The industrial metaverse will allow you to build digital twins of whole environments! Ganesh Sivaraman, Director of Products, Omniverse Digital Twins at NVIDIA, joins Ryan Chacon on the IoT For All Podcast to discuss how NVIDIA is building the industrial metaverse with the NVIDIA Omniverse platform. They explore how NVIDIA and NVIDIA Omniverse relate to IoT, industrial metaverse trends, leading use cases for the industrial metaverse, why some IoT and edge device initiatives fail, IoT data challenges, metaverse adoption, and the future of the industrial metaverse.

About Ganesh

Ganesh Sivaraman has 20+ years of experience building high-value, high-impact products at Nokia, startups, and large enterprises solving customer problems. Prior to joining NVIDIA, he was responsible for building indoor positioning and asset tracking systems with embedded IoT in lighting for some of the major big box retailers. At NVIDIA, Ganesh is part of the Omniverse team, a platform for developing and operating metaverse applications. His focus includes Omniverse Digital Twin and IIoT. With Omniverse Digital Twin, customers build massive large-scale virtual worlds and connect systems and machines from the physical world to monitor, control, and simulate with Omniverse to drive automation and operational efficiency.

Interested in connecting with Ganesh? Reach out on LinkedIn!

About NVIDIA

Since its founding in 1993, NVIDIA has been a pioneer in accelerated computing. The company’s invention of the GPU in 1999 sparked the growth of the PC gaming market, redefined computer graphics, ignited the era of modern AI, and is fueling the creation of the metaverse. NVIDIA is now a full-stack computing company with data-center-scale offerings that are reshaping industry. NVIDIA Omniverse is a computing platform for building and operating metaverse applications.

Key Questions and Topics from this Episode:

(00:49) Introduction to Ganesh, NVIDIA, and Omniverse

(02:22) How NVIDIA and NVIDIA Omniverse relate to IoT

(08:09) Industrial metaverse trends

(10:20) What are the leading use cases for the industrial metaverse?

(12:13) Why do some IoT and edge device initiatives fail?

(15:38) IoT data challenges

(17:54) Metaverse adoption and future

(21:43) Can Omniverse be used before any physical deployment?

(24:00) Learn more and follow up

Transcript:

- [Ryan] Hello everyone and welcome to another episode of the IoT For All Podcast. I'm Ryan Chacon, and on today's episode, we are going to talk about industrial metaverse trends as well as why 70% of edge initiatives fail and what can be done to increase the likelihood of success on that front. With me today is Ganesh Sivaraman, the Director of Products and Omniverse Digital Twins at NVIDIA, a pioneer in accelerated computing, and now are a full stack computing company with data center scale offerings that are really focused on reshaping the industry. Prior to jumping into this conversation, please subscribe to the channel if you have not done so already. Hit that bell icon so you get the latest episodes as soon as they are out and give this video a thumbs up and we'd truly appreciate it. But other than that, onto the episode. Welcome Ganesh to the IoT For All Podcast. Thanks for being here this week.

- [Ganesh] Thank you so much, good to be here.

- [Ryan] Yeah, it's great to have you, excited about this conversation and I wanted to kick this off by having you give a quick introduction about yourself and the company to our audience.

- [Ganesh] That's great, so, thank you for having me, Ryan. So, my name is Ganesh Sivaraman, been in the tech industry for 20 plus years ranging from, like, mobile telecom background, all the way to doing my own startup. And then, most recently before joining NVIDIA was focused on building IoT infrastructures into the LED lighting market and rolling that out as like into location services and asset tracking services to some of the big box retailers. I recently joined NVIDIA just a little under a year, and my role here is primarily to be part of the Omniverse platform, which we'll go into a little bit more. And NVIDIA, as people know, has been around for, you know, since the nineties, and they've been very well known for everything that they do with gaming and GPUs in general.

- [Ryan] Right.

- [Ganesh] And over the years, NVIDIA has transformed itself to becoming, of course, a very AI centric company. It has lot of different products and tools that powers modern AI today. And then... And most recently, they have essentially kind of expanded that, you know, AI and GPU capability to start building virtual worlds and industrial metaverses and etcetera. So that's what Omniverse is all about, and I'm part of the Omniverse product team focused on the digital twin side and the IoT side.

- [Ryan] Fantastic, talk us through, just to give a little bit more context to our audience, more of, kind of the use case and applications as it relates to the IoT space that you all kind of play in, or how it kind of connects to the IoT space, the things you're doing directly in that realm, so our audience has some context. 'Cause as you mentioned, NVIDIA is very well known for a lot of things, but I'm not sure our audience as familiar with what you're doing as it relates to the IoT space. So maybe talking a little bit more in around deployments or applications or kind of things you're bringing kind of to that industry.

- [Ganesh] So, I think we need to kind of break that into two parts as NVIDIA and then also from Omniverse lens perspective. So, let me kind of take the NVIDIA perspective of things and how that relates to IoT and then I'll touch upon that how Omniverse and IoT kind of fit together as well. So, if you look at NVIDIA in general, because of all the compute capabilities that it has been bringing to the table for the last several years, one of the key areas that NVIDIA has been actively playing and continues to play is in the whole edge device side of things when it comes to IoT. And as we all know from the IoT landscape, edge gateways and edge devices play a very important role in bridging the OT, IT gap, the operational technologies and the information technologies. And with great advancements that's happening on the AI side, a lot of these inferences and models that are required to make certain decisions are now happening more and more at the edge. So, NVIDIA has been actually been very active in the IoT space through the tools and the chip sets that they've been providing to make that edge compute possible for several different use cases. Now, if you look at from an Omniverse perspective, which is the other side of things that NVIDIA is involved in, is that Omniverse is a platform, primarily, and it is mainly used for building large scale virtual worlds.

- [Ryan] Okay.

- [Ganesh] When you're talking about virtual worlds, it's basically nothing but Metaverse and industrial metaverse, or whatever different applications of virtual worlds that the customers want to build. They come to Omniverse to build that virtual world. One of the things that Omniverse has started to resonate well with a lot of customers is how do I create these digital twin, which is, you know, an exact identical replica of my physical world. And the physical world could be anything. You can imagine as a factory, as a warehouse, as a data center. That is either yet to be designed and deployed or that is already deployed, so you can think about it in two different ways. So, Omniverse actually allows for customers to create those virtual worlds. Now, when you're creating a virtual world, the other side of the virtual world is the physical world, which means that there are equipments, machines, systems, whatever that is operating has to essentially be plugged into the virtual world. That's where IoT comes in. So, how do you bring in those physical, continuous, real world data streams, IoT data streams from the physical world and connect that to the virtual world, so both of them, the virtual and the physical are living in perfect harmony. That's how the IoT plays from an perspective on both sides of things.

- [Ryan] Okay, fantastic. Yeah, I was actually, I had somebody on recently who was talking a lot more about, obviously, digital twins tied into this, but we were talking about kind of simulation and being able to run simulations early on in kind of setting up, kind of that digital world presence prior to any type of physical elements coming into it. But it sounds like this is actually bringing in physical elements as well on top of it from a data stream standpoint. Is that kind of accurate?

- [Ganesh] That's correct, but you can... So, that actually is a good way to break down the reasons why we need IoT data streams into any applications that is related to metaverse or creating this industrial metaverse. So you could think about it in one of two different ways. One is I'm actually interested in simulating the world, exactly as you said. So, I have not necessarily built anything yet, or I've already built something, but I want to actually run scenarios or different scenarios or simulate them before I make any changes or optimizations to my processes or to my factory or to my warehouse. So, things are already running, and producing whatever it's producing, but we want to simulate and see what more further efficiencies or gain we can get without actually making any changes to the physical factory or physical warehouse. Because you have to stop it, there's a lot of stoppage loss, and so you cannot really make those decisions that easily when you are dealing with a physical world, so that's where the simulation comes in. So, when you're doing simulations, one of the things that you want to know the baseline, what is happening to my physical world as is. So that's where the IoT comes in. But then, the other aspect is just continuous monitoring and control. So, it's not just simulation, but with people actually building their 3D worlds using Omniverse. They're trying to actually bring a simple single way to have a single pane of glass view, so I want to overlay the 2D data sets. The 2D could be data sets coming from an IoT sensors or IoT systems, but also I want to overlay whatever enterprise data and applications and a spec sheets, data sheets, whatever that is of an equipment, and I want to interact with that. So, you bring in the non 3D and the 3D all in a single pane of glass and create that context for continuous monitoring and control as well. So it's simulation and control, monitoring and control.

- [Ryan] Fantastic, okay, that makes a lot of sense. And now, if we're talking about kind of, you mentioned the metaverse and kind of the industrial metaverse, what are some of the trends that you're seeing when it comes to the industrial metaverse? If we're talking about, let's say, a macro level kind of thing, what are some of those trends that you're seeing as it becomes more and more of a, not just a hot, but also a relevant topic as it connects to IoT as well?

- [Ganesh] Yeah, so, I think, the term metaverse kind of became really popular with, you know, Facebook changing its name to Meta. And a tremendous amount of drive that came in. But that's more focused on the consumer side of things. What in media and what we are doing in Omniverse is more towards the industrial, but of course Omniverse being a platform it can be used for other applications as well. But I think one of the biggest trend and the drive that people are seeing is how automation efficiencies, operational efficiencies and automations can be gained by building virtual worlds of whatever processes that people are trying to imagine. So, if there's a brand new factory that someone is considering, especially with the EV phenomenon that's going around, a lot of battery companies are coming up. So, if you're building a massive, say, a battery factory, you probably want to do that first in a metaverse. Go through that whole design iteration testing process. And then, once you have that blueprint saying that, yes, everything, you know, people, process, factory, equipments, everything is as expected, then you go ahead and build it. So, the room for error, room for last minute surprises dramatically reduces. So essentially, you are gaining operational efficiencies and then also leveraging massive amounts of automations that's going to come through AI. So Omniverse also enables, you know, all of that when you're building virtual roles and taking into account all these latest and the greatest technology stack that we are providing. And we are still just in the first innings of all of this and we are really seeing a tremendous amount of interest from some of our major customers.

- [Ryan] And what is it... When we talk about interest from your customers, what is it that they're most looking to kind of have you help them do? Obviously, the efficiency side, and there's different kind of elements there, but just like from an application side, where is the big demand right now in these early innings as you say, and where do you kind of see this kind of growing into?

- [Ganesh] So, I think it's the whole notion of building like a virtual factory or a virtual facility. So, you want to basically design all of that and, you know, bring maximum amount of efficiencies in that planning phase. And then one of the things that Omniverse does really well is allowing for collaborations between all facets of your building process. You know, you have people working in different tools and applications and Omniverse actually brings all of that through USD, independent or whatever applications you're using, whatever formats you're using. It brings collaboration capabilities as well. So, when you take all that into account, you know, planning phases and efficiencies plays a key role for customers to start adopting Omniverse.

- [Ryan] Yeah, now it sounds like a fantastic kind of just evolution of everything we've been talking about over the last... At least personally with me over the last number of years when we've talked about digital twins more on kind of replicating the physical devices, but now we're talking about the actual full environment and being able to bring that to life to help with decision making, especially earlier on in the deployment process, which I know is one of the big challenges in this space is that a lot of IoT initiatives, really, they fail. They don't, you know, they fail once they kind of get deployed in the pilot stage very often. And there's a lot of work that could be done earlier on, or done maybe throughout, which it sounds like the Omniverse and as, you know, we're talking about the metaverse as well, could help prevent or at least make better decisions around. From your perspective, what are the reasons that a lot of those initiatives are failing? 'Cause obviously, I'm sure a lot of that drove Omniverse being created.

- [Ganesh] So, when you're talking about failing, I think, you are more talking about in general IoT deployments, or are you talking about specifically metaverse or digital twin initiatives? Which one are you referring to?

- [Ryan] Yeah, more so, the IoT initiatives and deployments. So, you know, edge initiatives, things that are being deployed physically.

- [Ganesh] Correct.

- [Ryan] There's a reason that they are failing and there's this virtual side of it, you know, the metaverse, which, you know, I was talking the Omniverse. The simulation... All these different things that are being created and launched now, I think help avoid or kind of increase likelihood of success, I guess I should say. So, I also wanted to get your thoughts on kind of why you think a lot of these initiatives are failing and how companies really can be approaching it to increase the likelihood that they don't fail.

- [Ganesh] Yeah, so, when we talk about IoT initiatives, in general, failing, it's essentially the fragmented landscape that IoT is. It's basically, a multi-vendor, multi-protocol, multisystem initiatives typically that the customers and customers end up buying. And when we talk about any you modern applications or the next generations of applications, like Omniverse, or anything that the customers want to embrace the edge device and whatever they are doing at the shop floor level or the local site level becomes very, very important and critical. And if you have a massively large global company that has multisite that is spread across different regions and countries and has multi-vendor, multi-protocol, you just made that problem much, much bigger. So, how do you solve that? And that's one of the reasons why a lot of customers, when they start with an IoT initiative, they start with a very small local site, one vendor, one system, and they try to see that whether it works, and in most cases it does. But when they try to go from there to a global scale and say that let's implement this, that's where they find a lot of challenges. And the reason why those challenges are because IoT in general is very, very high data intensive. It's like small, it's high frequency, low volume, I wouldn't say low volume, but it's basically, you know, it's not a massively, we are not talking about streaming videos. It's a small amounts of data, but it's at a high frequency. But you're bringing that data from so many different systems, so you need to have a good way to aggregate and that's where the edge devices that NVIDIA is building is trying to solve for those problems where you can bring data from wherever you want in whatever protocols or however you're aggregating it. But then, we are essentially making sure that some of these, you know, fragmented approach that typically happens today on the edge devices can be addressed. And we are helping that by directly making those edge devices connect over USD and essentially bring that data sets from the edge into Omniverse directly natively into USD. So, these are some things that we are trying to address as future items, which will make creation of digital twin a lot more easy for our customers.

- [Ryan] Yeah, that makes a lot of sense. I was gonna ask you kind of as it relates to a lot of these companies now have access to a lot more data than they had before. You know, they adopt IoT in some way, or just in general, the ability to bring in new data, but they're struggling to drive intelligence from it, insights from it, but there's lots of different sources. There's, you know, and there's a difficulty in aggregating that content, and it sounds like you all have a lot, or you're taking that as a pretty big initiative to help companies solve. When you talk with companies that work with them, what other challenges, maybe even as it relates to what we were just talking about, do you see a lot of companies have with that influx or access to new data that they maybe did not have before?

- [Ganesh] The most of the datas are typically siloed 'cause, you know, remember we are talking to an industry that has been around for decades. And some of these companies have already started to embark on digital transformation, maybe a decade or a decade and a half ago. And that's how they started to embrace IoT and wanted to bring all of the smartness and connected systems so that they can bridge that OT, IT. Most of the times all of these systems were very local, very site level centric, and they were not all connected and talking at a global level. And IoT started to change all of that. And it started to kind of break down some of those silos and the whole fragmented multi-vendor, multi-protocol. Some of that was actually breaking down, but with a goal that a company has that everything actually will be first built in the virtual world and then we'll be built in the physical world. These new data sets that they are thinking about and imagining is going to just unlock tremendous amount of potentials and that ultimately drives automation and operation efficiencies. So, it definitely is the whole move to an industrial metaverse is essentially helping customers tap into their existing investments that they've done for IoT initiatives as part of the digital transformation. And it actually tries to build that ramp from their IoT initiatives into building digital twins and Omniverse. So, it's a good thing because they don't have to reinvent the wheel, they can just leverage their existing investments and infrastructure.

- [Ryan] One thing I wanted to ask you just kind of came up as you were answering that question is, what has been the reception, or what has this reception been like for the companies that you talked to in more in the industrial sector? And the reason I ask is, 'cause a lot of the companies I've spoken with recently, have mentioned that the industrial sector companies that work in manufacturing and factories has kind of been, sometimes they run into big challenges in getting them to adopt new technologies. And I'm just curious when we're talking about Omniverse and things like, you know, talking about the metaverse at a high level, that's a new kind of world for them to even kind of wrap their head around. So, how are you kind of seeing the companies that you all work with or talk to versus like, adopt this new technology and kind of where do you see this going as we kind of expand out through the rest of this year, into the future? Because this is gonna only become more and more of a widely adopted thing and something that provides more and more value to the industry.

- [Ganesh] So, I think every industry, whether it's data centers, whether it is logistics and transportations, whether it is even smart city, whether it's even telecom, whether it's even space or deep science, and we just talked about industrial, but I'm just trying to expand that. So, every vertical in industry that you can imagine is essentially going to start building things first in the virtual world. And it is something that most of the industry, if not all, have been already doing that with their 3D modeling, with CAD drawings and other mechanical drawings and designs. They have been already doing that. That's the way where architects build a building and how mechanical engineers actually build and design any kind of systems and equipments. That has been already happening, so that's not a consumer behavioral change, it's a must have. That's the only way where things get actually built. What industrial metaverse - What metaverse in general, not industrial metaverse, metaverse in general, is actually helping, is bringing all of that skillsets through different ways and applications and standards that already exist into Omniverse. And they don't have to make any changes through that. We bring all of that content in whatever way they're designing today and then creating it as a single platform to collaborate. And now, when you are actually designing everything off Omniverse as a baseline, you can add several different other capabilities that Omniverse brings natively. For example, you can add a physics engine, you can add materials. So, I can actually know that this is plastic versus of metal. And of course, the physics behind a plastic and a metal different, changes dramatically. So, you can do a whole lot of stuff when you are bringing capabilities to build virtual worlds that are photorealistic, full fitter visualization, and identical replica of the physical world. And then, you start applying AI techniques, and you do simulations, which is exactly how a real world will behave. You're actually assimilating it. It's not animation, it's simulation. You're simulating it in the virtual world. So, you basically build a blueprint even before you actually go break ground. And that actually is extremely exciting for a lot of customers. And that's why they've tried in, you know, tested different things that Omniverse can provide and they believe that there is a tremendous amount of potential to go that way. So, that's where I think, you know, in the coming months, and I know at GTC there is a lot of different things that's going to be shared. And in the coming months, in the coming years, you will see metaverse in general and how Omniverse actually participating in that.

- [Ryan] And one of the things you mentioned, one of the big values initially was talking about how you can kind of plug in that real world data into the Omniverse solution. Is that required or can this... It sounds like, can this all be done kind of prior to any actually physical deployment, but you can kind of set up the environment to replicate the real world prior to potentially any new technology or deployments going on. Is that the case?

- [Ganesh] It could be seen both ways. So, you could see a brownfield as well as a greenfield. So, when I talk about, you know, designing everything from scratch first in the virtual world and a customer is basically thinking about building a brand new factory or a brand new warehouse or something that doesn't exist yet. That's your greenfield side of the business. But then, you can also do the same thing for your brownfield, your existing business that is running and operating today. You can bring that and build that as well into Omniverse and creating an industrial metaverse. It's for both those worlds. But the key here is that you actually have a replica of whatever that physical world is in the metaverse. And now, you're using Omniverse to simulate it. That's why I said it's, you can think about it, monitor and control and simulation. And monitor and control, falls into things that you have here and now. And I want to know what's happening live. And you're bringing all of that data through IoT into Omniverse.

- [Ryan] Great, no this has been... This is super exciting stuff to kind of learn more about. We've had a couple guests on talk about the metaverse since applications in the IoT space, but never to this extent to really break down exactly-

- [Ganesh] I'm glad.

- [Ryan] Yeah, what a product can actually do. We talk about digital twins all the time, but this is kind of taking this all to a whole new level on what you can actually create virtually prior to or throughout the deployment of any kind of new correct solution or initiative, right? And it's going to be something that I imagine saves lots of time, energy, mistakes being made, costs, things like that. Being able to handle this all through software as opposed to the physical deployment and the hardware, which is obviously where a lot of the costs come in. And the mistakes are often amplified because of what has to go into fixing them, so-

- [Ganesh] That's correct.

- [Ryan] So, yeah, I really appreciate you taking the time to kind of talk about this a lot. This is awesome. For our audience who wants to learn more about this topic, learn more about the Omniverse, learn more about what you all are doing, just maybe follow up with anything connected to this conversation, what's the best way they can do that?

- [Ganesh] So, I think, you know, you can of course go to our NVIDIA's Omniverse website and a lot of wealth of information is already available, tremendous amount of easy getting started applications. You can even download Omniverse to your Windows workstation and get started that way. And then of course, I highly encourage our GTC that's coming up this month and you can essentially tune in. There are several different tracks, both technical and non-technical, and some major announcements are being planned, which is gonna happen next week, so stay tuned for that as well.

- [Ryan] Fantastic, well, really appreciate your time. Again, thank you so much for coming on and talking to our audience about this. Very exciting stuff to kind of follow along and see how it's gonna continue to