The AI theme on Wall Street could eventually end in a bubble and crash but it’s not there yet.

This is the conclusion of Alger’s star portfolio manager Ankur Crawford.

I’ll add to that idea.

For starters, saying we’re not in a bubble is not the same thing as saying the stocks related to the AI theme won’t ever come down. Too many people speaking on the markets deal in absolutes. As Obi-Wan warned us in Episode III, “only the Sith deal in absolutes.”

So if I say it’s not a bubble, I’m not saying there won’t be a drop in share prices. I am only saying the vast majority of investors are not recklessly speculating on technology bets that are extremely improbable to have a payoff. The payoff from AI-related capex spend is already in progress. Not every project will have proven to be successful in the long-term, but what other endeavor could you ever make this claim for.

The reality is that we are still short computing power given the explosive use of generative AI and there is no end in sight. Yet. And until there is, the spend will continue. Ankur doesn’t think - given already announced capex plans and the current supply-demand imbalance for chips - that we will see this market go into any sort of surplus through at least 2028.

What is more likely, at this stage in the game, is that the technology will advance sufficiently so that companies will begin saying they are able to accomplish more inferencing with their existing infrastructure so that the cost of all this compute can come down. They will not be saying the demand has come down. They’ll just get better at satiating it. This is a positive, not a negative. If you disbelieve this idea, then your only out is a world in which we close Pandora’s box, everyone stops using AI tools in their work and play, and we pretend it never happened.

Which of these scenarios seems most likely - hyperscalers get better at prioritizing chip and electricity use or people just stop generating tokens and delete the app? If it seems obvious, it should be.

And again, for the people in the back, two things can be true at the same time. The AI theme could be perfectly rational AND the related stocks could fall. Here’s Micron’s drawdown chart over the last three years. Don’t tell anyone, but this company absolutely destroyed the analysts’ earnings expectations the last time they reported, gave increased guidance out as far as the eye could see and then the stock promptly fell over 30% (bottom pane):

It’s since recovered, but this is the point. Stocks don’t always obey the fundamentals of the backdrop. Because they are owned by people, traded by algorithms created by people and perceptions change as quickly as human emotions.

This is what investors have to put up with if they’re going to win. Many cannot.

One year ago today Micron was selling for $68 per share. It closed Friday at $455. Riding a stock for a a septuple (7x) over the course of a year means enduring periodic corrections that could be mindblowing in scope. Anyone who says this is easy is lying.

Someday, Micron, Ciena, Dell, Sandisk, Nvidia and the others will fall 30% and just keep on falling. Everyone knows it. This spring it didn’t happen. Maybe this summer. Maybe in the fall. Maybe three years from now when the world has finally caught up on the demand for compute and nobody needs needs the next optical router or server rack cooling system or graphical processing unit or shipment of DRAM.

We’re not there yet. That’s the message of the markets this month. That’s the message we got from Ankur on this week’s brand new episode of The Compound and Friends.

The Nasdaq has just recorded its 13th consecutive positive daily close. That’s something that has only happened three times in history. From the March 30th low, the index is up 16%. If you’re watching from the sidelines, it’s easy to say “bubble.” It’ll definitely make you feel better about yourself.

Then what?

There are lots of reasons to be cautious right now (there always are). While it’s true that valuations for mega-cap tech have come down since the start of the year, it’s not like they’re pricing in worst-case scenarios surrounding the naval conflict in the Strait of Hormuz. Nobody believes in this war ending cleanly except for the software programs that run the Dow Jones up a thousand points every time Trump insists it’s a done deal. If you want something to worry about, worry about that instead.

Because the AI revolution isn’t going to be turned back. In many ways it has barely even begun. Autonomous driving is in the first inning. The Low Earth Orbit (LEO) economy is in the first inning. Humanoid robots are in the first inning. OpenAI, SpaceX and Anthopic haven’t even come public yet. We may end up with a bubble (because we always do), but let’s not fade something that hasn’t even arrived yet.

One thing we can say for certain is that none of this is going back into the box.

You can watch or listen to Ankur’s appearance on the show at the links below. I guarantee you’re going to learn a lot from her just as Michael and I have. I’m biased, of course,, but this episode is incredible.

Speaking of the Low Earth Orbit economy, this is a chart of Rocket Lab, the best publicly traded proxy we have for space-related investor enthusiasm. With the SpaceX IPO potentially coming as soon as this June, investors who haven’t gotten access to pre-IPO shares of Elon Musk’s rocket company are buying this one instead. Rocket Lab has re-usable rockets and contracts to launch both commercial and government satellites into space. They’re also building and operating the equipment necessary to tow satellites and other orbital devices into their proper place. The company is based in California and launches its rockets from two places, New Zealand and Virginia. They’re winning new contracts and iterating their technology at a rapid rate. It’s a stock that should be on your screen. You’ve heard me talking about Planet Labs (PL) and Joby (JOBY) and Archer (ACHR). Here’s another name in the same theme.

THE COMPOUND & FRIENDS

Michael Batnick⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ and ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Downtown Josh Brown⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ are joined by Dr. Ankur Crawford to discuss: the strongest names in AI, the future of the Mag 7, the semiconductor market, the case for Amazon, and much more!