NVIDIA Corporation (NVDA) Earnings Call Transcript & Summary

Hello all. We're back. Welcome back. Happy New Year. I hope you had a good holidays. Richard took a break in the snow. He took a break in the rain. And depending on where we are, I guess, you could tell probably from that statement. But I hope you all had a good holidays and took the time to rest relax, enjoy the food. And maybe some art work while you're at it. We're back this month with Level up within NVIDIA RTX in a real engine 5.1. We're going to be talking about some, some developments that have been going on, what are we working on? And some interesting things I know Richard is excited to share, which I will gladly let him speak to. But for those of you not aware, we're going to have episodes coming up in the next month and following as well so please keep your eyes open for future episodes and any of our marketing material coming out on them. And yes, so thank you again for joining, and I will let Richard take the floor about state of RTX today in 5.1 and what's going on?

What I wanted to do was start by talking about, at a very high level, what is the state of RTX because we're at the start of 2023 now, and real-time ray tracing began in 2018 in the Unreal Engine, and specifically unreal engine Relign 4 and a much older version than what we have that what we're working with right now. So a lot's changed in 4.5 years. And this is a really interesting topic, right, because in the -- yes, I think it was April 2018 when it came out. So in the approximately 4.5 years since ray tracing was introduced to the world. You could say that ray tracing RTX has gone into at least a second or maybe a third generation of refinement with new capabilities and like new levels of performance and features that weren't there originally. So yes, really, I just wanted to recap that, and talk a little bit about where this is going because as people might remember when real-time ray tracing started, it was really sort of -- it was very segmented features. We're talking about like ray trace shadows, and that improved shadow quality, but it didn't really give you any new capability. There was RTGI, if people remember that, ray traced GI, but it was very experimental and very much a beta, and never actually became like a final usable product. It informed other products. It informed other forms of global illumination like Lumin, but in itself was ultimately deprecated and we sort of moved on from that. So a lot of advancements in lighting and reflection in indirect lighting and performance and speed and capability since it originally started. And yes, you could say that we're heading into a second generation or maybe even a third generation refinement to what RTX is and what it's going to be going forward into the future.

In short, I think we're going to be talking about where we've been, where we currently are and where everything is going and what all this tech means in 2023?

Yes. RTXDI is really a big part of that. I mean when we're -- when we talk about retracing RTX and so forth. And what's the goal of RTX, what are we trying to do with it? I think you can probably say in a nutshell, we're trying to give you more realistic rendering, more realistic lighting, better graphics in a real-time sense, lighting and rendering that you don't have to bake down seems that are fully dynamic that open up new possibilities and new capabilities with the kind of scenes you might have wanted to do in the past, but weren't able to do. So RTX has sort of had that goal from the very beginning. But like I said, we're heading in -- we're in a phase now where we're, I would say, taking that capability even further. RTXDI is sort of like the ultimate expression of that because -- it -- not only does it give you very high-quality lighting, it gives you -- quality lights. It also gives you very high light counts, as you can see with the scene here. This is -- this is an unreal 5.1 scene. This is in the NBRTx branch. And so this is -- every light you see all these emissive point likes and so forth that are animated and moving. They are real lights. So there's a ton of them, there's a ton of shadow casting going on. This is a fully ray tracing too. So we're doing ray-traced foliage. And you can see like shadows from the grass the grass like moving and swing in the wind. This was a feature that you couldn't do when real-time retracing started, and we've resolved that as part of NBRTX. So a lot of -- and like I said, advancement in new features and new capabilities have come into it. Here, you can see another feature of NBRTX-5.1, and this is the translucent reflective surfaces. Here, we've got a glass surface that you can both see through, and it's doing like transition reflection on the surface. So I can see what's reflected below like, above it, I got a nice mirror image of what's happening in the scene, and I can still see through the glass to what's underneath. That's a 5.1 specific feature, and it's meant to work with Lumin, and it also works with hardware rate tracing and RTXDI. So all of these technologies work together to give you a more capable, dynamic real-time scene. And everything here, of course, is fully real time and taking advantage of the engine. This, by the way, too, the scene is going to be a showcasing that we'll be releasing to developers. There'll be a stand-alone build there'll also be project files. So people will be able to get into the source assets and see how all this is made, a better understanding for how everything is done here, and really terror to part and dissect it and get a lot of best practices type information out of it. Yes, just to show like how capable RTXDI is a rendering technology, I mean, we've talked about it before, you've seen it before. But here, you can see a bunch of animated moving lights, right? There's a ton of them on this -- that swinging back and forth. And there are different colors and they're animating different ways and so forth. And you can see the moving shadows from the objects. And it can do this with very high poly geometry, right? Like that's not -- in the old days, this would have been like an alpha texture, we would have faced that. But now with a technology like, we just modeled the geometry. And trace raise, trace shadow rates directly off of it. So yes, as the light moves over, you get very physically accurate, very close to photo realistic light and shadow behavior in a real-time sense.

I think that's the thing we're trying to aim for is the closest thing that we can get to accurate for the realism or at least in real time, especially it's kind of like the -- what everything is kind of trying to -- like the path that we're going down and the goal that you'd like to achieve and beyond that as well, things like path tracing.

I wanted to bring up a web page that we have published. This is on the NVIDIA website. And I'm sure we'll have this link for people. But this is something that we wrote last year. It's just -- it's a blog, but it defines what is path tracing. And if you -- this is actually an unreal 5 image here at the top, that's a rendered path-tracing Image. But if you step through the article, I want to highlight this to people because when we talk about RTXDI getting us like a step closer to real-time path tracing. Because what RTXDI does is that it happens up all your light sources and reduce them in a single pass. So it doesn't matter how many lights you have in the scene, right? You can have 10 lights, 100 lights, it doesn't matter. And then it renders all the retrace lighting at the quality of the path tracing, which is higher quality than traditional ray tracing. And what does that mean? Well, this comparative image is really good to help understand that because here's a path trace image or a trace image and a rasterized image. And you'll see that like the rasterized image, it sort of got the least amount of information, right? It's got like missing bits in the reflection and the geometry looks a little more flat, it's missing specular, highlights. You can get those things out of a rasterizer, but they will be approximated. They won't be real. So they might look wrong at certain angles and things like that. Or you'll just do a lot of like sort of handcrafting and try to get the correct look out of it. Ray tracing began to solve this, and this is really ray tracing from the past 2 to 4 years, right? So traditionally, ray traced images, they by firing rays into the scene that would then bounce and then strike objects behind the camera, you would begin to gather more information in the scene that would otherwise be missed by traditional rationalization. And the thing that needs to be understood about a ray trace damage, a real-time rate case image is that it's not getting rid of rasterization completely, right? -- was still used restorization basically as a starting point and then do ray-traced effects on top to make a more complete image. So it never abandoned traditional rasterization, right? -- it would do certain things ray trace and then keep the restorization and sort of combine it and try to give you a better looking at overall image. However, path trade images, we're not doing real-time path creating in on real 5 today. But as you can see with RTXDI we are incorporating more elements of real-time path tracing. In that case, we're doing the direct lighting like the light and the direct shadow, is basically path trace quality or very, very close to it, like 95% close. And when you have full path tracing, this is an offline path traced image here, you can see that it's getting much more information, right? Because path tracing is basically true path tracing would abandon traditional rasterization techniques and go for very high accounts of rays being fired into the scene and bounced around and sampled multiple times. And you do that in order to try to find what we call ground truth. That's what you're aiming for when you do a path traced image, you're trying to find the ground truth like what does each pixel, what is the correct color for each pixel that I'm attempting to render. And -- so that's traditional offline path tracing, that's what that is. But you can see like where we've been going is we've been trying to get closer to this offline path traced image. And right now today in Unreal 5, I would say that what we have is a sort of a -- with RTXDI to do your lighting and then Lumen, using hardware retrace lumen. The combination of those 2 things and then maybe also Nanite on top of it for very high complex geometry. The combination of those technologies makes it possible for you to have a sort of an approximated or baby path tracing, look to the game and doing a real time and at high frame rate. Now of course, these technologies do require more powerful hardware. But this like the sort of approximated path tracing or baby path tracing that we're going into right now is very performant already. And we have a bunch of technologies have been developing around this like DLSS, like SCR, like OMMs to basically speed it up. None of those technologies should like -- the output an image or change it fundamentally. They just give you a way to dramatically speed up your seat and make real-time path tracing more possible. So like while we're not doing like full real-time path tracing today, we're doing elements of it. And you could say that what's possible with RTXDI and Lumen right now today with NBRTX 5.1, which is all brand new, right? Like we're talking about code that is like less than a month old, right? So this is fresh off the presses. There is -- that gives you a form of baby path tracing in Unreal Engine 5, that you could make use of today that it gives you approximated path trace GI, but real -- or very close to path trace quality direct leading. And you get much closer to the offline path traced image than ever before. So that's where we're at right now. That is at the start of 2023 that's where we're at. And I really want to highlight this for developers because we're making -- this ground truth image is where we're trying to get to. We're always trying to get closer to this. Everything we're doing is we're trying to get closer to this image. So the question for us is like how do we get there? How do we make it more possible for you? How do we make that very scalable. And that's fundamentally what we're working on. Higher quality images, faster images, dynamic images, because if it's dynamic, and you don't have to make your scene or do what you want, you can throw in lots of dynamic objects. So you can have lots of destruction, and you can create -- have user-generated content and you can create a truly modern game or you're not tied down to the old ways of static environments and limited pre-generatid content and so forth. So yes, we're looking at sort of those sort of future capabilities are coming online rapidly.

Yes. So on the topic of baby path tracing, I mean, we are making movements into real-time papering. This is not an entirely new thing, right? It's -- in years past, it's been done with minecraft RTX that use the form of sort of simplified real-time path tracing. And like here with the Racer-RTX demo that's going to be coming out soon, this uses RTXDI as it's -- this is like how it's achieving this kind of look what direct lighting is with the technology at RTXDI because it gives you not only really high light counts, but real-time path rating, and they're doing lighting directly from emissive surfaces. And so you see a lit-up keyboard and you get just a little bit of GlowLight coming off that side there and all kinds of highly precise path trace type effects. RTXDI is the foundation for that. And in their case, they're using, I believe, probably resurge for the GI they're using a different GI solution that's not an Unreal Engine 5. But Lumen is very much the equivalent of that. I mean, HDI solution is going to have some differences and strengths and weaknesses. But any modern GI system and Lumen is a very good one. It's very strong. Lumen GI plus RTXDI and the Unreal Engine 5 is going to give you a very close look to what you could achieve here in the Racer RTX demo. And the same thing goes for the upcoming Cyberpunk RT overdrive mode. That's a formal baby path tracing. It uses RTXDI as a basis. Same thing as the RTX portal game that we just released right before the winter break that uses RTXDI and I believe also -- GI for its real-time global illumination. But like I said, the global illumination solution almost doesn't matter because any global -- any GI solution, any real-time GI solution today is sort of an approximation of like true path trace GI or say, path traced indirect lighting. And like I said, efforts are underway to get us closer to a true real-time path tracing, but you can see the kinds of visual effects that we're able to achieve even in like an approximated or baby path tracing mode like this, where we're getting very close. You could say within feet or inches of like what you would expect in an offline rendered scene. I mean just look at that like that lighting and shadow quality. It's a -- right? fantastic. It looks great. Look at that every day.

Right, exactly. So yes, I mean this sort of look is achievable in -- my point is the bottom line here is that this sort of look is achievable in unreal 5 today. If you get Unreal 5.1 NBRTX you can achieve this look right now. And you can do it in real time. You could put -- especially if you're running on a 40 Series car, but you wouldn't -- you don't have to be available. It will run on our 30 series. It will run on a 20 series. If you put DLSS on it, you can totally get that to 60 frames per second and make it scalable across a range of hardware. So yes, anyway, that's -- I just really wanted to talk about some of that real-time capability where we're at and the kind of lighting quality that we're looking at right now.

Yes, it's good to see it actually in a scene and how the letting quality where it's come and how far it is .

And we're not looking I don't want to totally reinforce. We're not all the way there -- we're not 100% of the way there in Unreal 5 right now, but more of these technologies are coming online for Unreal 5, and they are improving and getting better and more capable. So I expect that's not going to change. Like we're making strides to constantly improve that capability and that quality.

All the more reason for people to keep up with these webinars and anything else that we come up with logs write-offs that we're doing videos and stay on top of it. And maybe well across the finish line at some point and low past it. Year-over-year, there's new developments, things that we're working on, things that community is working on. It's not just like 1 solution. There's many different pieces that come together to make -- to build the larger piece of the puzzle. So there are a lot of pieces that come together. And those little pieces could change, someone could stay the same. But it's -- I think it's an exciting time, especially see what the next year or 2 brings in this space? .

Yes. And like I said, that's really what -- in talking about this overview of where we're at and where we've been. I just want to highlight that for people so that everybody is aware, these are the kind of technologies that we have available right now today and more are coming online, right? I haven't even talked about DLSS 3 or frame generation yet, like we're not even at that point. So even more capabilities are coming up. So yes, this is -- you can look at RTXDI as a massive software optimization, too, right, because for goodness, like, I don't know, 20, somewhere between 20 and 30 years. We've been dealing with the same light count limits in that whole time, right? Like going all the way back to like maybe Doom 3, where I believe -- I could be mistaken, but Doom 3 stencil shadows is sort of a very primitive form of retracing, and it I think it could handle no more than 4 shadow testing lights on screen at once. And basically, we've been living with that same 4 shadow testing like limit this whole time. RTXDI opens that up dramatically, blows the roof off of it. So if you want 100 lighting you're seeing, you can do it. This scene here has over 200. Like I said, every light on these objects is a real light.

Speaking about RTXDI we just -- we have some training material that we'll release soon, and we have 1 out there already that kind of goes through how to get started with it. And that's something that's pretty exciting as Richard talked about, allows you to take dynamic like counts from single digits to plus hundreds. And it's definitely an exciting technology.

Yes. And 1 thing I wanted to highlight about this too is like right now, the scene is using -- is using a bunch of Unreal rights? Like so there's light actors that are placed in the scene. And it's letting very much the traditional sort of way. In the RTXDI SDK today, not an unreal engine, right? But in the SDK, RTXDI has the capability of casting pixel precise direct lighting from emissive surfaces. So in theory, you could light an entire scene without ever placing a single light source. You do it all from PBR materials. Now like I said, we have that working in the SDK and certain other engines, I think the Cyberpunk RT overdrive mode that's about to come out. They take advantage of that technology. They're not using the Unreal engine, right? So they've got their own engine. They've taken the SDK into their engine, and they have that capability, where like even emissive like Neon strips, right, can cast direct lighting and shadow off of that emissive. So we're building this in a certain way today, but in the future, once we get that feature brought into Unreal Engine, it -- I mean, that's huge, right? Like potentially, you're talking virtually unlimited lighting. Because there are some limits right now, what we're dealing with today. It's worth noting. I mean RTXDI in the unreal engine right now today has a limit of 256 lights in any given scene. Normally, you would never go over 256 lights, why would you do that, right? No scene can handle that many lights and not crash in the process. But we're working on raising that limit for NvRTX. And then if in the future, we can bring that SDK feature of light casting firmamissives then you don't even to worry about light counts. So just -- it becomes completely -- control.

I mean putting the power in the hands of the artist is kind of where ?

Where it all happens. Right. Yes, I don't know. It's so big to think about these things because they're -- we're talking about capabilities that were not possible in the days of rasterization. And ray tracing started by improving quality. It didn't -- in some ways, it didn't fundamentally change features, maybe with reflections that data gave you like being reflected what's behind you. That's new. But -- but like in the case of shadows, no fundamental new features, it was just improved quality. Now we're looking at not only improving quality in a dramatic way, but improving what you can do with it, like on a manyfold over what's possible with the technology with rendering with real-time rendering.

How about we just kind of briefly touch on it just a little while ago, but exactly what does maybe patch tracing mean?

Sure. That's what I -- I'll be honest. That's what I call it, and that's what some people at NVIDIA have called it. You can call it approximated path tracing. I think it means a technology like RTXDI, which gives you path trace quality direct lighting, right? That solves 1 part of the path tracing equation. And then a reflection system and a global illumination system. In the case of Unreal 5.1, we think that's a lumen. If you if you do hardware ray trace lumen for your GI and your reflections and you combine that with RTXDI, then you basically have an unlimited lighting system. And then if you merge that with a very high poly geometry system like nanite, you have a virtually unlimited rendering system completely. So you can have millions and millions of triangles potentially billions of triangles in your scene, and a very high light count, maybe not completely unlimited, but very high light count, and GI reflections everywhere. And it -- that's more unlimited rendering than we've ever had before, in a render in a real-time sense. So I think that's -- something like that is the sort of the modern definition of this halfway step to true real-time path tracing. I mean, like I said, we didn't call it baby path tracing, which is for short, right. And really, if you just -- if it's RTXDI plus almost any high-frequency GI system. Lumen is a good example of that. It's got excellent hardware RT support. It's got -- they can do high-frequency detail. So -- and it's been improved over version 5.0, not just in performance, but in quality. So that it looks much closer to the off-line path tracer. In fact, I could just -- I can show that real quickly here. Yes, let me pause things for a little bit, so over to this. So yes, this is the real-time scene, right? I mean when I say path tracer, we're talking about the off-line path tracer in Unreal 5. So I can be in lit mode and then I can switch to the path tracer. And you also -- I'm not going to wait for it to resolve because this is going to take too long. But this is the offline path tracer, right? And this is what it does. It's recursively aspiring millions and millions of rays into the scene and it's bouncing them around in a near infinite way bouncing and resampling every pixel multiple times to try to find what is the true life value, the true color value of each pixel on the screen. So that's a very expensive operation. And you can see from the thermometer bar this takes a long time to do, right? -- just to generate 1 frame. So if I switch back to lit mode, it's not exactly the same image, but it's very close, right, aside from some temperature and heat differences, maybe because path tracing, we'll do like infinite bounces or near infinite bounces on GI, whereas your lit mode rendering, maybe path tracing, if you will, will -- it's much more limited and like the number of bounces and samples that are going on. So you might lose a little bit of heat and color. But in terms of overall quality, the images are usually 80%, 90% similar. You're getting very close. The difference is that 1 is real time and the other is off-line, right? But that path traced image, like I said, there are lots of important reasons for it, right? It's useful as an artist to look at your -- and make sure that your relates are doing what you want, like I want really high-quality shadows and lighting like this is all path traced quality down here. I can switch to the path tracer and make sure that it's actually doing what I wanted to do. everything looks -- I'm getting a nice. I'm getting all the geometry is tracing the way it should switch back to lit mode. And as long as I'm maybe like I said, like 90% similar then we're basically there. And for us at NVIDIA, any place where the off-line path tracer and the real-time image differ are areas of research for us. Like how can we get closer to the real-time image or the off-line path traced image, how can we get closer to ground truth. So we're constantly looking at that and sort of analyzing that and it's becoming very informative for us to like understand how to improve real-time rendering.

I think that's a good point also bringing up how close the 2 comparisons are like the offline path tracing versus the real time. Like Richard said, we're getting -- I mean you can tell the difference, but there's not a lot of -- it's not like a 50%, 60% difference, right? It's getting. And I think that's the thing that's really exciting is especially from coming from official effects and type of graphics game development backgrounds. It's for the audience it's really something to be excited about like we are every day behind the scenes exactly.

If what you're doing is -- I mean -- could be huge for you if you're doing like some type of rendering or video production, because you want your real time seem to be as close to your final output as possible. If we can get the real-time output to 90% of the quality of the final output, then that's incredible by itself, right? Because usually, what your realtime output, you're looking at simply shaded triangles or stuff with the texture or whatever. So we're definitely not -- typically not like this sort of like real-time lighting complexity. So yes, we're -- yes, this has many uses beyond just games, right? Applications, VFX and really any kind of real-time rendering scenario where you need to try to get the highest quality lighting possible.

And I think it's all about optimization in the end as well. And things like DLSS, SCR kind of further optimize even O&Ms?

SCR in the NvRTX branch. I can show you where that's set out. If you pull down the NvRTX, you can get the SCRs for state execution reordering. And if you have a 40 Series card, what it does is it optimizes complex, like retracing calculation data into a more optimized workflow. So that -- the 4 Series card can execute it much faster. So that's basically a shader execution reordering. It is taking shader related data and reordering the data and executing in a much more fluid way. And that also lets you know, too, that SCR is not going to -- it's not necessarily going to speed up everything 100%, right? We say SCR is going to speed up retracing operations, retracing and path racing operations by up to 40%. Now that doesn't mean you're going to get 40% more frame rate, right? That means that your retracing operation can be up to 40% faster. And like since it's shader specific or shader related, that means it will speed up things like translucency or reflection or anything we're doing a ray tracing operation or path-tracing operation involves a shader calculation. It typically won't speed up things like direct shadows, right? So we do a lot of optimization already on RTXDI to make sure it runs fast. So SCR doesn't have like a direct benefit on certain ray tracing effects like that. I just want people to be clear on that, but it's not 40% faster frame rate automatically. But depending on the ray tracing operation you're doing, and the more ray racing you do, the more benefit you get out of it, like the most benefit you can get out of SCR is at an offline path traced image because that's your most complex sort of set of ray operations. And SCR has a dramatic improvement on offline path tracing. So if you're doing video work and you're baking out images, SCR can have a huge uplift if you have 40 series card. But anyway, I think by default, we have that feature off in NvRTX, and to turn it on, it's -- you can do a search for it. Just by typing the word reorder. And you'll see our various SCR commands console commands right here. You can build these into your default engine for your project. But for -- just for example, you'll see like we do various speed ups on Lumen because if you're using hardware rate trace lumen, it can take advantage of SCR. And this is another reason why we think Lumen is definitely the way to go, right, is the most robust, the most complete GI and reflection system in the Unreal engine and it's built into it. And like if you're using NvRTX, you got an additional speed boost on it. So you can see like here, I can speed up the operation for Lumen's screen probe gather when hardware ray tracing is on and just by toggling a bit. And you'll see it frees us for a moment and then it -- it just continues. And if SCR is working, it is completely transparent to the end user, right? You will never see anything visually wrong with the scene. There shouldn't be anything visually different there's not much to show about it, right? It's working because I turned it on. And if you look at stat GPU, you might see some millisecond savings in certain operations. It just kind of depends on like I said, like how complex you're seeing is. But yes, just stepping through them, let's go back to that. Yes. Like I said, the good way to find it is just type the word reorder, and you'll find all of them. There's 1 for path tracing. So that will speed up off-line path tracing. And that's all you got to do. And now you're getting that speed boost off your card.

And just to reiterate, like the hardware and software, what hardware and software is optimized for SCR, so you need a 4 Series card?

Yes. Yes. Right now, today, it works on 40 Series hardware, right? So we just released 40-70 TI so work on that end up. But in theory, it works on any 40 series hardware. For a current game or an application to take advantage of it, you have to get the SCR optimized code that's in NvRTX, and like rebuild your binary with the latest code. If you're good at managing your own code, look if that's your normal thing, then you can just cherry-pick the code out of our branch and get to work with that if you want, or if you want to pull down NvRTX and work with that, that works too.

All right. So heading to our last part of this session, you can talk about micro meshes and the role that they -- could potentially play in this workflow like whatever you're viewing?

Yes. So we don't have them in Unreal 5 yet. We do have a developer page up on it, so you can understand what we're talking about. As NVIDIA micro mesh page. If you show down a bit, you'll see micro meshes is really 2 parts. It's displaced micro mesh and the opacity micro maps. So DMMs and OMMs. Now spoiler alert, I'll just say right off the bat, the MMs probably won't be coming into Unreal 5. And that's because largely, they do the same thing. Well, they achieve a very sort of similar end result to Nanite. So it may not make much sense to have them in Unreal 5. I think we're still kind of evaluating that. But like just at a very high level, DMMs and Nanite, you could say, both aim to achieve much higher triangle scenes in a more optimized way, right? They do it in totally different ways. The technique and method is for rendering is 100% different from each other. But the end result is very similar. So for that reason, because Unreal fiber already has a solution for very high poly scenes. DMMs may not make much sense for Unreal 5. It's still an SDK, it makes a lot of sense for other rendering engines, just maybe not this one. Opacity micro maps on the other hand, will have a great benefit in Unreal 5. And we don't have it in the engine yet, but it is coming. And OMMs -- so opacity micro map and the whole idea is that it's -- when you have a you have like a bit of foliage and you'll have like a large triangle covering a space like this. If you just look at it from a rate tracing standpoint, it has -- the ray tracer has to test against all of these translucent parts of the triangle to determine whether or not it should be rendered. And that can be very wasteful. So OMMs will give you a way to sort of eliminate that excess triangle space and make it so that the ray tracer can operate on tested, alpha tested geometry much faster. So their impact in Unreal 5 should be very outsized, especially in flat for -- scenes or any type of scene where like you might have a lot of cloth or anything like that. So Yes, I'm sure when we get to a point where we can announce OMMs for Unreal 5, we'll do that. But yes, as of today, we don't have it in NvRTX yet.

Great Well, I think you've kind of touched on a bit of stuff today. We are having our Q&A as usual after this. So please stick around, ask us questions if you have any. Thank you for showing up again, joining us in the new year. Excited to have trying to plan out what's going to happen for us for the next few episodes and the rest of the year as well. And hopefully, we'll have some exciting things planned guests and partnerships. So looking forward to sharing that with everyone as time gets closer. But as for this episode, thank you for coming.