The Metaverse Hype Died. BMW and NVIDIA’s Factory Bet Didn’t

By NeuralWired Staff · Updated July 2026 · 11 min read

Microsoft shut down Mesh, its 3D collaboration app, on December 1, 2025. Meta froze new features on Horizon Worlds four months later, then reversed course within days. If you’re an IT leader who budgeted for “the metaverse” in 2023, both moves probably feel like a warning. Here’s the part the headlines are missing: while the consumer metaverse was collapsing, BMW quietly used NVIDIA’s Omniverse platform to cut its factory planning costs by up to 30%, and digital twin patent filings jumped 600% since 2017. The enterprise metaverse didn’t die. It just stopped pretending to be a social network.

What Actually Died: The Consumer Metaverse

Start with the timeline, because it matters. Microsoft didn’t just quietly deprecate a feature. It retired the standalone Mesh app, shut down mesh.cloud.microsoft, and pulled the avatar-based “Immersive spaces (3D)” view out of Teams entirely, according to Computerworld’s reporting. In its place: “immersive events in Teams,” a narrower tool built for scheduled gatherings like training sessions and product showcases, not everyday meetings.

That follows Microsoft’s earlier decision to stop making HoloLens 2, despite a reported $22 billion U.S. Army headset contract still on the books. Read those two decisions together and the message is blunt: general-purpose 3D avatar meetings never found a real audience inside the enterprise.

Meta’s retreat was messier. In March 2026 the company said it would stop adding new features to Horizon Worlds, effectively freezing its flagship consumer metaverse product, then partially reversed the decision days later, promising the platform would stay available on Quest. Reality Labs, Meta’s metaverse division, has lost more than $70 billion since 2021. The money didn’t vanish. It moved. Meta is directing over $115 billion toward AI infrastructure in 2026 alone, mostly data centers.

“They glommed on to the term ‘metaverse’ without really understanding the concept. Their efforts on their metaverse strategy seemed completely indifferent to what previous platforms had learned.” Wagner James Au, author of “Making a Metaverse That Matters,” via reporting republished by The Cool Down, March 2026

Here’s the thing worth saying plainly: you’ll see headline stats claiming Decentraland has as few as 38 daily active users, or citing wildly specific engagement collapses for Horizon Worlds. Treat those as reported, not confirmed. They trace back to secondary aggregators, not platform disclosures. The confirmed story, Mesh’s shutdown and Reality Labs’ documented losses, is damning enough on its own without inflating it further.

What Survived: BMW, NVIDIA, and the Digital Twin

While Mesh was winding down, BMW Group was scaling up. The automaker runs FactoryExplorer, a digital twin platform built on NVIDIA Omniverse, across its Debrecen plant and others. According to BMW’s own press materials and a joint case study with NVIDIA, the platform is projected to cut production planning costs by up to 30% and has already shrunk a collision-check process from four weeks down to three days.

That’s a name-brand company, publishing its own numbers, about a use case with nothing to do with avatars or virtual meetings. It’s industrial simulation: modeling a real factory floor before anyone moves a single robot arm.

Digital twins as a category are growing faster, and more quietly, than “metaverse” as a whole ever did. The global digital twin market sat at roughly $33.97 billion in 2026, with projections putting it near $384.79 billion by 2034, a 35.4% compound annual growth rate, per PatSnap’s industry analysis. More telling than any market-size projection: digital twin patent filings rose 600% between 2017 and 2025, with 2,451 filed in 2025 alone. Patents are harder to fabricate than survey numbers. That’s real engineering work, not marketing spend.

Consumer Metaverse vs. Enterprise Digital Twin: What Changed

Where the Real ROI Is: VR Training at UPS, Boeing, and Shell

Metrigy’s late-2024 survey of roughly 400 companies found that 16.5% planned to invest in VR or AR by the end of 2025. Irwin Lazar, the firm’s president and principal analyst, doesn’t dress that up as a boom.

“Use cases tend to be very targeted around training, product demonstrations, engineering and design, and customer engagement rather than for general purpose meetings. We expect to see slow continued growth, but I don’t see these kinds of virtual reality tools being more than a niche market going forward.” Irwin Lazar, President and Principal Analyst, Metrigy, quoted in Computerworld, December 2025

Niche isn’t the same as fake. It’s the difference between a platform strategy and a tool that solves one specific, expensive problem. UPS used VR to cut driver safety training time from eight hours to two, a 75% reduction, per its case study with ArborXR. Boeing’s published case study materials report a 75% cut in training time alongside first-attempt assembly accuracy improving from 50% to 90%. Shell reports a 30% reduction in VR-related training costs. Three different industries, three specific, named, attributable results.

The Uncomfortable Part: Pilots That Never Scale

Digital twins are the strongest enterprise metaverse use case on paper. They’re also the clearest example of a gap between piloting something and actually running it. Research cited by Gartner found that 75% of organizations that piloted digital twins struggled to scale past that initial pilot stage, even as 70% or more of manufacturers in aerospace, automotive, electronics, and energy are actively piloting or deploying the technology.

Remember Gartner’s own 2022 prediction that 25% of people would spend at least an hour a day in the metaverse by 2026? Search results as of mid-2026 show no corroborating usage data anywhere near that figure in enterprise contexts. That prediction should be read as unmet, not as a forecast still quietly ticking toward true.

“The current hardware suffers from limitations like a small field of view, heavy designs, motion sickness and poor graphics. The future of the metaverse lies in the hands of technologies in AI, 5G, edge computing, and display like microLEDs and better optics that are still to come before it can be fully realized.” Bob Gourley, CTO, OODA, quoted in Live Science, March 2026

Not everyone reads the pullback as failure. Futurist Mark van Rijmenam frames it as a maturing phase rather than an ending.

“It’s maturing into something more meaningful than the hype once promised. What felt like abandonment was actually a pivot beneath the surface. It’s being rebuilt with purpose, not PR, and with technology that’s actually ready for the spatial internet.” Mark van Rijmenam, futurist and author, quoted in Live Science, March 2026

Our read: both things are true at once. BMW’s numbers are real. So is the 75% pilot-to-scale failure rate. If digital twin platforms end up following Mesh’s pattern, real vendor investment, genuine flagship wins, but no generalization past a handful of marquee customers, the “enterprise metaverse survived” story could look premature within 18 to 24 months. Worth watching, not worth ignoring.

What to Actually Budget For in 2026 and 2027

If you’re the one signing off on next year’s spatial computing line item, here’s the practical shift. Stop buying “a metaverse strategy.” Start buying three separate, narrower things:

• Digital twin infrastructure for a single production line or facility, evaluated against a specific cost or downtime metric, not a company-wide platform rollout.
• VR training modules for one high-cost, high-risk skill (driver safety, assembly precision, hazardous-environment procedures), measured against your own current training time and error rates, not a vendor’s recycled 2020 case study.
• Scheduled immersive events, Microsoft’s own replacement category, for large training sessions or product showcases, not daily team meetings.

McKinsey’s research on predictive-maintenance digital twins found downtime reductions of 30% to 50% and maintenance cost cuts of 10% to 40%, a wide but credible range rather than one suspiciously precise headline number. That’s the pattern to look for in any vendor pitch you get this year: named companies, ranges instead of round numbers, and a use case narrow enough that you could measure it in one quarter.

Frequently Asked Questions

The Bottom Line

The metaverse, as a single procurement category, is over. What’s left is three separate, defensible technology bets: digital twins with documented industrial ROI, VR training for specific high-cost skills, and scheduled immersive events for large-scale gatherings. None of them need avatars. None of them need a headset strategy company-wide. Over the next six to eighteen months, watch three things: whether digital twin adoption starts closing that 75% pilot-to-scale gap, whether Meta’s AI-infrastructure pivot quietly starves Horizon Worlds for good, and whether mid-market manufacturers start showing up in NVIDIA and Siemens case studies alongside BMW. That last one is the real test of whether “enterprise metaverse” is a durable category or just a slower-motion version of the same hype cycle.

The Metaverse Hype Died. BMW and NVIDIA’s Factory Bet Didn’t

By NeuralWired Staff · Updated July 2026 · 11 min read

Microsoft shut down Mesh, its 3D collaboration app, on December 1, 2025. Meta froze new features on Horizon Worlds four months later, then reversed course within days. If you’re an IT leader who budgeted for “the metaverse” in 2023, both moves probably feel like a warning. Here’s the part the headlines are missing: while the consumer metaverse was collapsing, BMW quietly used NVIDIA’s Omniverse platform to cut its factory planning costs by up to 30%, and digital twin patent filings jumped 600% since 2017. The enterprise metaverse didn’t die. It just stopped pretending to be a social network.

What Actually Died: The Consumer Metaverse

Start with the timeline, because it matters. Microsoft didn’t just quietly deprecate a feature. It retired the standalone Mesh app, shut down mesh.cloud.microsoft, and pulled the avatar-based “Immersive spaces (3D)” view out of Teams entirely, according to Computerworld’s reporting. In its place: “immersive events in Teams,” a narrower tool built for scheduled gatherings like training sessions and product showcases, not everyday meetings.

That follows Microsoft’s earlier decision to stop making HoloLens 2, despite a reported $22 billion U.S. Army headset contract still on the books. Read those two decisions together and the message is blunt: general-purpose 3D avatar meetings never found a real audience inside the enterprise.

Meta’s retreat was messier. In March 2026 the company said it would stop adding new features to Horizon Worlds, effectively freezing its flagship consumer metaverse product, then partially reversed the decision days later, promising the platform would stay available on Quest. Reality Labs, Meta’s metaverse division, has lost more than $70 billion since 2021. The money didn’t vanish. It moved. Meta is directing over $115 billion toward AI infrastructure in 2026 alone, mostly data centers.

“They glommed on to the term ‘metaverse’ without really understanding the concept. Their efforts on their metaverse strategy seemed completely indifferent to what previous platforms had learned.” Wagner James Au, author of “Making a Metaverse That Matters,” via reporting republished by The Cool Down, March 2026

Here’s the thing worth saying plainly: you’ll see headline stats claiming Decentraland has as few as 38 daily active users, or citing wildly specific engagement collapses for Horizon Worlds. Treat those as reported, not confirmed. They trace back to secondary aggregators, not platform disclosures. The confirmed story, Mesh’s shutdown and Reality Labs’ documented losses, is damning enough on its own without inflating it further.

What Survived: BMW, NVIDIA, and the Digital Twin

While Mesh was winding down, BMW Group was scaling up. The automaker runs FactoryExplorer, a digital twin platform built on NVIDIA Omniverse, across its Debrecen plant and others. According to BMW’s own press materials and a joint case study with NVIDIA, the platform is projected to cut production planning costs by up to 30% and has already shrunk a collision-check process from four weeks down to three days.

That’s a name-brand company, publishing its own numbers, about a use case with nothing to do with avatars or virtual meetings. It’s industrial simulation: modeling a real factory floor before anyone moves a single robot arm.

Digital twins as a category are growing faster, and more quietly, than “metaverse” as a whole ever did. The global digital twin market sat at roughly $33.97 billion in 2026, with projections putting it near $384.79 billion by 2034, a 35.4% compound annual growth rate, per PatSnap’s industry analysis. More telling than any market-size projection: digital twin patent filings rose 600% between 2017 and 2025, with 2,451 filed in 2025 alone. Patents are harder to fabricate than survey numbers. That’s real engineering work, not marketing spend.

Consumer Metaverse vs. Enterprise Digital Twin: What Changed

Where the Real ROI Is: VR Training at UPS, Boeing, and Shell

Metrigy’s late-2024 survey of roughly 400 companies found that 16.5% planned to invest in VR or AR by the end of 2025. Irwin Lazar, the firm’s president and principal analyst, doesn’t dress that up as a boom.

“Use cases tend to be very targeted around training, product demonstrations, engineering and design, and customer engagement rather than for general purpose meetings. We expect to see slow continued growth, but I don’t see these kinds of virtual reality tools being more than a niche market going forward.” Irwin Lazar, President and Principal Analyst, Metrigy, quoted in Computerworld, December 2025

Niche isn’t the same as fake. It’s the difference between a platform strategy and a tool that solves one specific, expensive problem. UPS used VR to cut driver safety training time from eight hours to two, a 75% reduction, per its case study with ArborXR. Boeing’s published case study materials report a 75% cut in training time alongside first-attempt assembly accuracy improving from 50% to 90%. Shell reports a 30% reduction in VR-related training costs. Three different industries, three specific, named, attributable results.

The Uncomfortable Part: Pilots That Never Scale

Digital twins are the strongest enterprise metaverse use case on paper. They’re also the clearest example of a gap between piloting something and actually running it. Research cited by Gartner found that 75% of organizations that piloted digital twins struggled to scale past that initial pilot stage, even as 70% or more of manufacturers in aerospace, automotive, electronics, and energy are actively piloting or deploying the technology.

Remember Gartner’s own 2022 prediction that 25% of people would spend at least an hour a day in the metaverse by 2026? Search results as of mid-2026 show no corroborating usage data anywhere near that figure in enterprise contexts. That prediction should be read as unmet, not as a forecast still quietly ticking toward true.

“The current hardware suffers from limitations like a small field of view, heavy designs, motion sickness and poor graphics. The future of the metaverse lies in the hands of technologies in AI, 5G, edge computing, and display like microLEDs and better optics that are still to come before it can be fully realized.” Bob Gourley, CTO, OODA, quoted in Live Science, March 2026

Not everyone reads the pullback as failure. Futurist Mark van Rijmenam frames it as a maturing phase rather than an ending.

“It’s maturing into something more meaningful than the hype once promised. What felt like abandonment was actually a pivot beneath the surface. It’s being rebuilt with purpose, not PR, and with technology that’s actually ready for the spatial internet.” Mark van Rijmenam, futurist and author, quoted in Live Science, March 2026

Our read: both things are true at once. BMW’s numbers are real. So is the 75% pilot-to-scale failure rate. If digital twin platforms end up following Mesh’s pattern, real vendor investment, genuine flagship wins, but no generalization past a handful of marquee customers, the “enterprise metaverse survived” story could look premature within 18 to 24 months. Worth watching, not worth ignoring.

What to Actually Budget For in 2026 and 2027

If you’re the one signing off on next year’s spatial computing line item, here’s the practical shift. Stop buying “a metaverse strategy.” Start buying three separate, narrower things:

• Digital twin infrastructure for a single production line or facility, evaluated against a specific cost or downtime metric, not a company-wide platform rollout.
• VR training modules for one high-cost, high-risk skill (driver safety, assembly precision, hazardous-environment procedures), measured against your own current training time and error rates, not a vendor’s recycled 2020 case study.
• Scheduled immersive events, Microsoft’s own replacement category, for large training sessions or product showcases, not daily team meetings.

McKinsey’s research on predictive-maintenance digital twins found downtime reductions of 30% to 50% and maintenance cost cuts of 10% to 40%, a wide but credible range rather than one suspiciously precise headline number. That’s the pattern to look for in any vendor pitch you get this year: named companies, ranges instead of round numbers, and a use case narrow enough that you could measure it in one quarter.

Frequently Asked Questions

The Bottom Line

The metaverse, as a single procurement category, is over. What’s left is three separate, defensible technology bets: digital twins with documented industrial ROI, VR training for specific high-cost skills, and scheduled immersive events for large-scale gatherings. None of them need avatars. None of them need a headset strategy company-wide. Over the next six to eighteen months, watch three things: whether digital twin adoption starts closing that 75% pilot-to-scale gap, whether Meta’s AI-infrastructure pivot quietly starves Horizon Worlds for good, and whether mid-market manufacturers start showing up in NVIDIA and Siemens case studies alongside BMW. That last one is the real test of whether “enterprise metaverse” is a durable category or just a slower-motion version of the same hype cycle.