In 2005, a lone investor sifted through thousands of mortgages and uncovered a ticking time bomb.
He bet against the housing market – and by 2008, Michael Burry’s foresight paid off as the financial system imploded.
Now, nearly two decades after his “Big Short” fame, Burry is sounding another alarm – this time on the AI boom that’s captivated markets and minted trillion-dollar titans.
His claim?
The biggest tech rally of our era might be built on an accounting illusion – a hidden $176 billion mirage of profits – and it could all come crashing down.
If Burry is right, investors riding the AI wave today could face a reckoning tomorrow.
Wall Street is enamored with AI giants posting record earnings – but what if those profits are overstated by 20% or more?
Burry points to a classic financial trick. Tech behemoths like Meta (META) and Oracle (ORCL) have quietly doubled the “useful life” of AI hardware on their books from two–three years to five–six years.
That move slashes yearly depreciation expense and artificially boosts earnings.
Burry’s analysis suggests Big Tech may collectively understate depreciation by $176 billion between 2026 and 2028, inflating profits and misleading investors.
He even speculates Oracle could be padding profits by 27% and Meta by 21% through this sleight-of-hand.
In his words, “understating depreciation by extending asset life” is one of the great frauds of modern finance.
It’s a contrarian call in a market obsessed with AI’s limitless potential. Burry essentially asks: Are the dazzling earnings of the AI era a lie?
If so, today’s AI darlings could become tomorrow’s ticking bombs – much like those subprime CDOs he shorted in 2008.
In this article, we’ll unpack Burry’s warning and its far-reaching implications.
We’ll also explore a startling twist: What if this time the bubble doesn’t burst … because governments and tech kingpins refuse to let it?
Burry may be crying wolf, but the AI revolution has some very powerful guardians.
Burry’s Big Warning: Inflated Profits in the AI Gold Rush
Michael Burry isn’t known for idle chatter. When he speaks, markets listen.
Recently, the famed contrarian – best known for predicting the 2008 crash – took to social media to warn that major AI and cloud companies are cooking their books.
According to Burry, the hyperscalers — think Meta, Alphabet (GOOGL, GOOG), Microsoft (MSFT), Oracle, Amazon (AMZN) — have been stretching the lifespans of their AI servers far beyond reality.
Why does that matter? Imagine a $90 million AI supercomputer that, in truth, will be obsolete in three years.
Instead of charging $30 million of depreciation per year over three years, a company might declare it will last 6 years – expensing only $15 million per year.
Result: an extra $15 million paper profit annually, per machine. Do this at scale, and suddenly earnings look fantastic – perhaps too fantastic.
Burry’s data shows this is exactly what’s happening.
Since 2020, tech giants have roughly doubled the assumed life of their networking and AI hardware. Meta’s gear went from three years to ~5.5, Google’s from three to six, Oracle’s from five to six, and so on.
That halved their depreciation costs per unit, juicing net income. Burry calls it out as essentially “earnings inflation.”
Burry estimates by 2028, these tricks could make Oracle’s profits 27% higher (on paper) than they would be under traditional accounting, and Meta’s about 21% higher.
In short, the AI boom’s stellar earnings might be built on sand. If and when those data centers need a refresh sooner than claimed, companies could get slammed with extra costs.
Think of it as an “AI hardware hangover.”
In a few years, capex bills surge, profit margins shrink, and today’s lofty valuations come back to earth.
Burry’s calculation: a 20% hit to earnings and a 20% cut to market multiples could easily knock 40–50% off Big Tech stock prices.
For an investor who lives by asymmetry (big downside bets for huge payoffs), this smells like the next Big Short.
Will the AI Bubble Pop… or Is This Time Different?
Burry’s alarm has investors asking if the unthinkable could happen: could the mighty AI trade – which minted the world’s first $5 trillion company and sent stocks like Nvidia soaring 1,200% – suddenly collapse?
History is full of booms turned bust. The dot-com bubble didn’t burst because people stopped using the internet (they didn’t); it burst when financial reality caught up with overhyped dreams.
Burry suggests a similar fate could befall AI: demand for AI isn’t going anywhere, but a financial gotcha could still prick the bubble.
However, a growing chorus argues “this time is different.” We count ourselves amongst them, believing Burry’s 2025 warning will not play out like his 2008 heroics.
Why?
The core of Burry’s thesis might be flawed.
If tech companies are genuinely getting more life out of their AI hardware, then extending depreciation isn’t fraud – it’s just good accounting.
And indeed, there are reasons to think today’s AI gear does last longer:
- Unprecedented demand: In the current AI gold rush, new chips aren’t replacing old ones – they’re adding to the fleet. Cloud giants need every processor they can get. Even older GPUs find use chugging away on less demanding tasks. In a world starved for AI compute, nothing goes to waste. Thus, an Nvidia accelerator bought in 2021 might still be running jobs in 2026, well past the old three-year norm. Scarcity has extended these machines’ useful lives organically.
- Technical improvements: Better cooling, power management, and software optimizations mean hardware endures less stress and can stay productive for longer. Data-center innovation isn’t just about raw speed; it’s also about longevity. If engineering advances have doubled the effective lifespan of servers, then booking five to six years of use isn’t “cooking” anything … it’s reality.
From this perspective, Burry’s dire forecast could prove a false alarm.
If depreciation schedules are accurate, there’s no “refresh cliff” looming to wreck profits.
The AI giants would keep minting cash, and Burry’s short positions would bleed losses.
It wouldn’t be the first time a bold prophet turned out wrong; even a legendary investor can call 15 of the last two recessions.
Too Big to Fail? The Wild Card of Government and OpenAI
Even if Burry’s right about the numbers, the AI boom has an ace up its sleeve that 2000 or 2008 didn’t – massive, determined support from governments and mega-investors.
This isn’t just another tech fad; it’s being treated as critical national infrastructure.
Think Manhattan Project, not Pets.com.
Consider this: SoftBank’s CEO Masayoshi Son – arguably the world’s savviest tech investor – just liquidated a $5.8 billion stake in Nvidia.
Not because he’s souring on AI, but to double down on it.
He’s reallocating that capital into OpenAI (the creator of ChatGPT) and a colossal $500 billion “Project Stargate” to expand U.S. AI data centers.
In Son’s view, the real opportunity isn’t owning chip stocks at record highs – it’s backing the next phase of AI’s evolution directly.
His message: the AI revolution is just getting started, and he’s all-in.
Then there’s Uncle Sam.
The U.S. government has made it abundantly clear: we must win the AI race against China. That means pouring incentives, contracts, and perhaps even backstop funding into the sector.
When OpenAI’s CEO (Sam Altman) muses about needing government help to finance $1-plus trillion in AI projects, officials don’t flinch – they invite him to the White House.
We’re witnessing an unprecedented public-private alliance to ensure AI succeeds on American soil.
In practical terms, that could mean safety nets for AI companies that get in trouble, preferential treatment, and ongoing capital infusions that defy normal market logic.
In other words, AI might be “too big to fail.”
In the 1940s, nobody in Washington asked if the Manhattan Project was hitting quarterly earnings targets – they just knew it had to succeed.
Similarly, today’s leaders (and deep-pocketed investors) may not care if an AI firm’s depreciation math is squishy.
They’re willing to overlook some red ink or creative accounting to keep the innovation engine roaring.
Growth at any cost, backed by the full faith and credit of those who shape the economy.
The Bottom Line: Caution or Full Throttle?
So, where does that leave investors?
On the one hand, you have Michael Burry – a proven seer of market cracks – waving a big red flag.
He sees fake profits puffing up a bubble, and he’s positioned to profit if it pops.
His warning shouldn’t be dismissed; even those who disagree acknowledge he’s highlighted a real issue in tech accounting.
At minimum, due diligence is warranted.
Blindly chasing AI stocks higher, assuming they can only go up, is a dangerous game.
If the “depreciation deception” unravels, the comedown could be ugly for overhyped names.
On the other hand, the power dynamics behind this AI boom are unlike anything before.
We’re talking about trillion-dollar companies, visionary billionaires, and governments all aligned to propel AI forward.
They’re investing in AI like the future depends on it – because, in their view, it does. That kind of momentum can steamroll a lot of short-term concerns.
Just look at user adoption: ChatGPT hit one million users in five days of launch (faster than any platform in history), and it’s already serving 300-plus million weekly users a year later.
The scale and speed of AI adoption are unprecedented – and perhaps unstoppable.
In the end, this might be one of those rare showdowns where financial cynicism meets technological inevitability.
Is the AI boom a bubble ripe to burst, or a rocket ship that will leave skeptics like Burry behind?
The prudent path for investors is to stay informed and balanced.
Keep an eye on those earnings quality metrics – depreciation, free cash flow, capital expenditures – especially as we approach 2026–2027 when the “refresh cycle” verdict will come due.
At the same time, recognize that we’re in a new era where traditional metrics can be overshadowed by strategic importance.
Michael Burry has issued a stark reminder that no boom is without risks.
But unlike 2008, there are giants holding up the AI sky, determined not to let it fall. For hypergrowth investors, the takeaway is clear: approach the AI revolution with eyes wide open.
Heed the warnings, but don’t lose sight of the bigger picture. The next few years will reveal whether Burry’s dark prophecy comes true – or if this revolution truly is different.
Either way, fortunes will be made and lost in the balance.
Stay tuned, stay savvy, and buckle up – the ride isn’t over yet.
Burry’s warning is a reminder: every boom creates winners and losers. The key is knowing which side you’re on.
If Michael Burry is right, parts of the AI boom could face a painful reality check.
But that’s exactly why now is the moment to understand where the real profit engines are forming — not in hype, but in the hard tech powering the next phase of AI.
That’s why I recently teamed up with trading pro Jonathan Rose for The Profit Surge Event — a deep dive into the second leg of the AI revolution, where physical robots, sensors, actuators, magnets, and robot-grade energy systems become the new battleground for exponential growth.
While I map out the multi-trillion-dollar megatrends reshaping the economy, Jonathan zeros in on the surge points — the moments where adoption curves go vertical and certain supply-chain stocks explode. It’s the same approach behind his gains of +959% on Albemarle, +534% on MP Materials, and +233% on Rigetti.
Together, we reveal how to pair long-term conviction with short-term precision — the exact formula investors need as humanoid robots move from factory pilots … to warehouse fleets … and soon, into millions of homes.
If the Physical AI era is the true “main event,” this is the blueprint for capturing its biggest windfalls.
Click here to watch the replay and position yourself for the next humanoid-AI surge.
