Virtual reality (VR)* training is now firmly past the hype cycle.

In 2022, Goldman Sachs and Morgan Stanley projected the metaverse opportunity at $8 trillion. Ironically, funding of VR, augmented reality (AR) and metaverse companies peaked at about $4 billion in 2021 in about 400 companies. As investors awoke to the market realities of limited scaled adoption, the number of successful raises fell exponentially every year since to 50 companies this year, collectively raising under a billion.

The funding trends raise important questions for learning and development (L&D) leaders: Where does VR outperform other learning approaches? What vendor risks should buyers consider more urgently? And what have we learned about the limitations of VR in practice?

It’s a timely opportunity to take stock as new breakthroughs follow the funding correction, driven by generative artificial intelligence (AI) authoring and advances in VR glasses. This article helps level-set on takeaways from the last hype cycle, the implications for L&D procurement and strategy today, and how to stay ahead of this next wave.

Shift 1: Where VR Has Underperformed and Where to Apply It Going Forward

Immersive learning via VR was expected to transform all types of learning environments and reach learners from high school students to trades workers and surgeons.

Where VR consistently delivers, the pattern is by now well-evidenced. Stanford’s DICE framework provides a practical filter: scenarios that are Dangerous, Impossible, Counterproductive or Expensive to replicate traditionally. These explain where VR has had traction and tangible impact.

In manufacturing, Boeing reported 90% improvement in first-time quality using VR, while Airbus used it to place seat markers six times faster with zero errors in one month. The common thread is gross-motor procedural work, spatial-procedural training and high-consequence safety, where the cost of failure is concrete and rehearsing in real life is dangerous or operationally impossible.

Digital twins are another promising application, particularly in manufacturing and critical infrastructure, where VR can create simulations that reflect the actual equipment, procedures and operating conditions employees encounter on the job.

On the other hand, VR’s comparative strength for learning social skills is under pressure. Generative AI has changed the substitution calculus. AI avatars and browser-based role-play now prepare employees for crucial conversations, are increasingly price competitive and offer more dynamic interactivity than the branching scripts most VR soft skills modules rely on. The bottom line on conversational realism is no longer obviously in VR’s favor.

Shift 2: Vendor Solvency and R&D Investment Risk Should Change Your Procurement Priorities

The direct risk for L&D is vendor solvency and research and development (R&D) capacity.

The underfunded middle tier of VR companies is selling, pivoting or shutting down, and the political cost of a vendor failure lands on the L&D leader who championed the program. Many VR companies are relatively small content studios with revenue concentrated among a few customers, increasing the risk that a major customer loss or technology shift could destabilize the business.

For L&D leaders, that means vendor evaluation can no longer stop at product quality. The company’s financial health and ability to continue investing in its technology should be part of the learning technology business case.

The buyer test: Prioritize vendor diligence to understand the company’s last fundraise, its profitability, customer concentration risk and revenue churn, and it’s plan to fund its roadmap. Consider tightening protective terms for your intellectual property and tech infrastructure. Smaller vendors will accept this diligence to close or renew a deal. Some of the larger VR companies may push back, and that itself will be a signal.

Shift 3: Generative AI Authoring Is Collapsing Customization Costs

AI-powered authoring tools now let users build VR content faster and increasingly through natural language rather than traditional code and scene-building workflows. This matters because these capabilities serve two very different users to different ends. For L&D teams without engineering resources, they make creating and editing existing scenarios possible: swap dialogue, adjust branching logic, localize content and refresh outdated procedures, all without filing an engineering ticket.

The buyer test: In-house custom content is now realistic for mid-sized companies, which reframes the procurement question in two ways. The first is pricing. Custom and refreshed modules should cost less in 2026 than they did 18 months ago, because of the productivity gain from AI authoring. If a vendor’s per-module pricing has not moved while development costs have collapsed, that is a renegotiation conversation worth having.

The second is what you are buying. The choice is no longer just “buy or build content,” but a three-way decision: what you buy off the shelf, what you build in-house and new hybrid arrangements. The shift worth planning for is from procuring complete packaged modules to procuring a foundational scenario your team and operational leaders can edit and extend themselves.

When evaluating these platforms, ask whether your team can reuse and adapt core scenarios rather than paying to rebuild each experience from scratch. Also test how much natural-language editing the platform actually supports and whether those changes preserve the quality of the simulation.

For L&D teams, this also changes the balance between customization and standardization. Instead of paying a vendor to build every variation, teams may be able to own more of the content lifecycle themselves, provided they have the right authoring capabilities and governance in place.

Shift 4: Glasses and BYOD Could Reshape Your Future Stack

AI-powered glasses could change how organizations deliver learning and performance support in the flow of work. Meta shipped more than 7 million AI glasses in 2025, more than triple the prior year, and holds over 80% of global smart glasses shipments. Meta’s full AR glasses are reported to be launched in 2027. Google announced AI-powered glasses with Gemini for 2026. Glasses could become a more important immersive form factor than headsets, particularly as employee-owned devices become more capable.

Glasses allow learning in the flow of work, just-in-time learning, rather than learning by stepping out of work for just-in-case learning. The pattern is already being adopted in frontline industrial contexts, where connected-worker and AI-powered guidance tools provide employees with information and support while they work. Glasses extend this from tablets and phones to hands-free, in-context overlays.

The buyer test: This is a structural decision about your learning stack, not a hardware refresh. Don’t standardize on a single device family for at least another 18 months. Prioritize platforms that can support multiple devices rather than locking your organization into a single hardware ecosystem. As employee-owned devices become part of the mix, L&D will need to work more closely with IT on device management and security. At the same time, tool support may increasingly need to be planned for and funded separately from training budgets, as these tools increasingly serve varied purposes.

Shift 5: Integrated Learning Journeys Are the Durability Test

Organizations don’t pay for completion. They pay for productive workers, and productivity is a function of the whole learning journey, not one module. When budgets tighten, that distinction can determine which programs get cut and which get renewed. The strongest VR programs are not enabling standalone experiences. They are integrated into competency frameworks, workflows, onboarding journeys and tied to operational key performance indicators (KPIs).

That creates a different test for VR vendors. Standout partners have deep industry context and or have capabilities to include other learning modalities.

Going deep means building enough expertise in an industry, such as HVAC, rail, nuclear, surgical or distribution operations, that the vendor already understands the context, competencies and impact measurement. That depth separates an embedded competence partner from a content supplier. This way, your subject matter experts spend less time educating the vendor about your business and more time validating the learning experience.

Going broad means expanding beyond VR to make it part of a larger learning journey that can include eLearning, AI role-play, performance support and assessment. Interplay, for example, has expanded its offering to include the full scope from eLearning to in-person learning for skilled trades training. However, building that breadth in-house is expensive, and most VR companies will not be able to do it effectively or finance the expansion. For L&D leaders, either path can result in a more contextually relevant partner than the standalone VR vendor of 2022.

Standalone VR loses. The question is whether your current vendor has a credible path to becoming part of a broader learning strategy, either through deeper expertise in your industry or through a wider learning offering, or whether it is still selling you a cool tool.

The buyer test: Look beyond the VR experience itself. How will it fit into your existing training architecture? What should learners do before and after the simulation? How will managers and operations reinforce what learners practice? And which business outcomes should it influence: time to competence, safety incidents, error rates or throughput? Answering those questions requires multistakeholder sponsorship across L&D, technology, operations and finance.

Conclusion

The first wave of VR training was sold on immersion and bought on hope. It completed its hype cycle in 2025, and the correction separated the real use cases from the oversold ones. The second wave is already forming, and it will be bought differently: on evidence, integration and outcomes that both L&D and chief financial officers can defend with data.

For L&D leaders, the opportunity is to use this reset to reassess where VR belongs in the learning portfolio. Start by identifying the learning problems where VR has a clear advantage over other approaches, particularly when employees need to practice physical procedures, work in high-risk environments or experience situations that are expensive or impossible to replicate. Then evaluate whether your current vendors can support those use cases without locking you into outdated technology or costly customization.

That means asking sharper questions during your next technology review: Does VR measurably outperform a lower-cost alternative? Is the vendor financially stable and able to invest in its purported roadmap? Can your team adapt and reuse content as your needs change? Will the content work across emerging devices? And how will the experience connect to the broader learning journey and the business outcomes you need to influence?

VR/AR remains a tool with genuine promise whose broad application is still years out. But its narrow application is here, defensible and getting stronger. Buy where it probably wins, keep your technology and content portable, protect your intellectual property and integrate VR into the learning journey rather than bolting it on beside it.

The goal is to know where VR earns its place in your learning strategy and to build the flexibility to adapt as the technology evolves.

*VR is used throughout as shorthand for extended reality, which includes virtual, augmented and mixed reality. Funding figures reflect this combined category.