“The pure Nvidia numbers/guidance and strategic vision shows the AI Revolution is NOT a Bubble… instead its Year 3 of a 10-year build out of this 4th Industrial Revolution in our view,” Wedbush tech analyst Dan Ives said recently. For Meta, that revolution has become the new focus, supplanting an expensive bet on the Metaverse that has siphoned off more than $73 billion over five years.
Mark Zuckerberg’s move to slice the budget of Reality Labs by as much as 30% marks a retreat from the virtual reality dream that defined Meta’s rebrand in 2021. That the division’s operating loss reached $4.4 billion in Q3 2025 alone underlined the scale of the financial strain. While Oculus headsets and Horizon Worlds have found niche appreciation, broader adoption of VR hardware has been constrained by persistent limitations: headset bulk, motion sickness risks, and the limited willingness or ability of consumers to spend hours at a time in immersive environments. Even with VR revenue rising 74% year-on-year to $470 million ahead of the holiday season, the unit remained deeply unprofitable.
Investor reaction to the cuts was swift. Meta’s shares leaped 5.5%, adding $93 billion in market value, as Wall Street saw the move as a pivot into higher-margin growth areas. Resources are being redirected squarely at artificial intelligence, where Meta is racing to match or better rivals like OpenAI. It has introduced AI chatbots on Instagram and WhatsApp, as well as video generation tools, positioning AI as the next big engagement driver for its 3 billion monthly active users.
The pivot carries enormous infrastructure implications. Zuckerberg has pledged $600 billion in U.S. data center investment by 2028, part of a wider hyperscaler buildout that analysts at Morgan Stanley estimate will see Big Tech spend $3 trillion on AI infrastructure through 2028. This spending spree mirrors industrywide capital flows, with companies like Amazon, Google, and Microsoft collectively sinking around $400 billion into AI this year alone. Much of this investment is targeted at GPU-rich facilities capable of training and serving large-scale AI models, whose computational demands rise in direct proportion to user interactions. As MIT’s Paul Kedrosky noted, Every time you prompt an AI model, it eats up costs to maintain and cool servers. Those costs rise with the number of users. That’s a problem.
Meta’s own financing structures reflect the sector’s appetite for capacity without overburdening balance sheets. Deals such as the Blue Owl Capital–backed Louisiana data center, where Meta owns 20% but leases all computing output, keep multi-billion-dollar loans off official debt tallies. While such “special purpose vehicle” arrangements provide flexibility, they echo financial engineering tactics from the dot-com era, raising questions about long-term risk if AI demand fails to meet projections.
There’s increasing economic pressure on AI to turn a profit sooner rather than later. Despite explosive user growth for products like ChatGPT, roughly 95% of businesses investing in AI have yet to monetize the technology, according to one MIT study. And the energy and hardware intensity of AI models means scaling them lacks the near-zero marginal cost advantage of traditional software. Andrew Odlyzko, emeritus mathematics professor at the University of Minnesota, cautioned that the revenue required to justify trillion-dollar data center investments “gets into figures larger than total revenues of Google.”
The bet for Meta is that AI-enhanced recommendation systems and generative tools will deepen engagement across its ad-driven platforms, translating into measurable revenue uplift. Nvidia’s Colette Kress has already highlighted the fact that Meta’s AI recommendation systems are leading to “more time spent on apps such as Facebook and Threads.” If sustained, such metrics could validate the strategic shift and justify the infrastructure buildout. Yet the specter of the Metaverse looms, cautioning against this. More than hardware discomfort contributed to VR’s sluggish adoption; privacy concerns, regulatory uncertainty, and ambiguous consumer value each played their part.
Wearable AR/VR devices face challenges from state-level biometric data laws, distracted-use restrictions, and a need for clear social norms to govern recording features. These frictions delayed monetization and kept the technology from breaking into mass-market profitability. By cutting its losses in VR and channeling capital into AI, Meta is aligning with the prevailing growth narrative in Silicon Valley. But as history shows, infrastructure-heavy bets demand both sustained demand and clear monetization pathways. In the absence of either, the risk is not just overcapacity-it’s the possibility of repeating the cycle that once brought the dot-com boom to a halt.
