5 reasons why 'Meta' could be the BIGGEST BUBBLE of AI era

The broader AI market is being called a bubble by some and a revolution by others. But there is a separate question worth asking specifically about Meta: is this stock, right now, a bubble of its own making, inflated by hype around an AI bet that has yet to show a measurable intelligence of its own?

The Capex Number That Shocked Wall Street

When Meta reported its Q1 2026 earnings, the headline revenue numbers were strong, $56.31 billion in quarterly revenue, a 33% year-on-year jump, with advertising revenue of $55.02 billion driving almost all of it. But what moved markets was something else entirely: Meta announced it was raising its 2026 capital expenditure forecast to between $125 billion and $145 billion. That is up from its January guidance of $115 to $135 billion, and represents an increase of over 60% compared to what it spent the previous year. Meta's stock fell 8 to 10% in early trading the following day. Investors were not celebrating the revenue beat. They were staring at a capex number that is, by any measure, staggering.

The Human Cost Nobody Is Pricing Into Meta's Stock

There is another dimension to Meta's AI story that Wall Street is largely ignoring: the catastrophic real-world consequences of deploying a cheap, low-quality AI system at billion-user scale with essentially no human oversight left to catch its errors. Meta has systematically laid off its human content moderation teams, the people who used to review appeals, reinstate wrongly banned accounts, and apply judgment to edge cases. What has replaced them is an AI system that creators, journalists, and small business owners across the world describe as erratic, opaque, and unaccountable. Accounts with years of original content, verified identities, and legitimate income streams are being terminated or restricted without warning and without any realistic path to appeal, while some reporting that fake accounts with blue tick verification badges continue to operate freely on the same platforms. The engineers building this system, referred to internally in some circles as working in conditions resembling a content gulag, are reportedly producing training tasks so rudimentary they resemble puzzles designed for toddlers. That is the training data feeding an AI that is now making consequential, life-altering decisions about whether a creator's account survives, whether a small business keeps its primary sales channel, whether a journalist's archive disappears overnight. The gap between the simplicity of what is going into Meta AI and the seriousness of what it is doing to real people is not a technical footnote. It is a liability, legal, reputational, and financial, that does not yet appear anywhere in Meta's $145 billion capex plan.

Spending More Than It Earns On AI, With No Cloud To Show For It

The core problem with Meta's AI spending is structural. Google, Amazon, and Microsoft are also pouring hundreds of billions into AI infrastructure, but each of them has a cloud computing business that can immediately monetise that infrastructure by selling compute capacity to other companies. Meta has no such business. Its revenue is almost entirely advertising. That means every dollar it spends building data centres and buying Nvidia chips is a bet that AI will eventually make its ads more targeted, its platforms more engaging, and its users more sticky. That may well happen. But right now, it is a hypothesis, not a revenue line.

The Zuckerberg ROI Quote That Should Concern Investors

When Meta's Q1 2026 earnings call touched on what all this AI spending was actually delivering in terms of return on investment, Mark Zuckerberg's response was notable. Asked about measurable ROI from the AI build-out, Zuckerberg reportedly described it as "a very technical question." For a company spending $145 billion in a single year on capital projects, that answer is not reassuring. And the earnings quality underneath the headline numbers is softer than it appears: Meta's reported GAAP EPS of $10.44 was significantly inflated by a one-time $8.03 billion tax benefit linked to the Trump administration's tax policy. Strip that out and the underlying EPS was approximately $7.31, a number that tells a much more cautious story about actual profitability growth.

Michael Burry's Warning

The investor made famous by The Big Short, Michael Burry, has specifically flagged Big Tech's accounting treatment of AI hardware as a concern. His argument is that companies including Meta are extending the stated useful life of their computing assets, primarily Nvidia chips, which reduces annual depreciation charges and makes earnings look better than they are. Burry's estimates suggest Meta could be overstating its earnings by as much as 20.8% by 2028 if this accounting treatment continues. Whether or not that proves correct, the concern highlights that Meta's profitability picture may be softer than its headline numbers suggest.

Where The Stock Stands

Meta's stock hit an all-time high of $796.25 per share in August 2025. As of mid-June 2026, it is trading at $577.22, a decline of approximately 27.5% from that peak. At a market capitalisation of around $1.46 trillion and a business that remains almost entirely dependent on advertising revenue, Meta is not cheaply priced for a company whose AI bet has yet to prove itself operationally as well ass financially. The AI market at large may or may not be a bubble. But Meta's specific combination of enormous spending, advertising-only revenue, unproven AI returns, and a moderation crisis affecting millions of real users makes it a stock with a bubble story entirely its own.