Steven Carlini, Vice President of Innovation and Data Center, Energy Management Business Unit, Schneider Electric.
Many purists insist the only genuine metaverse is the one that requires you to immerse your “life-like” alternate self (not your cartoon-like avatar) in a virtual world where you can conduct life as your computer-generated self. This simulacrum was telegraphed in 1983’s movie War Games, where the distinction between reality and computer-generated, simulated reality becomes blurred. Is this a game, or is it real?
When people predict the metaverse is 10 years away, they’re talking about a metaverse that requires headsets and wearables to go through multiple generations of evolution, extensive modeling and programming of the alternate world and data centers and networks to be put in place to produce the capacity and low latency required.
Given the delay, it seems fashionable, perhaps, to put the idea of the metaverse away and wait for a decade. But not so fast! I contend the immersive, alternate, life-like world metaverse is only one of three types of metaverses. Each one has its own maturity and development path.
Three Metaverse Types
1. Personal Metaverse
An individual’s immersive experience using VR for playing games, attending concerts and museums, spending cryptocurrencies, etc.
2. Enterprise Metaverse
Immersive, mixed reality experience using AR/VR/XR at work for participating in meetings and operating or maintaining business environments.
3. Industrial Metaverse
Leveraging digital twin technology to replicate industrial or manufacturing operations, systems or processes.
The industrial metaverse is one I talk about because it’s at a maturity level that, while still early, is closer to its end vision than the personal or enterprise metaverse. Companies are leveraging digital twins today that will act as the foundation for tomorrow’s industrial metaverse. An AVEVA survey found 85% of industrial businesses plan to increase their digital investment, and the latest analysis by Emergen Research predicts the global digital twin market is expected to reach $106.26 billion by 2028 at a robust revenue CAGR of 54.7%.
How The Industrial Metaverse Adds Value
Design engineers use digital twin technology to run simulations and optimize the design around specified parameters like efficiency, output, reliability and even sustainability. Value doesn’t stop there, as after the design is constructed in the physical world, live data can be fed into the models and operations can be optimized further. This technology is only as good as the data. Having harmonized and verified live data is important as well as reliable data acquisition. Digital collection eliminates concerns about accuracy and depending on manual entry.
How The Industrial Metaverse Is Built Today
High-capacity compute resources are required as the industrial digital twin metaverse is a computationally heavy endeavor. Detailed, photorealistic, and physically accurate depictions or emulations of environments need to be created. Imagine the complexity of environments like a million-square-foot warehouse, a semiconductor manufacturing plant or an entire rail system. Serious computational power is necessary—servers with 40 GPUs and 16 TB of flash storage in IT racks in rows or PODs in purpose-built data centers.
The power density of these servers can far surpass 15-20 kW per rack, the limit of data center cooling. This is driving more interest in liquid cooling to the server and even the chip. Another building block is “transformers,” which are deep learning models designed for large datasets in tasks like natural language processing used in digital twin modeling.
Most companies won’t endeavor to construct their own data center to model complex digital twins. They’ll outsource to cloud platforms with GPU architecture and “transformers” integrated into their offerings, from companies like Amazon Web Services, Google Cloud, Microsoft Azure, Alibaba, Baidu and Tencent Cloud, for example.
The Future Value Of The Industrial Metaverse
Digital twins will likely expand into smaller instances and become ubiquitous and connected. In theory, every asset or process can be digitally “twinned”—connecting everything in a company. The benefit would be that almost every process could happen digitally, and countless numbers of digital twins could be interoperable. These mass simulations would drive enterprise-level benefits as opposed to system- or process-level benefits. Theoretically, these could extend to city-, country-, and even global-level benefits.
Building The Industrial Metaverse In The Future
Benefits at these levels will require a lot of computer processing, and not just in a single digital twin data center. They will need a fleet of connected local edge digital twin data centers.
Evolving from a single digital twin to wide digital twins involves focusing on a core system and then expanding it by integrating smaller ones. Take a utility grid, for example, where a power plant would be designed to power a large town. A digital twin of the plant and distribution system could be built and optimized to minimize waste and inefficiencies. Operating real-time data could be used to refine the operation.
Now we can introduce renewable supply in the form of a microgrid with wind power, solar power and utility-scale battery storage. A digital twin would be made for the new, local instance microgrid, which a local edge data center could operate and control. The local digital twin would integrate into the digital twin of the original grid, and so on. Eventually, the large town industrial twin would be processing big amounts of live, local data to make or automate decisions and improve performance across the entire system.
Maintaining these grids will be an enormous challenge, and digital twins using photogrammetry to generate 3-D representations of equipment used in generation, transmission and distribution equipment to evaluate wear and performance will be connected through 5G, 6G or even 7G in the future.
Organizations in all industries are under pressure to deliver improved, innovative operating models or decision-making processes. Now, companies can turn to digital twin technology for individual, industrial and manufacturing operations and systems. In the future, we can expect enterprise-wide and country-wide digital twin models with interoperable digital twins. Although it will take time for this vision to fully materialize, iterating, evolving and adding twins is a journey that’s well underway.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?