improving the 64-bit and 32-bit emulators and so on. But I think there is still a lingering issue in the minds of folks on app compatibility. What do you have to say about that?
Nitin Kumar:
Understand the concern that you’re highlighting. In a summary, we’ve made tremendous, tremendous progress in that regard. Let me add a lot more details around that. So one, when you look at, and I’ll come to enterprise in a second, and from a consumer perspective, if you look at majority of the applications, we have looked at what are the top several hundreds of the applications and we have worked with ISV partners along with Microsoft to work with them and ensure there is full compatibility of those applications when they’re running on Windows on Snapdragon architecture.
And on top of that, a large percentage of those applications are actually going to be running native. Let me give you a full, or maybe a few examples on that. When you look at from a PC user perspective, all the Windows Office suite of applications are all 100 percent native. So whether you’re using PowerPoint, Word, Excel, Outlook, Microsoft Edge browser, all of that entire suite is 100 percent native. It runs beautifully. I’ve been using my Snapdragon PC personally for several months internally. It runs flawlessly. It’s incredible on that.
When you look at outside of that, Google Chrome is one example. That was announced publicly, I think a couple of months back now, that Google Chrome is also now available native application on that when you look at it from a consumer perspective. And that’s a very popular browser when you look at it from that perspective. And I’ll give you a few more examples. Other browsers like Opera, Firefox, a few other popular apps, Adobe Photoshop, Lightroom, Zoom, Dropbox, Amazon Video, Spotify, Facebook, Instagram. I picked a few names across different verticals.
When you look at these, these are all native applications that are available. So as you can see, there is a significantly large number of native applications that are already available on Windows on Snapdragon. And beyond that, we’ve made tremendous effort working with ISVs and working with Microsoft to ensure full compatibility, even if a small percentage of apps that are very commonly used or may not be native today. But from a compatibility perspective, you will see a tremendous improvement when you test your new Snapdragon X Elite PC, which I’m hoping you’ll be getting it soon on that front.
And then similar approach, Prakash, to answer on the enterprise and IT managers, as you said, and you’re absolutely correct. There is legacy application, legacy software, but we are working very closely with IT managers, enterprises, our customers, our OEM customers to ensure that they’re end customers, particularly on the enterprise side. And we are partnering with ISVs to ensure full compatibility and native app support that are commonly used across enterprises. So I’m very positive and bullish on how much we have come forward along with the help of our partners. And we are on an aggressive path to make that journey a complete seamless journey when you switch over from your legacy, let’s say legacy architecture machine to Windows on Snapdragon.
Prakash Sangam:
You know, if you have to take a guess for enterprise applications, what percentage of the top used enterprise applications you think are either native or running with emulator right now? It’s more than 70, 80, 90, 95, 99. If you have to take a guess, which one do you think will be the closest?
Nitin Kumar:
Yeah, it’s hard to put a number, Prakash and I can check with our enterprise team and maybe get back to you and get back to your listeners offline as well on that front. But because each enterprise is different, every enterprise has different set of requirements, but as a broad percentage, I don’t know the exact percentage on that, but we are confident that a significant portion, a high percentage number of those applications are native and beyond that, a very large number of those applications are compatible with our enterprise team.
Prakash Sangam:
Sure, no problem. So, and also when I talk to some of the app developers, AI PCs kind of give some growth in the overall PC market itself. Otherwise, it was basically kind of replacement for them. So they had to invest more to get to the same amount of market, right? It’s not just developing and porting, they have to manage two streams of these applications, both for X86 and ARM, right? So it’s quite a bit of investment they have to do. What kind of incentivization or encouragement that you are doing to the app developers to do quick porting and also support native apps on Windows on ARM?
Nitin Kumar:
I understand, Prakash. Let me maybe take a step back. Of course, one is the on-device AI capability and AI is a fundamental technology that is just going to transform multiple industries and create new applications that will truly, everybody will benefit out of that. So that by itself is a big enough motivation if you look at it to what the technology helps on that.So that’s number one.
But outside of on-device, leading edge, 45 tops, and PU capability on X Elite and X Plus, there are fundamental advantages that we provide on the platform, outside of the AI that we just talked about. And out of that, battery life is another key advantage that is just going to be of a disruptive value to the end user when you look at what X Elite and X Plus platform provide.
And let me take a couple of examples on that. When you look at our X Elite platform and you compare that to a latest from a competition, like an Intel Meteor Lake platform on a Core Ultra 7 type of a device, our internal measurements that we have done across a variety of use cases, could be a web browser use case, could be a video playback use case, a Teams call, a use case in an office environment use case, or another productivity application. The battery life improvement we are talking about is like anywhere in the range of 40-50% to up to 2X. And I’ll repeat that. Like battery life improvement in the range of 50% to up to 2X, and that is huge.
That just completely changes the paradigm of what you expect out of the device, how you look at your device, what you get to do with your device. You can truly get like multi-day battery life and you don’t need to think about carrying that charger brick with you or leave the charger behind and truly get that mobility experience from a mobile laptop device. And that enhances or provides a significant value for the devices powered by Snapdragon architecture.
Now, when you combine that with on-device AI capability that we talked about, significant improvement in battery life or performance efficiency, along with leading performance. When you look at all this combination, this is a unique disruptive value proposition that we are providing. And it helps the developer ecosystem understand and realize that value. And this is the industry shift. This is where the next generation of PCs are going to and moving to, as well as we provide a good tool set change to help them port their application onto
Windows on Snapdragon architecture. I think it makes the story sellable on its own, if you will.And Microsoft Visual Studio, as an example, is a very popular application development framework, runs native and can help you compile your application natively on Windows on Snapdragon architecture.
Prakash Sangam:
The PC business is not just about performance, right? You know, if you start to look, there’s a lot of market development firms, subsidies and others that play a key role as well. I can understand you cannot disclose a lot of details on a public forum like this, but what are you guys doing, you know, to make sure OEMs are on your side other than just the performance itself?
Nitin Kumar:
No, absolutely, Prakash, and yes, we understand the PC industry, we understand how the market operates. We are very proud of our channel partnerships. You’ve heard some of the announcements, you’ll hear more as you go through, and then the devices become available. And in a few weeks, you will see the broad reach that we have working through our broad set of partners. And again, we are very proud of those partnerships.
In a summary, yes, we understand the industry, how the industry operates. We are well aligned with that, and we are investing on all fronts to make this a success and to make sure that we succeed on our mission of driving the best experience PCs.
Prakash Sangam:
Yeah, I can understand you can’t disclose more details. Yeah, this is my last question. So, in terms of consumer recognition, I mean, Snapdragon and Qualcomm is not a well known name in the PC industry, right? So you’re launching so many devices. I’m sure your OEMs will do their part in terms of customer education and so on. But what specifically you guys are doing from Qualcomm perspective to make your brand known? Like, I mean, everybody knows Intel, Intel inside that kind of people take it for granted. So what do you guys are specifically doing to increase that brand awareness?
Nitin Kumar:
Absolutely, Prakash, yes. I think Snapdragon is a very recognized brand and Snapdragon as a correlation with premium experiences. We have a very strong brand presence on our mobile portfolio. And on the PC, you’ve seen, we have made several investments on that front already. We have added a new category over there, Snapdragon X Series with X Elite and X Plus product portfolio already announced. We’re investing in that brand.
We’re driving awareness. And our goal is to have the premium PC experience associated with Snapdragon X. And you will see many more announcements on that front as well in the future. But yes, we’re making those investments on Snapdragon X.
Prakash Sangam:
Yeah, I’d be looking forward to more details on that as the rollouts happen so that people, when they go to Best Buy or online, when they’re checking out PCs, they feel comfortable that, they know this is a known brand and they’re buying something that they fully believe in. Obviously on the mobile side, it’s there, but the awareness need is a little bit more on PC side. We’ll be interested in learning more there.
Okay, so this has been a great discussion, Nitin. Thank you very much for all the insights. I think this and next year is going to be an exciting and crucial time for Qualcomm computing. So all the best and once again, congratulations on this fabulous launch event with Microsoft.
Nitin Kumar:
Thank you so much, Prakash. Really enjoyed talking to you and absolutely, yes, we’re super excited and we’re on this journey and couldn’t be more proud. Thank you.
Prakash Sangam:
Thank you again. So folks, that’s all we have for now. Hope you found this discussion informative and useful. If so, please hit like and subscribe to the podcast on whatever platform you’re listening this on.
I’ll be back very soon with another episode, putting light on another interesting tech subject.
