NVIDIA GTC 2026 | Rev Lebaredian, NVIDIA & Varun Parmar, Adobe

Clips
More from NVIDIA GTC 2026

Varun Parmar

General Manager

Adobe

Rev Lebaredian

Vice President, Omniverse & Simulation Technology

NVIDIA

play_circle_outline NVIDIA-Adobe 3D Digital Twins: Omniverse Links CAD to Marketing for Pixel-Perfect, Scalable Branded Content

play_circle_outline Scaling Brand-Consistent Content: Adobe Firefly Boards, Analytics-Backed Optimization, and AEM Digital Twin Asset Management

play_circle_outline Alliance of Open USD: standards for carrying digital twins across enterprise workflows and tools.

play_circle_outline Simulation replaces physical photos: lighting, environments, and product configurators simulated for realism.

play_circle_outline Conversational, agentic interfaces enable non-experts to run digital twin campaigns in plain language.

play_circle_outline Availability: 3D digital twins public beta in Adobe Firefly Creative Production for enterprise.

Info
Transcript

Rev Lebaredian, NVIDIA & Varun Parmar, Adobe

Varun Parmar

General Manager Adobe

Rev Lebaredian

Vice President, Omniverse & Simulation Technology NVIDIA

In this interview from the Nvidia GTC AI Conference and Expo, Rev Lebaredian, vice president of Omniverse and simulation technology at NVIDIA, joins Varun Parmar, general manager at Adobe, to talk with theCUBE's John Furrier about how 3D digital twins are bridging the gap between product design and marketing content production. The pair detail a new joint solution that converts CAD files into pixel-perfect digital twins, enabling enterprises to carry a single source of truth from engineering through campaign activation. Parmar highlights that brands expect co... Read more

explore Keep Exploring

How does NVIDIA’s digital-twin (Omniverse) technology integrate with Adobe and generative-AI workflows to enable scalable, identity-preserving content production from CAD files? add

How is Adobe integrating AI to streamline the end-to-end creative and marketing workflow—covering ideation, scalable content production, enforcing on‑brand guidelines, and asset/version management for campaign activation? add

How can digital design data (CAD) and digital twins be reused across departments—such as using a car's CAD model for marketing and advertising—and what role do standards like USD and organizations like the Alliance of OpenUSD (with Adobe's involvement) play in unifying those workflows? add

What are the key marketing and technical considerations when introducing a new car (e.g., colors, pricing, features, branding, differentiation), and how can localized visual asset generation, product configurators, and rules-based automation be used to present dynamic, pixel-perfect product visuals and drive engagement and sales? add

How will you democratize access to the joint 3D digital-twin solution so that everyone in creative and marketing can use it? add

How can companies get involved with 3D digital twins — is the technology available now, how would they use it, and what is the onboarding/engagement process? add

bolt Powered by CUBE AI

Rev Lebaredian, NVIDIA & Varun Parmar, Adobe

search

John Furrier

>> Hi. I'm John Furrier with theCUBE. We are here at GTC for the NVIDIA's conference, 2026, AI infrastructure up and down the stack from energy to AI native applications. A lot is changing, accelerated computing, AI factories. So much innovation here. There's so many storylines, but a lot of them are coming out of the ecosystem. We have Adobe and NVIDIA here. Rev is back. He's the vice president of Omniverse and simulation technology at NVIDIA. And Varun, general manager of Adobe. Thanks for spending the time.

Varun Parmar

>> Excited to be here. Yeah.

John Furrier

>> Day three.

Rev Lebaredian

>> Thank you for inviting me.

John Furrier

>> Yeah. We love the Omniverse. Gave a lot of plugs on a couple of other panels this past year. Good plug for Omniverse. It's been really a huge success. Congratulations on that. You guys have some news with NVIDIA and Adobe. Involves digital twins and simulation and how AI intelligence is coming into the production of creativity, marketing. What's the news?

Varun Parmar

>> Yeah. So if you look at it virtually, every single brand is saying that the need for content is going to go up by 5x over the next two years. And we know generative AI is the solution for content scaling. However, there are certain industries where identity preservation is really important, where you want to make sure that there are exact pixels of the product that are represented and there's no hallucination. That's exactly where the partnership comes in. With NVIDIA, we're releasing this 3D digital twin solution that allows you to take your CAD file, convert it into a digital twin, and then pipe it through the content production workflow that takes the gen AI and scales the content relative to all the campaigns that you need to run.

John Furrier

>> You don't usually hear digital twins and marketing in the same language. And we did hear, Rev, from Jensen in the keynote, the vertical focus with the software and the libraries. Take us through the NVIDIA piece here. How does this tie together with Adobe? Because you got workflows. We love workflows with agents. But what's the AI piece? How does this come together? What is the NVIDIA piece?

Rev Lebaredian

>> Well, the first big piece is the digital twin technology behind this. This is what our Omniverse libraries, our Omniverse technologies enable. Digital twin quite simply is a digital representation, a simulation, if you will, of a thing that's in the physical world. We build digital twins of products, digital twins of spaces and factories. We can build digital twins of humans, digital twins of just about anything. It turns out that almost everything we create these days, the digital twin is actually born first. We design products, physical products that we manufacture in the digital world first before we realize them in the physical world. So this technology of being able to represent physical things inside a computer and do iterations on it and simulation to perfect your product and perfect the manufacturing of that product and perfect the user experience of the product, everything inside simulation is a superpower. Without it, we actually don't really have a chance at creating the complex products that we're creating today and into the future. What we're announcing here with Adobe is really amazing. We're taking these digital twins that we're using as part of the design cycle of these products and carrying those digital twins forward into the marketing and communications part of the same enterprise. You can imagine designing in the digital world first, and then you design the machine that's going to build that, essentially the manufacturing process and the factories that build a physical product. And that same digital twin can then be used to create all of the marketing materials, all of the communications materials around that. And it's really important that you use the actual digital twin because as great as AI is at creating content, at creating things that look great, it's really important when you represent your product that what you're showing is exactly what it's going to be like when people interact with it in the physical world. And you can't really promise that with a pure AI generated image.

John Furrier

>> Varun, talk about the Adobe side, you guys obviously successful in market leading here on this. AI has been a great tailwind for making content. I mean, everyone's a better writer with the frontier models. There's more ideation going on, creativity. What's the focus and what challenge are you guys seeing? Because after a while, it's like everyone's using this image generator. It looks the same, but it's still elementary. So take us through some of the breakthroughs that you guys see here, because I can imagine quality, consistency, brand image. These are things that are hard today if we go in and tell a model, "Hey, be on brand."

Varun Parmar

>> Yeah.

John Furrier

>> Things like that. Where's your head at on this?

Varun Parmar

>> Absolutely. So I think when you think about Adobe, creativity is a big focus area for us, so is marketing. And I think what we've done now is brought in AI to actually really streamline the end-to-end workflow. So right from to your point, starting off with ideation and concepting. So we have something called this Firefly Boards that really allows you to use the Firefly models that Adobe has, but also third party models, so that you can actually ideate and come up with concepts. Once you have that concept and you're like, "I'm ready to go with this," you can actually start to use Firefly creative production for running scale content automation workflows. And in that, to your point, what's really important is that in order to have activation ready content, you need to make sure that it is on brand. So that's an area over the last couple of years, we've actually invested a lot to make sure how can you really instruct all of these generative AI models to understand things like your brand colors, your image guidelines, your illustration guidelines, your font, so you're putting text and so on and so forth. So that's another big aspect of what we are focused on. And then, I think content is a means to an end. What you really want to do is drive outcomes against your campaign. So I think this is where you can take that content, put it as part of a campaign. You need to do asset management. So Adobe has Adobe Experience Manager where you can actually manage and version these things. To Rev's point, it's really important when you create this digital twin, you have this USDZ file. There are different versions of it. How do you make sure that the creative and the marketer and the 3D artists have the latest version? So these are deep enterprise problems that we are solving. And once you've created the content, you need to go activate it. You have your own properties, your website, you have email where you're sending this content, you have paid media. So direct integrations into your activation channel becomes really important. And then, last but definitely not the least is, once you activate this content, you need to know what's working and what's not working. So this is where some of the investments that we've made over the years around content analytics, and then also bringing all of this back into the creation process and saying like, "Hey, there are certain objects, there are certain call to action, there are certain background and foreground colors that are working," and making sure that you're optimizing this process and automating it so that you're actually getting the outcome is where the big value add is.

John Furrier

>> It's interesting. I was talking to a big enterprise in the marketing department and I asked, "What's your biggest data problem?" They said, "Photos." "What do you mean?" "We hire freelancers and they take pictures of the product and there's a gazillion versions of it." And they don't know which one is best because they're all a little bit different. And that's a small, weird example, but that points to a data problem in a marketing use case where his interest was getting a photo, matching it. Give an example of some of the things that you're solving. That's a simple example, but give an example of how you see this playing out. You mentioned campaigns, you mentioned paid marketing. Those are complex workflows and there are many of them.

Varun Parmar

>> Yeah. A simple one is starting off with a campaign brief. So typically what we find is that folks are starting off for the blank piece of paper. And what you could do is, as an organization, as a brand, you've been running all of these campaigns. So wouldn't it be great, actually, that you can ideate with a system that tells you for a given product, let's say Photoshop in the case of Adobe, every year you're running multiple campaigns. What worked and what didn't work. So it's what worked for Thanksgiving campaign and what was the discount that actually worked? What channels worked and performed better? What geographies actually drove higher ROI? And so on and so forth. So that's simple example where you can actually use our technology to actually come up with the first brief for the campaign. You can take that campaign brief and say, "Are there assets that are available in my content management repository that I could leverage for this campaign?" Instead of actually going in and creating net new assets, which oftentimes is very expensive, you could actually use our technology to take a campaign and say, "Give me all of the assets that will perform really well for this audience, for this channel, in this geography," based on a campaign brief. And we would actually surface that. And once you have that, to the earlier point we were making, then you try to understand what's really working at the atomic level. It's not just, oh, an image or a video or an animation worked better, but it's actually the constituents of that actually is really what we're trying to optimize for.

John Furrier

>> So on simulation and Omniverse, take us through this connection because manufacturing, life science, we've seen great examples. That's a lot of compute. What's the goal? And it's not like... You can't do ... AB test was old, like do an AB test. Now you have unlimited simulations that you can do. Can you share how you see that playing out? How detailed is it? Are there agents involved? What's going on in the Omniverse piece here?

Rev Lebaredian

>> If you can take a physical thing and create an exact replica of it in simulation, then the things that you can do with the digital version is essentially all the things you can do with the physical version. So when we're designing a product, when we're integrating a product into a bigger system, when we're trying to communicate the product to your potential customers, all these things are things we do with the physical product typically. You mentioned earlier about taking photos of the product. Essentially, if you have a digital twin and you can simulate what it looks like in different lighting conditions, then you no longer have to be limited by the speed of taking photos in the real world, the speed at which you can set up the sets and all of the lighting and the environments. You can do it all inside the computer. Now, we could do this with generative AI, but we don't have the control through generative AI that's required for these real enterprise use cases. When you are designing a product, you need to reproduce exactly what it is, what it looks like, how it behaves. There's actually a brand promise that you have with your product and you cannot violate that. So using simulation and digital twins for that purpose is key. And we're seeing one aspect of what you can do with the digital twins throughout all of the Adobe offerings. I'm really excited about this. I think it's going to be a game changer for all enterprises that produce physical things to have a workflow where they can carry their digital twins of everything they do throughout every part of their company. This part, the marketing part, the part where you communicate what your product is, actually hasn't taken advantage of the fact that the products are designed in the digital world to begin with. Whenever an automotive company wants to create a campaign around their new car, their new model, even though the CAD for that car exists, and it exists in many places throughout their company, when it comes to creating the advertisements for it, they have to go reconstruct the car essentially from scratch, because the tools and the pipelines and the workflows for this department is very different from the department that's designing it. We are now unifying it all together through common standards. The one we've all settled on and with Adobe, we formed a standards organization called the Alliance of Open USD. And so this is the basis by which we can define these digital twins and carry them all the way through. It's really exciting to see the adoption, and it's exciting to see Adobe leading the charge in this segment, in this industry.

John Furrier

>> Varun, give an example because the car thing's a good one because the CAD builds it and then marketing does its thing. One of the things I love about the keynotes is the graphics are amazing. So I can imagine that the creativity would be off the charts, because even at the end of the keynote, Jensen had that song play. That basically was the best summary of the event. This is where creativity is unlocked. What's your vision on that? And take us through an example of how this is playing out with people that are looking at this now you're working with.

Varun Parmar

>> Yeah. Yeah. So I think just building on what Rev just mentioned. So I think typically if you take a campaign, you're introducing a new car, it's a new product introduction workflow. And as a marketer, you have a bunch of questions you need to answer. So I think this is before you get into the marketing aspect of it, like what should be the colors you should have for this car? What should be the pricing of this car? Are there certain things that customers are describing the shape of the car, the features of the car that you should use in terms of the branding, in terms of the positioning, in terms of the differentiation relative to the other categories of products that are there? So that's upfront in terms of the simulation and the questions that you can answer. Once you have that car, then you want to make sure that it is shown in the environment, in the local geography where your customers are. So if you're going to target specific geographies where this car is going to appeal, how do you make sure that you can actually generate the background relative to the pixel perfect car, that Rev was mentioning about, and have all of these versions. So that someone can go to your website in the context of a product configurator, have all of these options available where dynamically you can actually show them relative to all of the options that they're selecting, what's the visual representation of that. And then, codify all of the rules. So one of the things we can do is codify all of the rules in terms of like, "Hey, this product comes in five versions in a specific country, but it's only three versions in another one. For those five versions, these are all the accessories that are available." So imagine all of these rules that you can automate and simplify and deliver as an end user experience as part of your overall application or your website that people can start to consume. And those kind of experiences really drive the conversion in sales, drive the engagement. And then you can start to share that. You can think about how a social element of that would also look like.

Rev Lebaredian

>> I'd like to add to that. In many cases, you want to start marketing or advertising your product before it actually exists in the physical world. You want to start the campaign before manufacturing has really scaled up there. You mentioned our keynote and all the impressive graphics. For the longest time, I think over 10 years, basically every physical product that NVIDIA creates, our chips, our systems, the PCB boards that you see inside Jensen's keynote, all of those have been digital twins that we simulate what they would look like. And in most cases, the physical one doesn't exist yet. We've taped out the chip and maybe it's at TSMC or Foxconn or Wistron is building the systems, but we need to communicate to our audience what this product is going to be, what it's going to look like. And we're certain that it's actually going to be exactly what we're showing, because the simulation of it, the simulation of how light is interacting with this object is exactly correct. And it's from the actual CAD that we use to go build the product so we don't have to worry about misinforming our audience, telling them it's going to look and behave a certain way when it's not.

John Furrier

>> A mock-up or something. This changes the workflow. So that was a great point. I want to just double click on that. Because what that does is, what you just said is, people can get a headstart. So okay, TCMC is taping out the chip, everyone's waiting around the old way, "Hey, where's the model? Where's the prototype?" You're getting ahead of the game. So this changes the workflow. So I guess my question for both of you, and Varun we'll start with you, is what's the scope of time savings or innovation or value that's unlocked from this? Because can you put the order of magnitude out there? Because old way would be, I can imagine, meetings, waiting around. Talk about this, because this is a significant change.

Varun Parmar

>> No, no. I think the savings in terms of the time to market for certain campaigns is dramatic. We've seen things it takes 28 to 30 days for a typical campaign to have assets available. Now you're talking about literally in a matter of minutes or hours, you're able to do it. And to Rev's point, if you can get that headstart relative to your competition, where you can start to talk about your product before others can, you can just start to imagine how you can influence your potential buyers in terms of leaning towards your product versus others. So this is really, really dramatic.

John Furrier

>> Just ballpark, months, weeks, because-

Varun Parmar

>> Yeah, like I said, a month can be minutes, right?

Rev Lebaredian

>> Yeah.

Varun Parmar

>> Yeah. Months to minutes is what... Yeah. Yeah.

John Furrier

>> If they can get a system from six months deployed to six days, these racks, we're talking about major value.

Rev Lebaredian

>> Yeah. And it takes months to actually manufacture them. So you can start before they're actually manufactured. So it saves you months at least in my case.

John Furrier

>> Yeah. You get feedback too. You get feedback potentially. All kinds of new data comes in, feed-

Rev Lebaredian

>> That's right....

John Furrier

>> that right back into the simulation.

Rev Lebaredian

>> In the case of the automotive industry, we were talking about cars. It can be a year or more from the point at which they know exactly what the car is supposed to be like, because the design is finalized and they're manufacturing it, til they have a final production car out. You can get a head start and be ready day zero.

John Furrier

>> So this is generative AI intelligence impacting content production basically. And changes the scale and the scope and the workflows for the clients. Okay. So I have to ask the question because I'm curious. Jensen was all over OpenClaw and said everyone should have an OpenClaw strategy. I won't ask you your OpenClaw strategy, but I want to tie in because my takeaway was if agents are the workers, OpenClaw is the operating system, and AI factories is the infrastructure, that's the new flywheel. So how does this segue into this? Because they'll be agentic, because again, some of the things going on in the workflows once they get identified now the workers are also agents. And then OpenClaw's a wildcard because there could probably be some cool things going on there, coding a new prototype or... What's your thoughts on this, Rev, on that paradigm?

Rev Lebaredian

>> Yeah. There's a lot of questions out there about what OpenClaw and this agents first idea means for traditional software. I think this is a question in the context of Adobe. For us, it's very clear. Agents are the equivalent of human workers. These digital agents are the equivalent of human knowledge workers. They can do many of the things that only humans could do before, but they do it by using tools. And they do it by having specialized expertise in a particular domain. And so, the companies that are best positioned to create the best agents to go be the best specialist workers in that area, are the ones that are creating the tools for the humans that are doing those things today. For the creators, for marketers, Adobe is the undisputed leader in this area. The knowledge of how these workflows work, what works inside this domain, is trapped inside the minds of all of the people at Adobe. So they're best positioned to go create the agents that will be specialists, that they could then rent out, contract out to those that need these specialists. And their agents are going to use the tools that they've been building for decades and decades. This model is going to work across all companies. It just doesn't make sense that any one company could build this super AI that has the proprietary knowledge that every other company has in every industry and to one big general model.

Varun Parmar

>> Yeah. I think just building on what Rev is mentioning, let's take a creative professional, and what do they do? It turns out that 60% of their time, three days out of five, they spend in just resizing, reframing images, videos, putting text on it, call to action and so on and so forth. And I think I would argue that nobody wants to go to design school to actually do resizing and reframing. So this is an example of a task where you can leverage an agent. This is a repetitive mundane task where you can say, "These are the tools you should be using," these are the models that have been certified by your specific brand and your organization. And then you can point an agent to it. And then what does it do for the creative is that it frees up a lot of time so that they can actually focus on the big C creative, the big idea that's going to drive your brand. So I think one of the challenges folks are talking about is, in this world of generative AI when everybody has access to all of these models out there, are we going to see a sea of sameness? It's like exactly everything looks the same-

John Furrier

>> It's all vanilla.

Varun Parmar

>> Exactly. And so, I think the need for creative professionals to come up with those ideas and big concepts is going to be bigger than ever.

John Furrier

>> You unleash them to be more creative.

Varun Parmar

>> Exactly. That's right.

John Furrier

>> Basically taking all the grunt work and all the toil and the heavy lifting away. In a way too, Adobe has a domain because that was a big theme of domain expertise in what you do, they have this AI technology. So I guess my question is, are you the domain's experts that are going to horizontally scale across the industry? Do you see Adobe taking your expertise and just putting it out there horizontally?

Varun Parmar

>> Yeah. Yeah. I think we have deep domain expertise, like you're saying, both on the creativity side as well as on the marketing side. Some of these workflows right from ideation to activation and with insights require deep understanding of how these enterprise systems work. What is the specific audiences that you're targeting? What are the objectives that you're trying to achieve, what has worked in the past? What will work? And coming up with all of those deep nuanced understanding of what's working and encapsulating that into the workflow. So I think our strategy is obviously to embrace agentic marketing and agentic creativity, and to use some of the platforms that NVIDIA announced and make it available to everyone.

John Furrier

>> All right. So here's a question. So the younger generation, the Gen Zs, I'll call them, maybe, I guess there's another one before that, but ease of use has always been a first principle of Adobe. How do you see the digital twins integrating into the workflows on the usability side? Because now you have, you mentioned resizing, that's a great example of taking a pain away. How about on the ease of use, creativity? I did a post on this show and I went to Nano Banana and they did a pretty good one. Jensen's in there. But it wasn't on brand. It was like a cartoon. But it did better than the last version of whatever I used. So it wasn't really great. But I'm not creative, I could have used some help there. So how do you see, if there was an easy to use tool, I'd imagine the digital twin and some of the things you guys are bringing will help.

Varun Parmar

>> Yeah. No, I think democratizing access to all of these amazing innovations that we are doing collectively with NVIDIA is the key design center for what we are doing. I think I've heard there's probably about 20 to 50,000 3D artists that are out there. So clearly the opportunity for us with the joint 3D digital twin solution is to make sure that everyone in your creative department and everyone in marketing has access to it. So then your question is, how will you make this happen? So I think one of the things that we've done in this solution that we've jointly developed, is that we have a conversational experience. This is an agentic experience where somebody can just go in and say, "Here's my USDZ file, here's my digital twin, here's the campaign I want to run for this specific channel." And then you just describe it in plain, simple English or whatever the language of choice is for you. And then off the system goes and we automatically stitch together all of the nodes and orchestrate in the right order. And then we do the brand checks, we make sure that it actually meets all of your internal guidelines, your governance guidelines and so on and so forth, and the workflow is ready. And then you just press run.

John Furrier

>> How do people use this? Take us through the deployment and consumption. I just love this idea of digital twins. I'm envisioning, can the digital twin edit videos for me and take all the highlights from NVIDIA GTC and roll it into a 24-minute program?

Rev Lebaredian

>> Well, it would be an agent that does that for you.

John Furrier

>> Okay.

Rev Lebaredian

>> But the agent could use the digital twin inside the campaign. So if you're making a campaign about your car, or if you're NVIDIA about the supercomputers we have in our racks, we would hand the digital twin over to the Adobe system. And the agent living inside the Adobe world, would operate on that digital twin to produce the image.

John Furrier

>> So you're saying this is a physical AI first digital twin. Agents come in and do the tasks.

Rev Lebaredian

>> That's right.

John Furrier

>> Okay. And how does companies get involved? Is it available? How do they use it? What's the onboarding? Give a plug for how to engage.

Varun Parmar

>> Yeah. No, 3D digital twins is available as public beta today. It's part of Adobe Firefly Creative Production for enterprise. So if you go to Firefly at Adobe.com, you can just sign up and then start using it.

John Furrier

>> Rev, every vertical is going to have Omniverse. That's the vision.

Rev Lebaredian

>> Every vertical that participates in the physical world will need to have a digital version of the physical things that they create. And in order to operate on those digital versions, you need simulation and you need a lot of specialized compute to do that well. And that's what we provide. NVIDIA's position in all of this is to provide the computing infrastructure, the very special computers to do the most impossible of all of the computing tasks. And simulation is one of those. And so, we're really thrilled to be partnering with Adobe to provide our technologies, our infrastructure at this layer, so that they can create the workflows, the tools, and now the agents that will supercharge the whole industry.

John Furrier

>> Yeah. That's great value. And the whole physical world still is not AI-fied. So more AI is coming to the physical world. So a lot more to talk about. Congratulations. Thanks for sharing.

Varun Parmar

>> Thank you so much. Yeah.

John Furrier

>> All right. I'm John Furrier with theCUBE here at GTC. Thanks for watching.