The AI bottleneck

Editor's note: The following text is a transcript of a podcast story. To listen to the story, click on the arrow beneath the headline above.

JENNY ROUGH, HOST: Coming up next on The World and Everything in It: The Monday Moneybeat.

NICK EICHER, HOST: It’s time now to talk business, markets, and the economy with financial analyst and advisor David Bahnsen. David heads up the wealth management firm the Bahnsen Group, and he is here now. Good morning to you.

DAVID BAHNSEN: Well, good morning, Nick. Good to be with you.

EICHER: Well, David, stocks seemed pretty optimistic, or did, as we went into the holiday weekend. At the same time, there was a story last week I wanted to get your thoughts on: long-term bond yields and the fact that they have climbed to levels we have not really seen since before the financial crisis. Can you talk about the significance of that story?

BAHNSEN: Well, it’s a big deal for investors if it were to stay, just in the sense that a higher bond yield means that risk-asset valuations have to come down.

The reason why there’s a lot of reason to believe these higher bond yields are not going to hold is that risk assets have not been dropping; they’ve been going higher. That’s very different than what you’d historically see, and it gives at least some prima facie evidence that the more likely reason bond yields are moving higher is the short-term phenomenon of higher oil prices.

Bond yields and oil prices have been very, very correlated directionally for the last two, two-and-a-half months. 

We don’t issue 30-year bonds anymore, so these are all older. There’s a lot less volume of the 30-year to get a market signal from.

The 10-year is the most important price in the world. Basically almost every asset on earth, globally and domestically, is priced off of the 10-year bond.

It is the number that I think indicates expectations for nominal GDP growth in the U.S. The 10-year’s moved up. It’s at four-and-a-half, but it’s been above four-and-a-half four different times in the last few years.

Four-and-a-half is a very minimal number. It’s around the level I’d like to see. I’d like to see inflation down to one-and-a-half to 2 percent, and I’d like to see real growth at two-and-a-half to 3 percent. So those two put together get us to about four-and-a-half.

Before the war started, it was a little over four, and it moved up to about four-and-a-quarter. In the last couple of weeks it’s gone from four-and-a-quarter to four-and-a-half.

We have a lot of data coming in the months ahead, and some of this is good, Nick, in that the employment data has been a little bit better than expected.

If you get bond yields dropping, that is a much bigger question, and it usually means that there’s an economic weakening. Right now I don’t think bond yields are going higher because of economic strengthening. I think they’re going higher because of oil prices.

That’s a more short-term phenomenon, presumably around the Strait of Hormuz, and therefore we need a little more time to digest this.

EICHER: All right, David, your Dividend Cafe this week really turned data centers into an infrastructure story and away from simply a technology story.

Communities all over the country are starting to push back, and push back really hard against these projects, and I wonder whether there’s a risk that AI development is going to start running into the same sort of not-in-my-backyard, NIMBY resistance that we have seen with oil pipelines or even things like nuclear power plants.

BAHNSEN: Well, there’s two different things there, Nick. Forgive me, but I want to parse this out the right way for our listeners, because people could resist data centers for stuff that has nothing to do with AI at all. It could literally be about the data center.

I think so far the resistance to data centers has almost entirely been immaterial to what is happening inside the data center.

It’s mostly that people don’t want the power depletion that it represents, the water resources, and then the eyesore of a massive digital warehouse—a huge physical facility that has a lot of digital infrastructure inside of it.

There may end up being a second tier, which is people resisting data centers because they’re resisting AI and what AI could end up doing to jobs, or what AI is doing in terms of deep fakes and other sinister activity.

I would love to say that all of this can be tackled all at once, but I don’t think it can be. I think each community has to deal with what they want their relationship to be with huge new warehouse facilities independently around their own zoning laws, their own entitlement, and their own voter preferences.

What I do not want, Nick, is the federal government to come in and intervene as to what different cities, counties, and localities ought to be doing.

I also think the data-center conversation and the AI conversation need to be separate, because if we end up getting a fair acceptance—I’m not a NIMBY guy whatsoever—if the only hostility to data centers were, "I am old and retired and don’t want this thing around me," that’s not good enough.

It’s all going so fast that I don’t feel society is wrestling with it the right way. The political side of it is totally confused.

We need to go about this one step at a time with first principles, and that should be attached to data centers as much as it should be attached to AI technology.

EICHER: David, there was another story last week that caught my attention, and it had to do with the coming SpaceX IPO.

For all of the futuristic technology involved in something like this, I was struck in the Wall Street Journal story about how the actual process of an initial public offering really sounds very old school—bankers on phones deciding who gets shares and who does not.

Could you take this opportunity, with that coming sometime in June with SpaceX, to walk us through how an IPO works and what it says about how markets really function?

BAHNSEN: Well, there are a lot of things about this that will be similar to other IPOs because they have to be, but there’s plenty that’s very, very different.

For one thing, this will be the largest size of an IPO in history and the smallest amount being given to the public in history. It’s set up in a way to create an imbalance between supply and demand, which makes prices jump higher and unfortunately leaves some people holding the bag.

I think when you’re talking about a company that they want to value between one-and-a-half and $2 trillion, and they’re going to bring to market $50 to $70 billion, there’s just going to be a very limited float, which is what it’s called—the amount available for the public to buy.

With very high demand and very low supply, it pushes the numbers up, and I think this is not always going to be the most sophisticated buyers dealing with that supply-demand imbalance.

That said, you’re certainly right. Investment bankers have to get on the phone and secure orders, because even if the percentage of the float is very, very small, $50 to $70 billion is still a lot.

Remember, when Amazon went public they brought $25 million—with an M.

Now that was a long time ago, but the largest IPO in history was Saudi Arabia bringing Aramco, their national oil company, and it was a fraction of this.

A typical IPO is usually dealing with a billion dollars, not $50 billion or $70 billion.

You’re going to have big hedge funds and big institutional investors having to take a lot of the size down, and that requires a lot of bankers.

This is not just one company leading it. You’re going to have almost every investment bank on Wall Street talking to a lot of investors, and not just U.S. investors. There’s going to be heavy foreign ownership of this as well.

It’s going to be a big, elaborate process.

I should disclose, just for the sake of disclosure, that I’m a shareholder and have been for a number of years, and a lot of our clients are.

My opinion on the process is that it’s Elon doing it the way he wishes he had done Tesla to begin with. He’s going to be fully in charge of everything.

Anyone buying in the IPO is not going to have any of the traditional shareholder rights. He’s going to have ten-times voting shares. He can’t be fired. There’s all kinds of stuff they’re doing, but that’s fully disclosed. Any investor buying it will know that.

They’ve incorporated in Texas, which is a little different than what Tesla was in Delaware.

Let’s remember, SpaceX is no longer just SpaceX. They merged with xAI, and so you have basically Twitter and Grok and SpaceX all together.

SpaceX itself owns Starlink, which is where most of the revenues come from.

So you have a telecom, satellite, social media, AI, data-center-in-space, rocket-ship play all merged together.

At the end of the day, investors either want to own Elon’s conglomerate of companies or they don’t.

EICHER: All right. David Bahnsen is founder, managing partner, and chief investment officer of The Bahnsen Group. He’s a writer for WORLD Opinions and he writes each week at Dividend Cafe. Hope you have a great holiday-shortened week, David.

BAHNSEN: Thanks so much, Nick.

WORLD Radio transcripts are created on a rush deadline. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of WORLD Radio programming is the audio record.