Why Metaverse Failed, AI Succeeds

“Metaverse” just never seemed to catch on, and the issue is “why?” While it is always possible to argue that the concept was simply “ahead of its time,” perhaps there were other issues as well. 

For starters, metaverse was a push toward more immersive, higher-fidelity digital environments. But as with other proposed advancements in digital media, it did not solve a broad, urgent problem for most users.

Television or movies presented in “three dimensions” also arguably are “more immersive” or “realistic,” but that never is enough to create demand. 

But some thought something more was at stake. The metaverse, some thought, would become the next computing platform, the successor to the mobile internet. 

But the technology was not compelling enough; the friction was too high; the value way too limited. The ecosystem, content base and network effects were not there. 

Compare metaverse to artificial intelligence, a general-purpose capability that can attach to almost any cognitive workflow and business process.

Use case	Metaverse value	AI value	Relative usefulness
Virtual meetings / collaboration	Moderate: better spatial presence, but often not better than video calls.	High: summarizes, transcribes, drafts follow-ups, and speeds decisions.	AI
Employee training / simulation	High in physical or risky environments, where immersion helps.	High: creates training content, coaches, quizzes, and personalizes learning.	Tie, slightly AI
Customer support	Low to moderate: immersive support is niche.	Very high: chatbots, agent assist, routing, and automated resolution.	AI
Sales / product demos	Moderate: strong for 3D visualization and experiential demos.	High: personalizes outreach, generates content, and qualifies leads.	AI
Entertainment / gaming	High: this is one of metaverse’s best-fit domains. dreamsoft4u	High: generates content, NPC behavior, personalization, and moderation.	Tie
Education	Moderate to high for immersive labs, historical reconstruction, or anatomy. dreamsoft4u	High: tutoring, summarization, feedback, and adaptive learning.	AI
Healthcare / therapy	Moderate: useful for exposure therapy, rehab, and visualization.	High: triage, documentation, diagnostics support, and patient messaging.	AI
Remote field assistance	Moderate: useful when a remote expert needs the user’s visual context. techtarget	High: guides workers, interprets data, and generates instructions.	AI
Marketing / brand experiences	Moderate: immersive campaigns can be memorable but narrow. dreamsoft4u	Very high: segmentation, content generation, ad optimization, and personalization.	AI
Commerce / shopping	Moderate: 3D storefronts help some categories like real estate or furniture.	Very high: recommendation, search, pricing, and conversational shopping.	AI
Design / visualization	High: strong when spatial understanding matters.	High: concept generation, variants, and analysis, though not always spatial.	Tie
Knowledge work / office tasks	Low to moderate: metaverse mainly changes the interface.	Very high: directly improves writing, coding, analysis, planning, and review.	AI

Where metaverse mainly extends what digital media can look and feel like, AI extends what software can do. 

A virtual world can make communication, entertainment, and simulation more realistic, but it still stays within the realm of mediated experience. 

AI, by contrast, is increasingly useful wherever humans are reasoning, drafting, classifying, predicting, summarizing, planning or making decisions.

AI adds value even when the interface stays ordinary, because it upgrades the work itself rather than just the container around the work.

Metaverse even if all the other issues had not been present) is “only” the next evolution of realism in electronic media, while AI is the next evolution of cognition. 

Perhaps virtual reality will someday deepen immersion for users. 

But AI, in principle, can affect almost every cognitive task because it can assist with language, judgment, memory, analysis, and creativity in almost every domain. 

And adoption barriers are quite low: people can use it right now, with low friction and no new hardware requirements. 

If metaverse was about the realism of the interface, AI  is about the augmentation of cognition itself. 

I don’t recall anybody arguing that metaverse was a general-purpose technology on the scale of electricity, in terms of impact,for example. It’s pretty hard to find anybody arguing AI is less than that.

J Curve and Solow Productivity Paradox are at Work with AI

Investors are going to keep challenging firms to show evidence their heavy artificial intelligence investments really are boosting productivity.

That is going to continue being a tough challenge, as history suggests the real output gains will take some time to develop.

So AI "productivity," or the "lack of quantifiable gains," are currently the most significant contemporary case of the Solow productivity paradox. 

In 1987, Nobel laureate Robert Solow famously remarked, "You can see the computer age everywhere but in the productivity statistics."

Recent research suggests productivity might actually decline for a time as firms deploy AI. 

The reason is the J curve. 

“We find causal evidence of J-curve-shaped returns, where short-term performance losses precede longer-term gains,” say economists Kristina McElheran; Mu-Jeung Yang; Zachary Kroff and Erik Brynjolfsson. “Consistent with costly adjustment taking place within core production processes, industrial AI use increases work-in-progress inventory, investment in industrial robots, and labor shedding,

while harming productivity and profitability in the short run.”

In other words, it takes time for enterprises to retool their business processes for the new technologies. And the more profound the innovations, perhaps the longer it takes to integrate those tools. 

Also, much of the reported AI adoption is horizontal rather than vertical; personal rather than systematic. In other words, individuals might be using chatbots, but workflows have yet to be transformed. 

So “personal productivity” has not yet been matched by an applied transformation of key work processes. And personal productivity gains are hard to measure, in terms of impact on firm performance. 

Agentic AI should help, as they can affect complex business processes. 

source: Forbe

Many have noted that  U.S. labor productivity significantly slowed in the 1970s and 1980s, despite rapid information technology investment. 

Then starting in the mid 1990s a decade of faster growth returned arguably because business process re-engineering had taken place.

A similar productivity paradox surrounds AI. As explained by economists Erik Brynjolfsson, Daniel Rock, and Chad Syverson in a 2017 working paper, AI and the Modern Productivity Paradox,” the paradox is primarily due to the time lag between technology advances and their impact on the economy. 

While technologies may advance rapidly, humans and our institutions change slowly. 

Moreover, the more transformative the technologies, the longer it takes for them to be embraced by companies and industries across the economy.

Translating technological advances into productivity gains requires major transformations, and therefore time.

Today, we see a "Modern AI Paradox": while Large Language Models (LLMs) and Generative AI are ubiquitous in headlines and corporate pilots, global aggregate productivity growth  remains sluggish.

Economists like Erik Brynjolfsson argue that the paradox isn't a failure of the technology, but a timing and structural issue. He identifies four main reasons for this lag:

1. Mismeasurement: AI often improves quality, variety, or speed in ways that traditional GDP (which tracks "units produced") fails to capture.

2. Redistribution: AI may be used for "rent-seeking" (competing for market share) rather than increasing total industry output.

3. Implementation Lags: Significant "General Purpose Technologies" (like electricity or the steam engine) require decades of organizational restructuring before they move the needle.

4. Mismanagement: Companies often use AI to automate old processes rather than inventing new, more efficient business models.

Study	Target Group	Productivity Impact Found	Notes on Enterprise Deployment Gaps
MIT/Stanford (NBER)	Customer Support Agents	14% increase in issues resolved per hour.	High-skilled workers saw less gain; impact was greatest on novices. Enterprises often fail to use AI as a "leveler" for training.
Harvard/BCG (SSRN)	Management Consultants	40% higher quality; 25% faster task completion.	"Jagged Frontier": AI failed spectacularly on certain logic tasks where humans over-relied on it, leading to "falling off the cliff" errors.
Microsoft/GitHub	Software Developers	55% faster at completing coding tasks.	Gains are often eaten by "code bloat" and increased technical debt if not managed by senior architects.
Goldman Sachs Research	Aggregate US Economy	Projected 1.5% annual increase over 10 years.	Real-world adoption is currently hindered by power grid constraints and data center infrastructure delays.
NBER / Brynjolfsson et al.	Generative AI & the "J-Curve"	Initial 0% or negative impact.	The "Productivity J-Curve": Measured productivity dips initially as firms invest in "intangible capital" (retraining, restructuring) before the payoff.

While individual tasks show gains, enterprise-wide productivity often remains flat for several reasons:

• The "Pilot Trap": According to recent Adobe/Business research, 86 percent of IT leaders see potential, but only a fraction have moved beyond "isolated experiments" to organization-wide workflows

• Inertial Workflows: Companies often use AI to "do the old thing faster" (e.g., writing more emails) rather than "doing the right thing" (e.g., eliminating the need for those emails entirely). This results in "Digital Overload"

• The Human Bottleneck: AI can generate a report in seconds, but a human still takes hours to verify, edit, and approve it. Without changing the governance and approval structures, the AI speed gain is neutralized

• Data Fragmentation: Most AI models are effective only if they can access clean, centralized data. Most enterprises still have "siloed" data, leading to AI hallucinations or irrelevant outputs

• Skills Gap: Enterprises frequently treat AI as a "plug-and-play" tool like a calculator, failing to realize it requires a new type of "AI Literacy" to prompt and integrate effectively into complex projects.

None of that will be too comforting for suppliers who must justify their heavy AI capital investment. 

But history suggests the payoff is coming. It just will take some time. It always does.

Robotaxi Demand: Saving Time Versus Saving Money?

One traditional way of valuing a particular technology is to create proxies for “time saved,” using typical wage rates per hour. Suppliers often must do so, even if buyers historically are skeptical of the methodology. 

source: Ark Invest 

The average U.S. adult spends nearly an hour per day driving, so the imputed labor cost of all that manual piloting runs in excess of $4 trillion per year, according to Ark Invest analyst Brett Winton. “In addition we pay $1.6 trillion annually for the actual service of driving point to point.”

So one way of modeling market size of robotaxis is to estimate time savings, then impute some sort of hourly value to that time. Perhaps key to that analysis is the assumption that the highest income earners will value their time (and tradeoffs between time and money) more than lower wage earners. 

source: Ark Invest 

Some will instinctively prefer methodologies that simply compare the cost of a product versus the revenue generated by using that product. 

source: Ark Invest

Of course, when a consumer decides to take a robotaxi they are not just trading time for money, they are also avoiding the cost of running their own vehicle. So some will model additional value there. 

source: Ark Invest

So one possible impact is fewer vehicle purchases. 

Also, it is possible that as much as 40 percent of the addressable opportunity (gross profit) is captured in the first 10 percent of metropolitan areas are commercialized. 

source: Ark Invest

Using any set of assumptions, though, among the most important would seem to be the concentration of desirable markets, which are the largest cities. 

The robotaxi market is seemingly one of those cases where the upside is enormous, but the path to capturing it is hard.

Category	Opportunities (Upside)	Challenges (Downside)
Market Size & Growth	Explosive growth potential: projected to scale from ~$0.6B (2025) to >$100B+ by early 2030s (Grand View Research)	Forecasts may be overly optimistic; profitability timelines uncertain and may take years (Business Insider)
Cost Structure	Elimination of human drivers → major long-term cost advantage vs. Uber/Lyft model	Very high upfront costs: ~$150K per vehicle; ~$8+/mile in early deployments (Business Insider)
Unit Economics (Long Term)	High utilization (24/7 vehicles) could drive strong margins once scaled	Current utilization constrained by regulation, geography, and demand density
Demand Trends	Shift to Mobility-as-a-Service (MaaS); declining car ownership in cities (Grand View Research)	Demand sensitive to price and trust; adoption depends heavily on perceived safety (arXiv)
Technology Advantage	Rapid AI, sensor, and compute improvements increasing safety and capability (Grand View Research)	Edge cases (weather, pedestrians, rare events) remain unsolved at scale
Electrification Synergy	EV robotaxis reduce fuel + maintenance costs, improving operating margins (Grand View Research)	Charging infrastructure, battery degradation, and downtime management are operational constraints
Scalability	Platform economics (like Uber) + autonomy could create winner-take-most markets	Scaling is slow and city-by-city due to regulation and mapping requirements
Regulatory Environment	Increasing government support, subsidies, and pilot programs (Grand View Research)	Regulatory fragmentation; approvals required per city; sudden shutdown risks after incidents (The Verge)
Business Models	Multiple revenue models: ride-hailing, B2B shuttles, logistics, partnerships	Unclear dominant model; ride-hailing margins historically thin
Partnership Ecosystems	Strong partnerships emerging (e.g., Uber + Nvidia + OEMs) (Reuters)	Complex value chain: tech + OEM + platform + city → coordination risk
Capital & Funding	Large capital inflows (billions raised) signal investor belief in long-term viability	Cash burn is extreme; many players have exited after losses (GM Cruise, Ford Argo) (Business Insider)
Competitive Dynamics	Market likely supports multiple players (platform + tech + fleet specialization) (Business Insider)	Intense competition from Big Tech, automakers, and startups → margin pressure
Urban Infrastructure Fit	Ideal for dense cities with congestion, parking scarcity, and high demand (Grand View Research)	Limited viability in low-density or suburban/rural areas without subsidies
Safety & Insurance	Potential long-term reduction in accidents vs. human drivers	Liability risk is massive; unclear insurance frameworks
Public Perception	Early adopters show growing acceptance in pilot cities	High-profile accidents can rapidly erode trust and trigger regulation
Operational Model	Fleet optimization, routing, and pricing can be algorithmically optimized	Real-world ops (maintenance, cleaning, repositioning fleets) are complex and costly

The key challenge seemingly is cost per mile compared to the use of human drivers for ride hailing. 

If robotaxis beat human-driven ride-hailing on cost, then adoption could be highly significant. If not, niche use cases will rule. 

And technology alone is not determinative. This is a huge physical infrastructure, regulatory, capital investment and operations challenge. Likely winners will combine:

• strong regulatory navigation

• efficient fleet operations

• smart partnership ecosystems

• disciplined capital deployment.

Layer	Margin Potential	Capital Intensity	Moat Strength	Likely Outcome
OEMs	Low (5–15%)	Very High	Low–Moderate	Commoditized unless integrated
Software	Very High (30–70%)	Very High (R&D)	Very High (if winner)	Few dominant players
Fleet Operators	Moderate (10–25%)	High	Moderate	Quiet long-term winners
Platforms	High (20–40%)	Low–Moderate	Very High (network effects)	Major value capture if dominant

Still, costs matter. And some of the analysis by Ark Invest suggests that robotaxi costs will keep falling. That should help fleet operators. 

source: Ark Invest

Still, the emergence of just a few big winners on the service provider part of the value chain; the platform and software supplier parts of the business. 

But you would already have guessed that.