Tom Lee: AI trade still in very good shape fundamentally

10/6/2025 2:52:07 PM	medium term up	20 days later	+3.95%	+3.95%
9/15/2025 11:21:47 PM	medium term up	20 days later	+2.90%	+2.90%
8/22/2025 6:28:53 PM	short term up	8 days later	-0.83%	-0.83%
8/13/2025 3:51:26 PM	short term up	5 days later	-1.88%	-1.88%
8/4/2025 10:55:29 PM	short term sharp up	6 days later	+2.21%	+3.31%
8/4/2025 4:19:52 PM	short term up	6 days later	+2.21%	+2.21%

Let's welcome it now, CNBC contributor. Unstratzed Tom Lee. It's good to have you. Welcome back.

Scott

Tom Lee

Yeah, thanks Scott. Good to see you.

So the market obviously assumes that this is all going to happen. The government's going to reopen and we're just going to get back on our merry way in the stock market. That's certainly what the trade says today. Am I right?

Scott

Tom Lee

Yeah, I think there's a certain measure of relief because the shutdown robbed us of visibility on the macro data and of course there's economic hit and there's suffering cause air traffic is selling. So I think for many reasons this shutdown was a problem for investors.

So what about now between you know now and the end of the year? It's the AI trade that really seems to be debated. How do you feel about it heading into this final stretch?

Scott

Tom Lee

I think it's still in really good shape fundamentally. We know there's a lot of visibility for AI spending, and we know there's still innovation and gain of function taking place with the models. So to me it makes sense that there's still a payoff coming. I think one of the issues that investors are going to grapple with is even though they may say AI stocks are expensive, Nvidia still only trades at 29 times forward earnings. I mean it's still cheaper than Costco. So it's hardly a bubble valuation.

No, but I don't think when they talk about the valuation of AI stocks, they're necessarily talking about Nvidia, I think you would agree with me on that. They're more so talking about names like Palantir and some others where you could easily say, 'Yes, the earnings have been great and the story is incredible, but at some point you're only willing to pay X for a stock and maybe that one's gotten a little bit too rich.' Using that as an example of the kind of names they're talking about.

Scott

Tom Lee

Yeah, I think Palantir is an example of what they call an N equals one company. You've got a founder business developing products that really seem like magic for the companies that adopt it. So should this just be getting a software multiple? I don't think so. Palantir is opening up so many markets and if it was a venture capital firm, they'd be getting potential carry for how they transform businesses. So to me, I think this is still a market discovering how to define the valuation model for Palantir. But to me that's not a bubble. People would have said that about Tesla in 2018 because it looked incredibly expensive and Tesla has been a huge performer since then.

But what was last week about then do you think at its heart?

Scott

Tom Lee

Well, I think at the heart is the market wobbles and so when the market wobbles people do two things. One, they start to take shots at the leaders because they try to take down the leaders thinking that's what leads us down. But second, it highlights the grim reality that we're about to have a third year of double digit gains. The S&P is up 80% over the past 3 years and funds are measured on three-year performance. Only 21% of fund managers are beating 80% over that period. It's one of the worst periods ever for fund managers. So they'd rather see the market fall so that it falls to their performance level. I think what's more likely is the market follows a seasonal pattern which means there's going to be a lot of chasing into year end.

Don't you have any concern at all about the level of money these companies are spending on their buildout and now the new wrinkle of tapping the debt market rather substantially to help do it?

Scott

Tom Lee

Yes and no. AI is proving to be a scale game. It's not like you and I could start a garage and make a model competitive with OpenAI. We need to have scale and a lot of chips. That's really the model required for telecom infrastructure, energy infrastructure, internet infrastructure, and fiber. So it makes sense a handful of players are playing a big role, like in the energy system with eight majors paying a smaller ecosystem. This is developing in the US and it's leaving other countries in the dust.

SignalTube