B (4:58)

That's awesome. So you know it's interesting because I an example I would use is kind of a mixed reality kind of what you said right is like for example you know, if you shop at all and you know, as you can tell I'm in my house wife likes to decorate but she was showing me something cool and it's been a while back but where you know she was looking at the couch she wanted to buy and was able to kind of put that couch in our living room to kind of see what it's like right. So see kind of it mix to see. So it was a very cool experience but that's like you know, kind of a consumer. What we're really here to talk about is kind of the, the enterprise piece of kind of overall xr. So you kind of touched on this, you know GPU compute obviously very important the Nvidia RTX Pro. But what are some of the other like core components that are required, you know, whether it be hardware, software, you know, licensing to have if you are XYZ company to set up and run an XR experience.

A (5:58)

Absolutely. So we need to consider the foundation requirements like both from software and a hardware perspective. And first of all most important thing it's all about providing the high fidelity and photorealistic experiences. And the kind of experience that I'm talking about is the one that truly trick your brain into believing that you are somewhere else or you are interacting with an object as if it is real. And it requires incredible powerful GPUs. Now we all know that standalone VR devices today in the market, they all have strong compute power, but they typically aren't sufficient for handling the massive data sets for enterprise grade XR applications. And this is where powerful workstations like Dell Pro Max with GPUs like Nvidia RTX Pro become it essential. But on the other side one, another important piece of the puzzle is a specialized software for specific industry workflows, whether it's for design review or for simulation or for training. And we see graphics quality getting better and better every day. And the user's enterprise users are fully immersed. So we talked about compute power, we talked about software. And the third component which is increasingly important is nvidnims, what we call inference microservices. They automate the task within XR environments and they can power intelligent virtual assistants, enable interactions with, with your natural voice and it goes beyond traditional controllers. So AI, I think we should know, we should open apprentices for that and talk more about AI in XR and XR for AI. But one another critical aspect in enterprise grade XR setting is also collaboration. Like you truly, in order for you to truly have the potential of XR in the enterprise setting, you need just more than just a single isolated experience because it means that you are completely free from cables and ability to collaborate with your colleagues from wherever they are in the world. And so this is where the streaming solutions like Nvidia Cloud XR becomes absolutely essential. So the compute power should be there, the software, the AI and also the collaboration aspects. So all these are critical parts of the one big puzzle.

B (8:52)

I mean, it makes sense, right? Like, I mean with any, anything that, you know, we cover on this show, there's always a hardware, software, kind of a data storage component. All of that's, you know, really included. And one thing I want to touch on really quick and we can definitely get into AI here in a little bit more in the episode. But you kind of touched on like the, the headset or the untethered experience, right, which is, is very unique. And I mean, tell us maybe a little bit more about kind of how XR is evolving away from quote unquote like a tethered headset to a more immersive experience.

A (9:27)

So here I think the most critical aspect is the collaboration. So for example, we were in a design automotive design conference in May in Germany and in autumn. For automotive designers, it is very important to have the design reviews with your calling colleagues within the same virtual environment. You cannot just have the design review session alone or you cannot just rate or comment on a design alone. So with that it is becoming so critical to have the streaming solutions that support like multiple users where the users could be immersed and have joint feedback on one virtual object.

B (10:10)

Okay, I mean that makes sense. I mean let's actually jump into an example, right? I mean the collaborative piece kind of makes sense from an enterprise where you're designing like a design prototype with a car. But let's dive a little bit more into that. So you said, you know, you were at the event, I think you said may, you know, an automotive show. Can you talk a little bit more about the workflow where let's say XYZ company is designing a car or you know, manufacturing companies designing X, maybe talk us through that workflow a little bit more.

A (10:39)

Of course, let's start with an automotive, as we just mentioned, so in automotive we are, is part of everyday workflow now and it's because it is a critical component of, of the design review process. So instead of building a super expensive physical clay model for the design review, designers are using VR to simulate car prototypes and they test materials and they evaluate the lighting conditions in photorealistic detail and definitely speeds up iterations and cut cost. Because thinking about those expensive clay models, it takes a week to get a new version and you know, especially the one to one claim models are super expensive. Instead you can just tweak the model in, in, in VR and then, then review it with your colleagues immediately. So that's that, that's a huge important workflow for automotive, but in another workflow that we see the ability to simulate the car design into in multiple environments, which is so important for designers. So here the stable diffusion models come in, you maintain the design consistency while you change the scene, the lighting conditions. It's a super powerful capability for the designers and we can talk about, you know, manufacturing and Asian. So many different workflows that are out there.

B (12:19)

That's interesting. I've never heard so obviously very familiar with, you know, diffusion models like flux and stable diffusion, et cetera. That's really interesting. So you're saying that within, you know, different scenes and you know, augmented environments, they're using stable diffusion models, you know, pick their choice of which one, but to change the scene Rotate kind of in real time to use AI to power their XL workflows. Is that what you said?

A (12:45)

Exactly. So I want to see my car design in a like different light condition. I want to see the car design in a desert or in an ocean back wheel or in a, you know, like a studio kind of version. So it's, it is a powerful, powerful workflow for them.

B (13:05)

That's awesome. Well, I know we could talk about workflows and we can get into. I do want to cover one that you and I kind of talked about before that I think is, is really, really interesting and is kind of within healthcare, kind of around the ability to plan and virtualize out and practice kind of a surgery before you actually have a person undergo that operation. Can you talk a little bit about that? Because I find that extremely fascinating.

A (13:30)

It's fascinating, Logan, because it touches the human lives, right? It could potentially save lives. You know, it can improve the perceived decision of the procedures. And so in healthcare, the most important workflow we see is surgical planning and medical training. So surgeons for example, can take a patient's MRI or CT scans and they transform this into a highly detailed 3D models which allow them to practice super complex procedures in a virtual space. And on the other side for medical students, Xar offers a realistic simulation environment where they practice surgeries and learn new techniques, which is so critical. It's, it's completely risk free setting. This is, this is where it touches the human like because it's eliminates all the risks associated with the real, with the operation right when you are getting trained. And this puts like confidence but also the precision of the, of the operations and absolutely improving the patient outcomes. What do you think, Logan?

B (14:51)

Yeah, I mean I, I think it's really interesting. That kind of twofold of that is what I find most interesting is one, I mean I obviously haven't been to med school, but I could imagine kind of the very first time they're going in doing a procedure, you're like a brain surgeon or heart surgeon or whatever. You're probably terrified because you've had some practice on, as morbid as it may be, some cadavers and stuff like that, but you've never really practiced like your first run is on some like a living human being. And I think that that is just fascinating. But to the other piece I find fascinating is being able to, and it's maybe not AI, but being able to inject. If I was undergoing heart surgery for them to be able to eject my body, my scan, my X rays, so when they are practicing, they are ultimately practicing on a visual augmented reality version of Logan, which I just find incredible. And I think that's what I think is a bit different about XR versus some of the industry and workflows that we talk about is it has real app and not the others don't. But like it has real applicable in everyday life where you might not even see it, but ultimately is benefiting you, which I find just honestly incredible. So you know, we kind of talked about, you know, some workflows, we've talked about kind of the software hardware requirements. But as you know, at GTC and then just recently, within the last, what, a month, probably month or so, we launched our dell Pro Max T2 towers with the Nvidia RTX Pro 6000 black wheels, which are 600 watt. You know, obviously that's very important for, you know, a lot of different workflows, whether it's simulation, AI training, I mean, whatever it may be. But for xr, why is that? What are the kind of key components or underlying features of that GPU that makes it so relevant to people who are building XR workflows?

A (16:38)

RTX4 6000 workstation edition here the level of power and performance is long awaited by the XR community from different fronts. First of all, the combination of 96 gigabyte of memory and 4000 AI tops enables users to run massive scenes and LLMs in parallel to the XR workloads. So now it's enabled, enables AI for XR and it's a huge opportunity for the developer community. Great experience for the enterprise users. And now also it offers up to 80% faster rendering performance compared to our previous generation GPUs which is tested in VRED. And one critical feature that comes with RTX Pro is now it enables the dual GPU support with VRSLI over PCIe Gen 5. So basically in simplest terms this means you can dedicate one GPU per eye and it provides enormous power for running high fidelity XR experiences. We put ourselves in a very tough test. We recently demonstrated real time ray tracing with 6,000 in an automative show enabled by Vulcan rendering. So real time ray traces tracing is so critical for the designers and the users could simulate like real time physics and the lighting and created incredibly realistic environments for the designers. And this was the very first public demonstration of its kind. And we received a super positive response from the OTA designers.

B (18:38)

That's really cool. I mean I think the dual GPU support, you know, and for those that don't know, like Nvidia RTX cards, obviously very powerful, whether we're talking ADA generation or Blackwell generation. But, but you know, those cards are very different in a workstation or Del Pro Max than they are in servers because they don't typically use NV link. Right. An NV link is just a connection that basically strings those GPUs together as one common thing. So even though you might have, and I'm not a mathematician, but two 48 gig GPUs times two, put them together, that's 96. But in workstations you can't necessarily put those together. Right. But being able to run those workflows, you know, per GPU is really fast, is really amazing to be honest, because you're basically not only doubling the amount of performance with Blackwell, but now you're in theory doubling it again purely because you're not splitting kind of eyes. But one thing you talked about, you know, outside of the ray tracing, which is awesome and then you know, how quickly things can render almost in real time in the AI capabilities. But let's talk about the frame buffer bed. Explain a little bit more about, because I've heard the term but I think it'd be good for the audience to hear like within, you know, an XR kind of like from like developer, but also from you know, a person immersed in an experience. Why does frame buffer matter so much?

A (19:56)

So the large frame buffer here is important because then it gives you the ability to run AI and XR at the same time. Of course it's, it's important for the rendering capability. But for us as now AI is, is becoming the part of the XR experiences. The frame buffer is becoming extremely critical here. The more AI we will run alongside with the XR then we will need more and more frame buffer for a good XR perspective. So with RTX Pro 6000 we offer 9006GB of memory with 4000 AI tops, which is an amazing performance.

B (20:42)

It's huge. I mean, it's huge. I mean I, and I have only been kind of doing this role for like a year and a half and you know, I came on kind of when ADA was kind of launching and just to see the improvements and the advancements in Blackwell have been fantastic. But you know, one thing I like to do on the show is, you know, I like to live in the here and now and what's possible. But I always like where our guests who are obviously experts domains in their field or experts in their domains is to kind of ask them looking forward questions. Right. So we'll start off with one, you know, you know, look at your magic eight ball and I guess the first question is, you know, what do you see in the future of XR in the enterprise space? Like fast forward a year or five years from now. I mean, what do you see some of the emerging trends or breakthroughs that we can see potentially in the coming years.

A (21:28)

Not an emerging trend anymore, but what excite excites us to most is AI for XR and XR for AI. Now AI and XR convergence has moved from just a hypothetical discussion into a real world impact. Because what we see and what we believe is that AI is becoming the UX for XR and the way it changes the way we engage with the digital worlds with our voice. Like imagine just moving beyond clumsy controllers and AI makes enterprise applications more accessible. It's enable conversational interaction within xr. A perfect example I just mentioned is Nvidia nims, our inference microservices for example. There will be more and more use cases with AI and xr. And what also what is so also exciting exciting for us is the how an AI enables the lightweight AR glasses. So just want to give you an example like assembly line visual guidance use cases with the perfect AI example on how it's used in the training settings like glasses can stream live video of the workers view to AI in AI overlays the digital instructions and tracks the progress and provide real time feedback and guidance. And we can now predict anomalies in any kind of a standard operation like through cameras and microphones, you continuously capture the data and AI models can just learn normal operations and anomalies trigger alerts to the supervision. So now AI is bringing lots of benefits to the table and it completely changed. It was becoming a completely different world from an XR perspective. So I think this is, this is the most important trend we see. The other side of the coin is now XR enables the real world training for AI because XR allows to teach the robots for the complex operations through imitation learning. Like a human can perform an action in VR and AI just learns these demonstrations and acquire new skills. So then we train the robots without real time real world consequences in a completely risk free environment like tools like Isaac Groot for example is an example. So the most important trend that I see more and more impacting and what will move the SAR to the next generation of experiences will be the AI in my opinion.

B (24:31)

I mean you're probably not wrong because I mean that's where everything is moving and you made a really good, a really Interesting comment where you're using AI like an XR using AI to kind of real time train. And we had not necessarily an enterprise level example of this, but at Dell World, one of the partners that I have is Dauntless xr and they'll be on an episode, upcoming episode. But basically what they did for Dell World is they took, there was a LEGO set which was building a Dell AI factory with Nvidia, so it had the LEGO parts. But what they did is they trained kind of an AI model to learn in the order and recognize the pieces in which order they needed to go where they needed to go. And then through that experience and that lens, they were able to overlay. So it was more mixed reality. But being able to have the person spot and recognize kind of with computer vision to be able to say, hey, that's the right piece. And then being able to say, okay, it goes here versus it goes here versus it goes here. And that, that is a really interesting use case, especially in, you know, I mean, that's just for fun, you know, at a trade show. But, you know, whether it be putting together like a complex piece of machinery or, I mean, sometimes I just put together some lawn furniture and good God, I would have loved to have it because it was the most complicated thing and the instructions were terrible. I would have much rather had I got an XR experience that would have guided me through it. But I love those examples and I love how it's a two way street. Like, you know, AI is kind of the UX for XR and XR is training AI and it's like this cyclical thing. I love it.

A (26:05)

Exactly. So instead of, for example, you just mentioned Logan, instead of reading a manual, you just ask the, you put on the glasses and you just ask, how do I do this? How do I do that? And then, you know, you interact with the AI model and ask questions in real time. It's an amazing capability that we all need, both in the Enterprise setting, but also in, in everyday lives, don't they?

B (26:33)

I agree. I absolutely agree. So, Belgium, this has been a fantastic episode. We're kind of coming up against it, but what I like to do at the end is, you know, for the guests to, you know, kind of pretend that someone watching has just come in at this part of the episode. You know, quickly take a minute or two and kind of recap maybe two or three, four of the key points that you'd want them to walk away.

A (26:59)

Learning from this episode, I want to highlight three points. The first thing is XR is evolving beyond just immersion into a truly special computing paradigm and digital intelligence just understands and interacts with our physical world and it unlocks the full enterprise value. Second thing that I want to highlight today is the convergence of AI and xr. It goes beyond a hypothetical discussion now and it's coming into today's reality. It transforms the way we interact with immersive worlds and leading to more natural, efficient and intelligent applications across all industries. But on the other side, XR is enabling the real world training for AI and finally, how do we deliver great experiences? Powerful hardware like Nvidia RTX Pro combined with Adele Pro Max workstations is enabling a high fidelity, collaborative and AI driven experience.

B (28:14)

So basically just to recap, XR is beyond visual immersion into spatial computing. AI is becoming XR, XR is becoming AI and you need powerful hardware Nvidia RTX Pro GPUs with the multi GPU support for each eye high frame buffer that you can find on Dell Promax workstations. You know whether it be the T2 or whether it be the Dell Pro Max Premium or Plus that enable empower those experiences. So Velgen, this has been fantastic. Before we go, can you take a second and tell everyone where to one kind of find you on LinkedIn and then where they if they wanted to go learn more on Nvidia.com, where they should go.

A (28:56)

Please check our developer page on Nvidia.com or XR page and also please find me on LinkedIn. My name is Beljinja Glorixer and happy to help them. Any questions?

B (29:07)

And yeah Belgian is very helpful. I can't tell you how many times I've reached out to her to ask a question and she is bad analogy but Johnny on the spot she's very quick with the answer. So Belgian, it's been fantastic. Really appreciate you taking the time to chat with me today. Like I said, I told you it was going to be a great episode on XR and we even got into a little bit of AI. So with that this is reshaping workflows with Delpro Max and Nvidia RTX GPUs. Until next time, keep those XR and AI workflows running locally on your Nvidia RTX and your Del Pro Max and we'll see you on the next one. This podcast was produced in partnership with Amaze Media Labs.