Why Smart Money Is DUMPING AI for Crypto

Despite a lot of scrutiny, the AI trade is not collapsing.

However, it's fair to say that at this juncture, it's a bit tired.

And in markets, those are very different.

Look at what happened last month when Nvidia reported earnings.

Revenue came in at $96 billion, beating consensus by over 4%, with earnings per share of $2.22, beating by over 6%, and management guiding it to 70% revenue growth for the following fiscal year.

An absolutely huge quarter for an already massive company.

But over the trailing month, the stock managed a gain of 2.8%.

Then there's Broadcom, which reported earlier this month.

Revenue up 86% year on year.

AI semiconductor revenue up 221% to $16.7 billion.

And the CEO raising multi-year guidance to $230 billion of AI revenue by 2028.

Then its shares fell.

after the print, and over the trailing month, the stock is down over 16%.

And that tells you everything that you need to know about the market's appetite.

When a trade is healthy, good news moves prices.

When a trade is exhausted, good news is already in the price, and the only thing left to trade is disappointment.

And the positioning data backs up this idea.

In July, Bank of America's Global Fund Manager survey had 82% of respondents calling long semiconductors the most crowded trade on the planet.

A record.

By August, that had fallen to 53%.

Roughly a third of the most consensus trade in the world stepped back within a single month.

So where did they go? Well, money that has spent two years in one trade does not go to cash when that trade stalls.

Cash is a decision in and of itself.

And allocators, who've been paid handsomely for owning volatility, did not suddenly develop a taste for 4.7% in treasuries.

What they do instead is look for the next asset that offers a similar volatility profile with a more compelling story for the moment.