Michael Burry has once again stepped into the middle of one of Wall Street's biggest debates.

The investor made famous by The Big Short is increasing his bearish exposure to some of the market's hottest artificial-intelligence and semiconductor names.

His latest targets include Nebius, Micron and Oracle, along with bearish exposure through a semiconductor ETF.

The message behind the trades is blunt:

Burry believes investors may be paying far too much for the AI boom.

According to the latest report, Burry added to his short position in Nebius at around $247, increased his Micron short around $924, and also added to bearish exposure involving Oracle and the semiconductor ETF.

The timing could hardly be more dramatic.

AI stocks have been delivering enormous gains.

Semiconductor companies are reporting record demand.

Data-center spending continues to surge.

And yet one of the investors most closely associated with identifying financial bubbles is positioning for a reversal.

Michael Burry became famous for recognizing the housing bubble before the 2008 financial crisis.

His bet against subprime mortgages became the central storyline of The Big Short.

That history has turned his investment decisions into market events.

Whenever Burry publicly reveals a bearish position, investors immediately ask whether he has identified another bubble.

But there is an important caveat.

Being right about a bubble eventually forming does not mean the timing will be right.

Markets can remain irrational longer than a short seller can remain solvent.

That lesson is especially relevant in today's AI market.

Nebius is one of the more aggressive AI infrastructure plays.

The company is building data-center capacity designed to serve the rapidly growing demand for AI computing.

That puts it directly inside one of the hottest investment themes in the world.

But it also creates enormous capital requirements.

AI infrastructure is expensive.

Companies need data centers, networking equipment, advanced processors and power.

The business can grow rapidly if demand remains strong.

But high growth does not automatically equal attractive stock returns.

That appears to be central to Burry's argument.

Recent reporting described Burry as viewing Nebius as resembling the late stages of an AI boom.

Burry's bearish position in Micron may surprise investors because the memory-chip industry is currently benefiting enormously from AI.

High-bandwidth memory has become one of the most important components in modern AI systems.

Demand is strong.

Supply is constrained.

Prices have risen.

And major memory producers are investing aggressively to expand capacity.

Micron is therefore one of the clearest beneficiaries of the AI infrastructure boom.

So why short it?

The answer may be valuation and cyclicality.

Semiconductors have historically been cyclical.

When demand is strong, companies expand production.

Eventually supply catches up.

Prices fall.

Margins compress.

And earnings decline.

A short seller may therefore believe that today's extraordinary memory profits are unsustainable.

The AI boom is creating genuine economic demand.

This isn't simply a story about companies putting "AI" in their presentations.

Data centers are actually being built.

Cloud providers are spending billions.

Semiconductor companies are reporting stronger revenue.

SMIC, for example, just reported second-quarter profit of approximately $479.2 million, more than triple the previous year, while revenue increased 36% amid strong AI-related demand.

That makes Burry's bet more difficult than simply calling AI a fraud.

The technology is real.

The demand is real.

The question is whether stock prices have moved further than fundamentals justify.

This is where bubble analysis becomes more subtle.

A company can be excellent and still be a bad investment at the wrong price.

That distinction has defined many market cycles.

During a boom, investors often stop asking whether a company will succeed.

Instead, they assume success and focus on how much bigger the company can become.

Valuations rise accordingly.

Eventually, expectations become so high that even outstanding earnings may disappoint.

A company can report record profits and see its stock decline because investors expected even more.

That appears to be the kind of risk Burry is positioning around.

Oracle has become increasingly associated with AI infrastructure because of its cloud business and massive data-center ambitions.

The company has benefited from rising demand for computing infrastructure.

But those ambitions also require enormous capital expenditure.

Investors must therefore consider the relationship between revenue growth and the cost of achieving that growth.

If AI demand remains explosive, Oracle could continue benefiting.

If customers slow spending, however, the economics of aggressive infrastructure expansion could become less attractive.

Burry's bearish position suggests he believes the market may be underestimating that risk.

Burry is not simply betting against individual companies.

He has also added bearish exposure to the broader semiconductor sector through options on the iShares Semiconductor ETF (SOXX).

That is significant.

A single-company short can be based on company-specific problems.

A semiconductor ETF position expresses a broader macro thesis.

It suggests Burry may believe the entire AI hardware trade has become vulnerable.

That would include memory.

Processors.

Networking.

Manufacturing equipment.

And other companies benefiting from data-center investment.

The strongest challenge to Burry's thesis is that AI demand continues to produce real earnings.

The semiconductor industry isn't simply trading on promises.

Companies are showing revenue growth.

Capacity is constrained.

Customers are signing enormous contracts.

And demand forecasts remain extraordinarily strong.

Omdia recently projected very strong semiconductor growth in 2026, driven heavily by AI and memory demand, with shortages expected to persist in key areas.

That makes an immediate collapse far from inevitable.

This may be the most important point.

A bearish thesis can be fundamentally correct and still lose money in the short term.

If AI infrastructure spending accelerates for another year or two, semiconductor companies could continue producing extraordinary earnings.

Their stock prices could continue rising.

Burry's short positions could therefore move sharply against him before the market ever reaches the conditions he anticipates.

The latest reports already show how dangerous that can be.

Nebius shares recently surged dramatically, leaving bearish traders under pressure.

That is the classic problem with short selling.

Losses can become enormous when a stock keeps climbing.

The conversation has moved beyond asking whether AI is real.

It obviously is.

The real questions are:

How much will AI earn?

How quickly will customers monetize it?

How long will infrastructure spending remain elevated?

How much competition will emerge?

And how much of that future growth is already reflected in stock prices?

Those questions are much harder.

They require investors to separate technological progress from financial returns.

Burry's bearish thesis does not necessarily require AI spending to disappear.

Stocks can fall even while companies continue growing.

All that is required is for growth to come in below expectations.

Suppose investors expect a company to double revenue.

If revenue rises 50%, the business is still performing extremely well.

But the stock could decline if investors had already priced in 100% growth.

This is why valuation matters.

The most useful lesson isn't necessarily "sell AI stocks."

It is to pay attention to expectations.

Burry's positions serve as a warning that the strongest part of the market may also be the part where expectations are most demanding.

Investors should examine:

A great company at an extreme valuation can become a dangerous trade.

Wall Street is now effectively watching two competing narratives.

The bullish argument says AI is still in the early stages of a technological revolution.

The bearish argument says investors have already priced in too much of that revolution.

Both can contain some truth.

AI can transform the economy while some AI stocks still decline.

AI can create trillions of dollars of economic value while investors lose money buying companies at inflated prices.

That distinction is crucial.

Ultimately, the market will decide who is right through financial results.

If AI infrastructure spending continues accelerating, Burry's shorts could remain under pressure.

If growth begins slowing while valuations remain extreme, his strategy could start looking increasingly prescient.

That is why his latest trades are worth watching—but not blindly copying.

The investor may have a long-term thesis.

The market may have a different timeline.

And short positions can be extraordinarily risky even when the underlying argument is eventually proven correct.

For now, Burry has planted his flag firmly on the bearish side of the AI debate.

The AI boom is producing real money. Michael Burry is betting that Wall Street has already priced in too much of it.

The next few earnings seasons may reveal which side has underestimated the future.