This article is part of a VB special issue. Read the full series here: The metaverse - How close are we?

Defining the “metaverse” is a difficult task, but one commonly accepted definition is a digital space populated by representations of people, places, and things. Through a combination of technologies including virtual reality (VR), augmented reality (AR), and AI, the metaverse that some futurists envision is an extension of the real world — albeit without the physical trappings.

Companies like Rockstar and Roblox have pitched the metaverse as the ideal platform for gaming, but there’s no limit to the potential applications in the enterprise. Because the underpinning technologies are adaptable in nature, they can be repurposed for other tasks — for example, enabling a manufacturer to predict how machinery might behave once it’s installed on the production line.

According to TrendForce, the metaverse could propel global “smart” manufacturing revenue to $540 billion by 2025. A recent Altair survey found that 86% of industrial manufacturing companies already use some form of simulation to reduce development times and risk, achieve higher operational efficiency, and increase machine performance and product quality.

Simulation of the real world

The metaverse could be a fit for manufacturing, in particular, where considerations like labor costs, inventory, and the speed of production dictate decision-making. As IoT For All’s Roshan Srinivasan writes, manufacturers generally aim to optimize production based on demand forecasts using signals like seasonality and market size. The main pain points are long lead times, quality control concerns, and risks in production design (e.g., mistakes in facility layout) — all of which could lead to faulty products and longer wait times for manufacturing.

“[T]he goal [is to simulate] experiences in the virtual world before moving into the physical world,” Danny Lange, senior VP of AI at Unity, told VentureBeat via email. “For example, enterprises can use the metaverse to explore and improve supply chains or production lines, so that everything runs smoothly going into real production. I think the ‘industrial metaverse’ is where simulation and reality meet. Data created in simulation leads to design and implementation improvements in the real world. Data created in the real world flows back into the simulation, thus creating a virtuous feedback loop.”

Some experts predict that the industrial metaverse, as it were, will require internet of things (IoT) sensors to collect data from real-world devices, enabling companies to accurately model them in the digital world. Duality, a startup developing what it describes as an “enterprise metaverse,” asserts that this is one of two distinct aspects of simulation. The other is a system driven by a simulation model that accurately mimics a physical system to understand its behavior under different circumstances.

Ericsson is testing 5G network signals in Omniverse.

Above: Ericsson is testing 5G network signals in Nvidia’s Omniverse platform.

Image Credit: Nvidia

Metaverse technologies could help to prevent breakdowns on oil rigs, for example, by piping in data from the physical world and synchronizing it with an auditable copy. Companies in industry and manufacturing, as well as energy and telecommunications, have already begun to embrace the metaverse for these purposes. Partnering with Talumis, a simulation platform provider, Unilever created a model of a production line to identify bottlenecks that could be minimized by adjusting the product and package type. Siemens Energy and Nvidia, leveraging real-time data like temperature and pressure, created a simulation to model the corrosive effects of heat, water, and other conditions on metal over time to fine-tune maintenance needs.

Elsewhere, Ericson teamed up with Nvidia to map cities with buildings and landmarks, so it can understand how wireless signals bounce around and reach customers. BMW built a copy of a car factory with cars, robots, and digital humans to plan the installation of manufacturing systems. And Mars — the candy, pet care, and food company — is using Microsoft’s metaverse stack to digitize its supply chain and optimize production through simulations.

In the future, Mars says that it hopes to use these the technologies to identify the optimal way to create products by creating simulations that take into account the climate and other situational conditions.

“Companies can use [metaverse] technology to optimize production, saving cost and time,” startup Blueprint shared in a blog post. “They can simulate regional demand patterns to identify new locations or simulate floor plans to determine the most efficient use of space. They can also make a digital copy of equipment to predict when it will break and what incremental changes could mean for production or staffing.”

Future metaverse use cases

Looking ahead, the metaverse could be adapted to use cases in retail and other industries currently less reliant on simulation. For example, a restaurant could create a store virtually by inputting third-party data, including demographics in that area, and running simulations based on minor tweaks to the parking lot and store design to prevent traffic issues. Or a store could apply metaverse technologies to pilot new concepts virtually, capturing data on how people interact with them.

As companies increasingly adopt automation and robotics technologies, the metaverse could play a role here, too. Unity, the platform for creating games and other 3D content, in November announced the launch of new products to improve modeling, testing, and training complex systems through AI. One of the offerings — Unity Simulation Pro — enables multiple GPUs to render simulations with million-square-foot warehouses, dozens of robots, and hundreds of sensors in realistic virtual worlds.

This is not a picture. It's a simulation in Unreal Engine 5.

Above: This is not a picture. It’s a simulation in Epic’s Unreal Engine 5.

Image Credit: Epic Games

One possible wrinkle is the computational requirements required for accurate simulations at scale. Intel’s Raja Koduri, VP and head of the accelerated computing systems group, believes that a 1,000 times increase in power is needed over the current collective computing capacity to realize the promise of the metaverse.

Lange agrees that the metaverse will be an “extremely data-hungry technology,” but offers a solution in synthetic datasets. A recent research paper coauthored by Unity engineers proposes a “synthetic data generator” for human avatars called PeopleSansPeople. Containing “simulation-ready” 3D human models, virtual lighting, and a camera system, it generates labels that can be used to train an AI system to recognize different poses, motion, and more.

“Synthetic data is an intrinsic part of simulations in the metaverse where a wide range of scenarios can safely be played out before impacting the physical world,” Lange said. “Synthetic data comes at a fraction of the cost of real-world hand-labeled data.”

Not everyone is convinced that synthetic data holds the key to a more sustainable metaverse. But in conjunction with algorithms and software improvements, it could make the industrial metaverse more attainable than it is today.

“So algorithms, architectures, neural net algorithms — some of this stuff has to play a role in increasing [the metaverse’s] efficiency,” Koduri told Quartz in an interview. “Algorithms will have to play a big role in getting the [necessary one-thousand times compute power]. We can do brute force. We can put more compute, like these Bitcoin farms. [B]ut that’s not energy efficient… [s]o it has to be that balance where we do energy-efficient compute and hardware, some better algorithms, better architectures.”


Read more from this VB Special Report:

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.

Author
Topics