Artificial Intelligence

Second Quarter Report 2024

The first half of the year has proven profitable for our accounts. We have been more cautious this year than in the past, and continue to be so even as the market has rallied. For the first time in nine years, we have raised cash by selling the bank stocks we own. We will elaborate on this decision in this letter, along with some thoughts on artificial intelligence.

Why sell the bank stocks? We have been owners of bank stocks for over a decade, and now we are not. Back when we owned them in 2009, Wall Street thought the world was headed into a depression, and the residential real estate holdings of the banks were going to be wiped out. We disagreed with that assessment and bet accordingly. Today, Wall Street thinks that commercial real estate will be okay, and the banks will dodge that mess. We aren’t so sure.

It is estimated that there will be $1.3 trillion in commercial real estate coming due in the next 24 months. Of that amount, roughly half ($670 billion) is viewed as distressed and in need of additional investment or a write-down. If someone is to increase their investment in a commercial office building in a large city today, they have to answer some questions about that commitment. Will workers return to the office in large cities? Will the population shift to the southern parts of the USA reverse? Will the taxes needed to be raised to support the budget deficits of the large cities make these investments not profitable? With interest rates up 4 points in the last 3 years, will the repricing of the debt make the investments not viable?

When we reviewed these questions, we came to the conclusion that we don’t know the answers to these problems. We think it’s likely that developers will walk away from their buildings rather than put up more money. In Los Angeles, the AON building sold for 45% less than it was purchased for in 2014. The second tallest building in Minneapolis was handed back to the lender. The 1760 Market Street building in Philadelphia sold for $11.5 million, which was $20 million less than they paid for it in 2018. These are not isolated sales; it’s a problem that isn’t going to be solved by ignoring it. The due bill is here, and we don’t know how it’s going to play out.

One last thought on the banks from a macro standpoint. We think that our economy is in stagflation. A common definition of stagflation is high inflation and low growth. We think that’s accurate. If you are a bank, you will be making fewer loans while inflation keeps your interest rates higher than you want them to be. We don’t view that as a good place to be right now. The distress from commercial real estate will be felt by the smaller banks first. Those are the ones you should watch to see how this plays out.

Artificial intelligence is the hot investment this year. The hottest of those investments is Nvidia. They make chips that help you get your answers quicker. We don’t own that stock, but we own a few that will benefit from AI. In a standard AI program, you have what is called training and inference. On the training side, you feed all of your data into a computer and it calculates the answer using AI as the software.

Once the answer is calculated using your data input, the inference part of the equation is using that knowledge to make predictions. Using AI involves massive amounts of data transfer over the internet and energy use to calculate the answer. The International Energy Agency (IEA) forecasted that global data center electricity demand will more than double from 2022 to 2026, with AI playing a major role in that increase. This energy demand is going to be difficult to supply. Currently, in Ashburn, Virginia (outside DC), a data center cannot get additional power until sometime in 2026. All investors are looking for cheaper power or any power at all.

We have a very large position in Tesla and Apple. Tesla has a huge battery and storage business that will benefit from power shortages, which we think are a real possibility. Apple will benefit from the unique data they have from their end users when they do AI training. What areas that investors have not thought about are the secondary effects of higher energy usage. We own a lot of oil stocks. The demand for oil as we consume more energy is going to be way above the trend line. We will need to increase our consumption of hydrocarbons. To build the faster data centers, we will need steel and the copper that goes into the power lines. We own Nucor to profit from that.

Finally, our last beneficiary who may be the biggest winner from AI is Lumen. You need fast fiber to transfer data from the data centers to the end user. They have it, and it is going to help their bottom line a lot. Lumen has been a crummy investment for us. We think they get a lifeline with AI.

Sincerely,

Mark Brueggemann IAR Kelly Smith IAR Brandon Robinson IAR