Meta Killed the Metaverse to Save It: Inside the $90 Billion Pivot From Horizon Worlds to AI Glasses
Reality Labs revenue fell. The losses widened past $90 billion. And somehow Meta stock investors are angrier about the spending than the failure.
Four hundred and two million dollars. That's what Meta's entire Reality Labs division brought in during the first three months of 2026 — against a $4.03 billion loss in the same stretch. Somewhere in Menlo Park, a decision is being made in real time about whether the metaverse was ever the point, or just the story Mark Zuckerberg needed investors to believe while AI caught up.
The Reversal Nobody Saw Coming — Then Everybody Saw Coming
Back in November 2025, Meta's quarterly numbers told a tidy redemption arc: Reality Labs revenue climbing, losses narrowing, Quest 3 selling briskly, and a CEO who looked vindicated after years of investor skepticism. That story didn't survive the winter. By January 2026, Meta had quietly cut roughly 1,000 to 1,500 employees from Reality Labs — shuttering in-house VR studios including Ouro Interactive, retiring the Workrooms meeting app, and pausing third-party development of Horizon OS headsets entirely. A New York Times report at the time indicated cuts could affect more than 10 percent of the roughly 15,000-person XR division.
Then came the part that actually made headlines: on March 18, 2026, Meta announced it was pulling Horizon Worlds — the social VR platform Zuckerberg once promised would host a billion people and hundreds of billions of dollars in digital commerce — off the Quest store entirely, converting it to a mobile-only app. Two days later, Andrew Bosworth reversed course on Instagram Stories, saying the company would keep Horizon Worlds functional in VR "for the foreseeable future" after fan backlash. Whiplash doesn't begin to describe it.
What hasn't reversed is the math. Cumulative Reality Labs losses now sit close to $90 billion since Meta began separating the unit's books — and depending on which quarter you're counting from, some estimates push that figure to $90 billion-plus once Q1 2026's $4.03 billion loss is folded in. That's not a rounding error in a tech budget. That's roughly the GDP of a mid-sized country, spent chasing a computing platform that, by Meta's own internal targets, never got close to its goals.
Is this what failure looks like, or just what an expensive pivot looks like from the outside? Probably both. Meta's overall Q1 2026 revenue still hit $56.3 billion, up 33% year-over-year — its fastest growth rate since 2021, and net income reached $26.77 billion, helped along by an $8.03 billion one-time tax benefit. The advertising engine isn't broken. It's the metaverse half of the bet that's being quietly dismantled, brick by brick, studio by studio.
Why the Spreadsheet Stopped Mattering as Much as the Optics
I've watched enough quarterly earnings cycles to know that investors rarely punish a company for one bad number. They punish a pattern. And the pattern here — narrowing losses in late 2025, then a sudden 1,000-person purge, then a platform shutdown, then a public reversal of that shutdown within 48 hours — reads less like strategy and more like an organization figuring out its own conviction in real time.
Reality Labs operating margin for the broader company came in at 41% in Q1 2026, down from a 48% peak in Q4 2024, and that compression is happening while Meta is simultaneously asking shareholders to absorb a much larger capital expenditure bill elsewhere. The metaverse didn't just fail to deliver — it became the thing Meta had to visibly cut in order to make room for a bet investors actually wanted: AI infrastructure.
The Smart Glasses Paradox: Meta's Cheapest Bet Is Outperforming Its Most Expensive One
Here's the part of this story that should embarrass every Reality Labs budget meeting from the last five years. EssilorLuxottica confirmed that more than seven million units of Meta's AI smart glasses were sold in 2025 alone — and demand kept accelerating into 2026. By January 2026, Meta was reportedly in talks to push annual manufacturing capacity for Ray-Ban Meta frames to 20 million units or more by year's end, a target that would have sounded absurd two years earlier.
The newer Ray-Ban Meta Display glasses, launched at $799 in September 2025 with a built-in waveguide display and Neural Band wrist controller, sold so fast that Meta actually halted new international shipments. By January 2026, the company paused planned rollouts to the UK, France, Italy, and Canada, citing inventory that couldn't keep pace with US demand, with waitlists stretching well into the year. The company also pushed its next-generation Phoenix mixed-reality glasses from late 2026 to early 2027, prioritizing — in Zuckerberg's own framing — sustainable manufacturing over rushed launches.
By the Q1 2026 earnings call, Zuckerberg told analysts that daily AI glasses users had tripled year-over-year — a growth curve that makes Reality Labs' VR hardware look almost stagnant by comparison. The contrast is stark: a $400 million quarterly business propped up almost entirely by glasses nobody needs a headset to wear, subsidizing a $4 billion quarterly loss tied to hardware fewer people are buying.
The Privacy Backlash Nobody at Meta Connect Mentioned
Success brought its own problems. The Electronic Privacy Information Center has urged the FTC to block Meta from deploying facial recognition features on the glasses, after reports surfaced of people being filmed without consent in public spaces. A federal lawsuit filed in early 2026 alleges Meta routed users' glasses footage to human contractors overseas rather than processing it purely through AI — a claim that, if substantiated, directly contradicts the privacy assurances Meta made when the original Ray-Ban Meta line launched in 2023.
This is the unglamorous side of consumer wearables that product demos never show: a camera worn on someone's face is, structurally, a surveillance device whether the wearer intends it that way or not. Meta is now managing a hardware hit and a regulatory headache simultaneously — a combination that historically forces companies to slow down right as momentum peaks.
With apps surging and receding, chasing one craze and moving on from others, and adding new features with each passing year, the FTC has understandably struggled to fix the boundaries of Meta's product market.
Meta Beat the FTC — Then Got Sued Again Almost Immediately
In November 2025, Meta won a five-year antitrust battle when a federal judge ruled the company does not currently hold a monopoly in social networking, rejecting the FTC's push to force a breakup of Instagram and WhatsApp. The judge noted that the competitive landscape had shifted dramatically since the original 2012 and 2014 acquisitions, with YouTube emerging as a dominant force the FTC's case hadn't adequately accounted for. It was, by most legal observers' accounts, a decisive win — and one that arrived just months after Google was found to hold an illegal monopoly in search.
The victory didn't end Meta's legal exposure — it just moved it. The FTC appealed to the DC Circuit in 2026, backed by 29 states and the District of Columbia. Separately, Texas Attorney General Ken Paxton filed a lawsuit on May 21, 2026, alleging that WhatsApp's end-to-end encryption promises were false from the start, a claim that lands awkwardly given Meta's own decision on May 8, 2026 to remove the optional end-to-end encryption feature from Instagram direct messages, citing low adoption.
Regulators rarely let a company breathe for long once they've decided to look closely. Meta won the headline case. It is now fighting a half-dozen smaller ones, any of which could chip away at the trust the company is trying to rebuild around its AI products — including the same smart glasses driving its current growth story.
The Capex Number That Spooked Wall Street More Than the Earnings Beat
Q1 2026 should have been an unambiguous win. Revenue grew 33% year-over-year to $56.3 billion, beating the $55.45 billion consensus estimate, and net income rose to $26.8 billion, or $10.44 per share — crushing analyst expectations of roughly $6.70. Instead, the stock dropped. Shares fell more than 6% in after-hours trading, not because of the results, but because Meta raised its 2026 capital expenditure guidance from a prior $115–135 billion range to $125–145 billion, citing higher Nvidia GPU pricing and ballooning data center construction costs.
| Metric | Q1 2026 | Q1 2025 (Prior Year) |
|---|---|---|
| Total Revenue | $56.3 billion | ~$42.3 billion |
| Reality Labs Revenue | $402 million | ~$412 million |
| Reality Labs Operating Loss | $4.03 billion | $4.21 billion |
| 2026 Capex Guidance | $125B–$145B | $115B–$135B (prior guide) |
| Daily Active People (Family of Apps) | 3.56 billion | 3.43 billion (+4% YoY) |
The irony is hard to miss: Meta is being penalized by markets for spending too aggressively on AI infrastructure in the same earnings call where it's also winding down spending on the metaverse infrastructure that defined its identity for half a decade. Total expenses grew 35% year-over-year, slightly outpacing the 33% revenue growth — a gap that, if it widens further, becomes the story that actually matters more than any single quarter's headline beat.
CFO Susan Li also disclosed something unprecedented in the same call: daily active people fell slightly on a sequential basis to 3.56 billion, the first such quarter-over-quarter decline in the combined family of apps' history, attributed to internet disruptions in Iran and a WhatsApp access restriction in Russia rather than any organic engagement drop. It's a geopolitical footnote rather than a product failure — but it's also a reminder that Meta's scale has made it newly exposed to events entirely outside its control.
Orion, Artemis, and the Glasses That Still Don't Exist
For all the noise around layoffs and shutdowns, Meta's actual long-term AR bet hasn't moved much. Orion, the true augmented-reality prototype Meta showed off at Meta Connect with its silicon carbide waveguide lenses and EMG wristband, remains exactly what it was always described as: a research vehicle, not a product. Each unit reportedly costs around $10,000 to manufacture, and Meta has been explicit that Orion itself will never ship to consumers.
What's coming instead is "Artemis" — a lighter, more manufacturable descendant of Orion's R&D. Industry reporting points to a 2027 launch window, likely timed to follow Meta Connect in late September, though Meta CTO Andrew Bosworth has only committed to a vague "couple years, but not decades" timeline. Developer access to Orion itself is expected to widen in 2026, suggesting Meta wants a software ecosystem ready well before Artemis hardware actually exists.
This is the tension running through everything Meta announced this year. The company wants to be seen as disciplined — cutting VR jobs, narrowing losses, pausing international glasses shipments to fix supply chains. It also wants to be seen as bold — raising capex by $10 billion mid-year, promising true AR glasses by 2027, tripling its AI glasses user base. Those two postures don't usually coexist comfortably inside one earnings call, and Meta is currently trying to hold both at once.
What This Actually Means Going Into Q2
Meta has guided Q2 2026 total revenue to a range of $58 billion to $61 billion, with full-year 2026 total expenses projected between $162 billion and $169 billion. Analyst sentiment remains broadly bullish despite the post-earnings stock dip — 38 analysts tracked by Public.com carry a consensus Buy rating with an average price target near $839, while a separate tally of 66 Wall Street analysts puts the median target at $825, ranging as high as $1,015 and as low as roughly $664.
That spread — nearly $350 between the most bullish and most cautious targets — tells you something honest about where Meta sits right now. Nobody disputes the advertising business is firing on every cylinder. What's genuinely contested is whether the AI infrastructure spend pays off fast enough to justify itself before investor patience runs out, and whether the smart glasses category can keep tripling its user base once the novelty wears off and the privacy lawsuits start landing in courtrooms instead of headlines.
Five years ago, Meta bet its entire corporate identity — including its name — on a virtual world that never found its audience. The company is now quietly unwinding that bet while insisting, in the same breath, that its next hardware platform will succeed where the last one didn't. Whether Ray-Ban Meta glasses are Reality Labs' redemption arc or just a more fashionable repeat of the same expensive pattern is a question this earnings cycle didn't answer. It's one the next four quarters almost certainly will.
Sources & References
- CNBC report on Meta's January 2026 Reality Labs layoffs, 2026
- CNBC report on the Horizon Worlds shutdown announcement, 2026
- CNBC coverage of Meta's FTC antitrust trial victory, 2025
- NPR analysis of Judge Boasberg's antitrust ruling, 2025
- Meta Platforms Q1 2026 10-Q quarterly filing summary, 2026
- Bloomberg report on Ray-Ban Meta glasses manufacturing expansion, 2026
- TipRanks analyst price target consensus for Meta Platforms, 2026
- UploadVR reporting on Orion AR glasses developer access plans, 2025