Fund Managers Dump Tech at Fastest Pace in 20 Years

A fund manager walking away from monitors showing red falling stock charts on a dark Wall Street trading floor

Michael Burry of The Big Short Is Betting Against the Market Again

Up through October, Wall Street fund managers were completely drunk on the tech party. They were piling into tech stocks with more enthusiasm than at any point since August 2020.

Then, just a month later in November, the mood suddenly froze over, and they started hammering the sell button faster than at any time since June 2006, almost 20 years ago.

✦ BofA Global Fund Manager Survey

November 2026 survey

Survey periodNov 6-12
Fund managers surveyed193
Total assets under management$526 billion
Tech net overweightOctober 37% → November 12%

Bank of America (BofA) surveyed 193 fund managers around the world (total assets under management of $526 billion) from November 6 to 12 for its Global Fund Manager Survey (FMS), and the results came out on November 17.

Here's what it shows. Take the share of managers holding more tech than the market average and subtract the share holding less (the net overweight).

That figure went from 37% in October to 12% in November, and it dropped a whopping 25 percentage points in just one month.

There are still 12 percentage points more managers overweight tech than underweight, but you can feel how much they cut their tech exposure in a single month.

Going by how fast they trimmed, it's a sudden change of stance that's hard to match in the past 20 years.

Coincidentally, on November 12, the day this survey closed, Michael Burry, famous as the real-life model for the movie The Big Short, made a move too.

That day, right after the Nasdaq 100 bounced off the 27,000 level, he posted an entry titled "Behold the bull-trap" in his paid newsletter, Cassandra Unchained.

He says he doubled his position in Nasdaq 100 put options expiring in early 2027 (a bearish bet that pays off when prices fall).

He even bought back the semiconductor ETF put options he'd sold earlier, and he likened the Nasdaq's October 6 peak to 'right before the 2000 dot-com bubble burst'.

The actual stock market feels very different from before, too. NVIDIA (NVDA) posted great earnings after the close on November 17, with quarterly revenue of $115.8 billion (+103% from a year earlier).

Still, it closed the November 18 regular session down 2.89% at $150.96.

The S&P 500 also slid for a fifth straight session, ending at 7,323.63.

〈Bloomberg report on the AI selloff〉
* From the official Bloomberg Podcasts channel.

Rechecking the AI Bubble Talk We're All Sick of Hearing

✦ Michael Burry's Moves

From November 2025 to November 2026

Early November 2025
Disclosed Palantir and NVIDIA put options in his last 13F filing
June 2026
Put Caterpillar and semiconductor stocks into his short basket
August 2026
Bought NVIDIA call options to hedge his bearish bets
Fall
Moved up the timing of the AI bubble bursting from 2028
November 12, 2026
Doubled Nasdaq 100 put options, bought back semiconductor ETF put options

So why did Wall Street fund managers and Michael Burry start doubting AI stocks at almost the same time? Once you look at what Burry has been digging into for the past year, it starts to make sense.

In his last 13F filing in early November 2025, Burry's hedge fund, Scion Asset Management, revealed some big bearish bets: Palantir put options (notional value of about $912.1 million) and NVIDIA put options (about $186.6 million).

After that, he opened up his newsletter and started pouring out his AI bubble warnings.

The sharpest thing he went after was Big Tech's 'depreciation accounting'. Say you buy a $40,000 AI chip and have to swap it for a newer one in two to three years. Then it's only normal to record a big expense on the books every year.

But Burry's claim is that giant cloud companies (hyperscalers) like Meta, Microsoft, Google and Oracle assumed a long useful life of five to six years for their servers and chips, which shrinks the expense they book each year.

✦ Spreading 3-Year Equipment Over 6 Years

The depreciation accounting Burry flagged

In reality
• Equipment you have to replace after about 3 years
In accounting
• Spread out as if it could be used for 6 years
• Near-term book profit looks much bigger
Burry's estimate
• Depreciation expense under-recorded from 2026 to 2028: about $176 billion

For example, say there's a piece of equipment you'd actually have to replace after about three years. Spread that over six years as if it'll last that long, and the near-term book profit looks a lot bigger.

Burry estimated that Big Tech's under-recorded depreciation expense between 2026 and 2028 alone would come to about $176 billion. The debate got even hotter when Amazon shortened the life of some of its servers from six years to five in its annual report.

People started asking whether the lifespans had been set way too long in the first place.

This past June, Burry put the heavy-equipment maker Caterpillar, which surged on AI data center demand, and semiconductor stocks into his short basket.

In the fall, he moved the timing up too, saying "the AI bubble could burst much sooner than the 2028 originally expected."

Of course, Burry isn't always right. Every time he's warned about a bubble since late last year, prices actually jumped even higher, and he struggled so much with timing that in August 2026 he even had to buy NVIDIA call options to hedge his bearish bets.

Still, his long-held suspicion that a single quarter of disappointing revenue could be enough to break the bubble is starting to look more and more real.

〈Legendary investors on the AI bubble〉
* From the SanDē Films YouTube channel.

Michael Burry: See? I Knew I'd Be Right. I'm The Big Short, The Big Short.

Right. 'AI bubble' isn't just Burry's lone cry anymore.

In the November BofA survey, 38% of fund managers answered "companies are overinvesting in AI," which is a record high for the survey.

On top of that, answers saying "AI hyperscalers will cut capital expenditure (capex) in the next six months" jumped from 30% in October to 52% in November. More than half the managers think a slowdown in Big Tech spending is right around the corner.

When asked where they want companies to spend the cash they earn, 42% picked improving financial health, such as paying down debt, and 33% picked shareholder returns, like dividends and buybacks.

The 'increase capex' answer shrank to 16%, the lowest since May 2020.

It seems like a market that ran ahead believing in AI's future has lost a lot of its confidence.

And the ranking of tail risks, the extreme risks the market fears most, flipped too.

✦ Tail Risks and Capex Outlook

October vs. November 2026, share of respondents

OctNov
Tail risk: AI bubble
Oct
23%
Nov
34% (+11pp)
Tail risk: AI hyperscaler credit event
Oct
11%
Nov
31% (+20pp)
Tail risk: second-wave inflation
Oct
20%
Nov
14% (-6pp)
Tail risk: disorderly rise in bond yields
Oct
31%
Nov
13% (-18pp)
Expected AI hyperscaler capex cut within 6 months
Oct
30%
Nov
52% (+22pp)

First off, in October the biggest worries were a disorderly rise in bond yields (31%) and second-wave inflation (20%), right?

But in November, worries about rates (13%) and inflation (14%) faded, and 'AI bubble' (34%) and 'AI hyperscaler credit problems' (31%) shot straight into first and second place. The two AI risks alone add up to 65%.

On the question of where a credit crisis that could rock the financial system would start, 'AI hyperscaler corporate bonds' (38%) and 'private credit' (30%), the non-bank lending market, ranked first and second as well.

Why? It used to be that Big Tech built data centers with their overflowing cash, but things are different now.

First, Oracle kept pouring money into AI infrastructure until its annual free cash flow swung to a $23.7 billion deficit, and this past July S&P downgraded it to 'BBB-', the very edge of investment grade.

The way Meta built Hyperion, its giant data center in Louisiana, looks a bit questionable too. It teamed up with Blue Owl Capital, a private credit manager, set up a Special Purpose Vehicle (SPV), and issued $27.3 billion in project bonds.

In other words, Meta pulled in money from outside its financial statements. Burry was probably watching exactly this kind of thing.

What Did Wall Street Fund Managers Buy After Dumping Tech?

If Wall Street fund managers dumped tech at the fastest pace in 20 years, did they sell all their stocks and go to cash? Nope.

The share of managers 'net overweight global equities overall' fell just 3 percentage points, from 44% in October to 41% in November, and 'average cash as a share of assets under management' actually dropped from 3.9% to 3.6%.

That's even the lowest since August.

So where did the money that left tech go? Most of it moved into bonds, neglected sectors, and outside the United States. Within equities, healthcare pushed tech aside to become the most favored sector.

Consumer companies got attention too, and as many as 71% of fund managers said they prefer companies with solid profits right now over 'dream stocks' with volatile earnings.

✦ October vs. November: What Changed

Category (net position or response share)October 2026November 2026Change from previous month
Tech net overweight+37%+12%-25pp (biggest drop since June 2006)
Healthcare net overweight+23%+34%+11pp (rose to most favored sector)
Bonds net underweight-37%-23%+13-14pp improvement (underweight narrowed)
Cash as share of assets under management3.9%3.6%-0.3pp (less buying power)
Expected AI capex cut within 6 months30%52%+22pp surge
Total AI-related tail risk*34%65%+31pp spike
* Sum of AI bubble (October 23% → November 34%) and AI hyperscaler credit event (October 11% → November 31%)

In particular, fund managers ranked 'global equities excluding the United States' (36%) ahead of US stocks (24%) and gold (17%) as the asset class that will perform best in 2027.

The Japanese yen (42%) was picked as the currency expected to be strongest, and the share saying gold is undervalued also hit a record high.

Anyway, the money that was piled into US Big Tech has started to scatter into assets around the world, like global equities and gold.

Maybe that's why the US stock market has been so lifeless lately.

Do You Trust Michael Burry? Or Is Now, With Everyone Shouting 'AI Bubble,' the Chance?

So with fund managers trimming tech and Burry stacking up put options, might now actually be a chance to buy the dip? BofA chief investment strategist Michael Hartnett's answer was firm.

Basically, he said "the AI boom and euphoria are over, so stay risk-averse for now, and fear hasn't built up enough in the market yet to start buying the dip." Let's look at why Hartnett sees it that way.

✦ Three Reasons Hartnett Gave

✔ First: No. 1 most popular trade, buying global semiconductors, 39%
✔ Second: Average cash level 3.6%, below the 4.0% sell threshold
✔ Third: Recession outlook 3%, Bull & Bear 7.4

First, the most popular trade fund managers named is still 'buying global semiconductors' (39%), and third place is 'buying the Magnificent 7' (12%).

It's cooled down from July, when the semiconductor crowding was at 82%, and it's true fund managers cut their tech exposure, but the heat is still there.

Second, the 3.6% average cash level is still below the 4.0% sell threshold that BofA watches.

Put simply, if the money sold during the recent tech plunge had been cashed out in large amounts, it could turn back into buying. But this time the cash level actually fell from 3.9% to 3.6%, right?

Right now, there isn't enough cash ammo to catch stocks when prices fall.

Third, the 12-month recession outlook is a mere 3% (soft landing 50%, no landing 42%), and BofA's contrarian indicator, the Bull & Bear Indicator, only slipped from 8.1 in October to 7.4 in November, far from 2.0 or below, the real buy-the-bottom signal.

It's an in-between stretch where everyone says they're nervous but nobody has actually panic-sold or thrown in the towel yet.

✦ Weighted-Average Targets for End of 2027

The outlook fund managers see

S&P 500 index
7,926 (about +8.2% from the November 18 close)
10-year US Treasury yield
4.9%
Gold price
$4,341 per ounce

The fund managers' weighted-average targets for the end of 2027 are 7,926 for the S&P 500 (about +8.2% from the November 18 close), 4.9% for the 10-year US Treasury yield, and $4,341 per ounce for gold.

The end of 2027 is a year out, and if they only expect the S&P 500 to gain 8.2% by then, they probably don't see anything like the strong rally of recent times happening.

So will Burry's short bet win in the end? Or will the AI bulls take control again? I'd like it to be the latter, but the market conditions keep seeming to side with Burry, and that scares me.

This isn't investment advice. You know that, right? Let's each protect our own hard-earned money.

That's all for this post. See you next time.

Related posts

No comments:

No comments yet. Want to leave the first one?