Michael Burry has changed his mind about one thing.

Not about the AI bubble itself — but about when it could crack.

The investor made famous by The Big Short has now shifted several bearish positions from outright short selling into put options, effectively giving his thesis more leverage and a more defined time window. And the message behind the move is pretty clear: Burry no longer wants to wait until 2028 to find out whether the AI boom has gone too far.

He thinks the correction could arrive sooner.

Burry is changing the clock

In his latest investment update, Burry said he was “moving timelines up” and wanted more leverage in his short positions.

That led him to replace bearish positions involving Micron, Nebius, Nvidia, Palantir and the iShares Semiconductor ETF (SOXX) with put options. Some of the new positions run into June 2027, while others extend to September 2027.

This is important because a put option gives Burry exposure to a decline without simply holding a conventional short position. If the underlying stock falls sufficiently before expiration, the option can become significantly more valuable.

In other words, Burry is not just saying, “I think these stocks are expensive.”

He is positioning for a move that could happen within a specific window.

And that makes the trade much more interesting.

But what exactly is he betting against?

This is where Burry's argument becomes bigger than Nvidia.

His concern is the enormous amount of money being poured into AI infrastructure: data centers, chips, networking equipment and computing capacity.

The obvious question is simple:

How much revenue will all of this infrastructure actually generate?

AI companies and hyperscalers are spending extraordinary amounts of capital today based on expectations of much larger revenues tomorrow. As long as those revenues keep accelerating, the spending can be justified.

But what happens if growth disappoints?

Ares Management has highlighted a similar risk: if AI revenues fail to meet expectations, corporate boards could eventually decide to slow capital expenditure and redirect money toward investments offering stronger conviction.

That is the weak point Burry appears to be watching.

The AI story does not necessarily need to collapse. It only needs expectations to become too optimistic.

And the market is still pushing the other way

Here is the fascinating part.

While Burry is increasing his downside exposure, Nvidia is sending investors a completely different message.

On September 28, Nvidia announced a $150 billion expansion of its share-buyback authorization, bringing the remaining total to $235 billion. The company continues to describe AI and accelerated computing as a massive long-term opportunity.

The Nasdaq also recently reached record levels, showing that investors have not suddenly abandoned the AI trade.

So we have two very different narratives unfolding at the same time.

One says AI is creating a once-in-a-generation technological shift.

The other asks whether the financial expectations surrounding that shift have become impossible to sustain.

And this is exactly why Burry's latest move caught my attention.

The real risk isn't that AI is fake

I don't think the interesting question is whether artificial intelligence will matter.

It almost certainly will.

The bigger question is whether investors are pricing too much future success into today's valuations.

We have seen versions of this story before. Revolutionary technology can be real, transformative and enormously profitable — while the stocks associated with it can still become overpriced.

That distinction matters.

A company can have incredible technology and still be a terrible investment at the wrong valuation.

And a sector can continue growing while its stocks experience a brutal correction.

That is the scenario Burry appears to be positioning for.

Could the AI bubble really burst before 2028?

Nobody knows.

Burry himself is not predicting an exact crash date. His latest move simply suggests that his previous 2028 base case has been brought forward after further research.

That does not mean an AI collapse is guaranteed.

It means the risk-reward has become interesting enough, in Burry's view, to use leveraged downside instruments with expirations concentrated in 2027.

And that creates a fascinating situation for markets.

If AI revenues continue exploding, Burry's puts could become very expensive lessons.

But if capital spending gets ahead of real economic returns, the same options could suddenly look extremely well timed.

For me, the most important signal isn't the word “bubble.”

It's the shift in timing.

Burry once pointed toward 2028.

Now he's looking closer.

And whenever an investor with Burry's history moves the clock forward, the question isn't necessarily “Will he be right?”

It's:

“What does he see that the market may be ignoring?” 🔎