Watch the Video
Our conversation began with a simple observation. A lot has happened in AI over the past week, and not all of it has been good. Lee opened by pointing directly at Oracle’s earnings report, which on its own would have looked solid for most technology companies. The problem, as we both agreed, is that Oracle is no longer being valued like most tech companies. It is being valued as a core pillar of the AI data center buildout.
When Oracle failed to meet revenue expectations and suggested that large-scale data centers may not be ready until 2028, it immediately chilled sentiment across the entire sector. On the same day the Dow pushed to new highs, the Nasdaq was hit hard, with anything even loosely tied to AI selling off aggressively. That divergence told us the market was reassessing the timeline, not the technology.
Why the data center story moved too fast
I argued that there are two realities investors are only now starting to price in. First, most AI companies are realizing they are not going to make meaningful money for several years. They will need to invest hundreds of billions of dollars before returns show up, even if some of that capital comes from private equity or bank financing. Second, building data centers is slow, politically complicated, and infrastructure-heavy.
Lee and I discussed the growing resistance to data centers at the local level. Water usage, electricity consumption, and rising power rates are becoming flashpoints. A single large data center can consume as much electricity as an entire country like Luxembourg. That kind of demand requires massive upgrades to grids, transformers, and transmission systems, many of which already face shortages and multi-year lead times.
The myth of rapid AI saturation
We both pushed back on the idea that the world could be blanketed with AI data centers in a year or two. Even with unlimited money, it simply is not possible. Transformers alone can take years to source and deploy. These bottlenecks mean that the AI revolution, while real, will unfold far more slowly than early bulls assumed.
Lee noted that AI enthusiasm has been building for roughly three years, dating back to OpenAI’s release of ChatGPT in late 2022. The narrative focused almost entirely on how AI would change everything, and we both believe it will. But as history shows, transformative technologies also produce waves of failed companies. Just as many internet pioneers disappeared after the dot-com boom, AI is likely to see a significant thinning of the herd within the next two to three years.
Valuations and the coming shakeout
I said bluntly that two years from now there will be blood in the streets. Competition is intensifying rapidly. Alphabet (NASDAQ: GOOGL) | GOOGL Price Prediction is already catching up in key areas, and new models from Anthropic and others continue to enter the field. OpenAI is not going out of business, but whether it ultimately justifies valuations approaching $500 billion remains an open question.
The China factor investors underestimate
One of the most important risks we discussed is China. Lee referenced recent reporting highlighting China’s ability to deploy power infrastructure at extraordinary speed. Solar arrays, nuclear plants, and effectively unlimited electricity give China a structural advantage in scaling AI data centers. I emphasized that great technology alone is not enough. If China can build data centers in every region without regulatory friction, US AI valuations could face serious pressure from overseas competition, not just from domestic rivals.
What investors should do now
Lee made the point that companies like Google and Oracle will survive, but the return on the staggering amounts already invested will take time. I agreed and offered practical advice. If an investor has tripled their money in Nvidia (NASDAQ: NVDA), they should take some profits. Half is a perfectly reasonable number. Markets have a long history of sharp corrections that erase gains faster than people expect.
We also touched on utilities and even speculative ideas like space-based solar power, but the conclusion was straightforward. AI is real, transformative, and here to stay. The timeline, however, is longer, more expensive, and more competitive than the market originally priced in. For investors sitting on large gains, caution and discipline matter more now than enthusiasm.
Transcript:
[00:00:04] Doug McIntyre: So, Lee, we’ve had a whole bunch of stuff go on in the world of AI in the last week, and not all of it’s been good. We have some of these weeks where everything is good. What’s happened in the last week or so that’s really not as good?
[00:00:19] Lee Jackson: Well, I think the most dynamic thing is, and it has really kind of pushed the whole sector down, and we’re on, we’re talking on Friday, in the week when Oracle (NYSE: ORCL) still posted, by most tech companies’ standards, strong numbers. Oh God. Yeah. They were great. It would’ve been really, really good.
[00:00:38] But Oracle is not most tech companies, and they were basically, people are starting to think that they didn’t hit the revenue numbers and now they’ve pushed out the data center readiness to 2028. Uh-oh. That’s not what everybody had on their scorecard.
[00:00:57] So I think it has really put a chill in the whole sector, and as today you can see, as the Dow Jones Industrial Average streams to new highs, the NASDAQ just gets absolutely hammered. Everything that’s even related to the industry is getting hammered. So I think it’s kind of, I guess the question is, is this just
[00:01:21] a hiccup in the bigger AI story, which I’m sure the real big AI bulls would say, or, is maybe the whole AI data center universe moving a little too fast?
[00:01:36] Doug McIntyre: I think there are two parts to it. Number one: I’m pretty sure most of these companies are starting to look at themselves and say, I’m not gonna make any real money on this for a couple of years.
[00:01:49] I’m gonna, I’m gonna have to invest literally hundreds of billions of dollars. I may get some of that from private equity or banks, so I may not be playing with my own money. The other thing is, and I want to stress this, it takes a long time to build a data center, and there are a lot of hurdles.
[00:02:08] One hurdle now is a lot of people don’t want data centers near where they live because of the water usage. But more importantly, it’s only natural: if a data center uses as much electricity as basically the entire country, the entire town or the country of Luxembourg, right? Exactly.
[00:02:31] People say to themselves, how can my rates not go up? Right. And it’s, you’ve gotta build that infrastructure. Things like transformers, which people say, oh, who cares, you need transformers to do this work. There is a shortage of transformers right now. Right, right. The lead time on building those transformers is a year or two, maybe.
[00:02:57] Yeah. So the whole idea that you’re gonna have the world plastered with data centers in a year or two, it’s not true. Even if all the money were available, it’s not gonna happen.
[00:03:09] Lee Jackson: Yeah, I don’t think it will either. And it was interesting, some of the people that were so bullish at the outset, and remember this has been rolling for like three years.
[00:03:22] OpenAI introduced Chat in, like, November of 2022. Yeah, so it’s been rolling for three years. But again, it was rolling mostly on the idea of how AI will change everything. And it will be a profound change. It’ll be more profound than perhaps the internet was 25 years ago.
[00:03:44] But it’s gonna take a lot of work. And as you and I both know, because we’ve been in the business so long, there were a lot of promising companies back then that are gone. Where’s Netscape? Some of the other big names from 20 years ago? They’re not around. There’s competition in that range, but I think there’s gonna be a real thinning of the herd in the next two to three years, certainly within two years, next year in 2027. I think there’ll be a big thinning of the herd and we’ll be looking at some different companies.
[00:04:17] Doug McIntyre: Two years from now, there’s gonna be blood in the streets. Very, very soon. Yeah. I think so. You’re starting to see Google’s (NASDAQ: GOOG) product catching up with OpenAI. Is OpenAI gonna go outta business? No. No. Is it worth $500 billion? Did the last people who put the money in that final round, are they gonna get their money back?
[00:04:37] I don’t know. You’ve now got, it’s unbelievable technology, but people are getting in too early.
[00:04:47] Lee Jackson: Yeah. And again, they’re all constantly iterating, ChatGPT came out, they have a new version five or whatever, but Gemini’s got a new one and there’s a new Claude from Anthropic.
[00:04:59] And so, they’re not all gonna make it. They’re not all gonna make it. The question is for the betting man and the investor: who stays and who goes in the next two or three years?
[00:05:10] Doug McIntyre: There’s another problem here. People don’t talk enough about, there was an article in the Wall Street Journal yesterday about the fact that China has effectively unlimited electricity because of the structure of its system.
[00:05:20] Yeah, of course. They say that their solar arrays span the mountains and that they build a nuclear plant in less time than it takes to get a permit in the United States, was one of the comments I believe. So we need to worry in the United States, and US companies in AI need to worry, about whether China is in the process of flanking us, because it isn’t just important to have great technology.
[00:05:49] It’s important if you can have virtually unlimited data centers, I mean a data center in like every village. I’m very worried about that. Unlike a Walmart. I am very worried that the valuations of US artificial intelligence companies are going to be hit by China. Yeah. And not just by competition in the United States.
[00:06:10] Lee Jackson: Yeah, I think you’re right. And again, Google, they’ll be around. Oracle, they’ll be around. But I mean, the sheer amount of money they’ve already put into this is staggering. It is staggering. And the return on equity and the return on investment, it’s gonna take a lot of work. It’s gonna take a lot of work.
[00:06:30] I mean, if I was long all of those kinds of stocks, with the exception of maybe Google, I’d trim some of it. And clearly people have already trimmed some Oracle ’cause it just exploded higher earlier this year. So there was a lot of profit taking there. And I read that Ellison lost $25 billion in like two days, or some staggering number like that.
[00:06:51] But I wrote that story. Yeah, I wrote that. Okay. Yeah.
[00:06:56] Doug McIntyre: The other thing about AI that I think investors, here’s my advice to somebody: if you’ve tripled your money in Nvidia (NASDAQ: NVDA), take some off the table. Take half. Half is a great number. Don’t assume it’s gonna keep going up. Lee and I have been around long enough to see at least five or six, and I don’t mean little corrections, I mean real corrections.
[00:07:23] Yeah. You hit that with Nvidia, it could be gone in five months. So yeah, it,
[00:07:28] Lee Jackson: it really could. It really could. And when I was a salesman at Bear Stearns and Lehman Brothers, I had people that made big money in Red Hat and all those names that came out then.
[00:07:41] And I recommended JDS Uniphase, if you remember that company. I had a client that had 250,000 shares of it because he was the first US employee. I told him flat out, I said, sell half. He did, but he kept the rest of it. And it went from a hundred and a quarter to two. So I think Doug’s advice is something to be counted on.
[00:08:02] Doug McIntyre: Yeah. I still like some of the utilities because, yeah, I know these guys talk about owning their own infrastructure. There was a story yesterday that there’s a company that’s going to put solar arrays on satellites in outer space and then beam the energy down.
[00:08:21] Lee Jackson: To the world. I read that too.
[00:08:23] In fact, I have that in my notes. And I thought that was outstanding, that they can grab power out there. Well, it’s because there’s no clouds.
[00:08:35] Doug McIntyre: Yeah. The explanation was that they like it because there aren’t any clouds. Yeah.
[00:08:39] Lee Jackson: Aether Flux is the player in space and their satellites will beam solar energy to Earth.
[00:08:50] Right. Yes. Well, I don’t know how they’re gonna do that, but that’s what they said. Don’t hold your breath. Well, anyway, I think on a final note on this, and we’ll get back to utilities soon, is: boy, be careful. And if you’ve got a huge staggering profit, take half, take half, or at least take a third.
[00:09:09] Yep. I agree.
Top Gaining Stocks
Top Losing Stocks