Regarding… Vol 3 no 29 – Before The Flood

One of my favourite plays on AI and its associated energy requirements is not doing well:

Despite having more than halved from its recent high of $11,760 per share, it is still a 20-bagger over the past 10 years. So, there’s that. Over that 10-period period, it has halved 6 times. So, there’s that, too.

But why do I like it, and why is it so weak now?

TPL owns vast tracts of land in the Permian Basin and earns royalties from companies that want to drill (frack) for oil on its land. As such, it is a capital-light business with operating margins that hover around 80%. However, the market recognises the superior business model, and it is trading at a high valuation, with a P/E ratio of over 40 times and an EV-to-sales ratio of 27 times.

It features three vectors of optionality.

• For every barrel of oil that is produced, 5 barrels of water come out. This water must be stored and treated, for which TPL earns a growing income stream. More importantly, this water can be used in cooling applications – such as those required for nuclear power stations or data centres, for instance. Here’s a nice article about this problem.
• TPL produces a large amount of natural gas, which is being flared off due to the lack of pipeline infrastructure to transport it elsewhere. This is inexpensive energy that could be used to power things like data centres, especially if the existing grid infrastructure becomes overloaded.
• TPL owns a lot of arid scrublands. Fracking happens below ground, but there is little use or value for the land itself – unless the NIMBYs start forcing datacenters to locate themselves in unpopulated areas. Areas that happen to be rich in cheap energy and cooling material.

I initially wrote about TPL in “Dodging the Curse” in October last year. Not much has changed in this short period, except that the company was admitted to the S&P 500 index, which caused a speculative increase and subsequent retreat. However, the core investment thesis remains the same. Here’s an article describing some developments in the Permian basin.

Key quote: “Backed by up to $1 billion from Five Point Infrastructure, PowerBridge is developing data centre parks in the Permian Basin with dedicated natural gas-fired power plants, broadband, and cooling infrastructure to attract technology firms. These projects include plans for wastewater desalination for cooling.”

My take: If you are bullish about AI, TPL could be a key beneficiary of the significant capital spending on data centres planned by big tech. Importantly, TPL itself does not need to spend very much at all to reap the benefits. The current low oil price and lingering doubts about the life of reserves in the Permian (which TPL says are still plenty) are creating a buying opportunity. TPL is a core holding in the “hard asset” component of the MWI Worldwide Flexible Fund (aka The Cockroach).

Last time I wrote about Palantir (in “Golden”, February of this year), I only half-jokingly said no one really knew what their business model was. Since then, the share price has nearly doubled, and more of us have become aware of it.

Simply put, they do AI software – a business model the market loves very, very much:

Palantir now trades at 80 times forecast revenue for the coming year. No big tech company has come close to that in the past 25 years. Tesla and Google’s parent, Alphabet, managed to achieve a peak valuation of about 20 times their revenue.

Put another way: when Meta Platforms had the $400 billion market capitalisation that Palantir now has, the Facebook owner was generating $40 billion in revenue. Palantir, in comparison, expects to generate just $5 billion in sales this year, albeit at a rapid growth rate. When Apple reached a $400 billion market value, its revenue exceeded $150 billion.

As they say in the classics: drive for the show, putt for the dough. Revenue is a good metric, but any business owner will want to know how much money was made. Over the 12 months ending March 2025, Palantir generated profit of $571 million, placing it on a P/E ratio of (almost) 1000x. When you get to these numbers, it may as well be 1000x!

But that’s not all, as they also say in the classics. For every $1 Palantir earns currently, it awards $2.50 in stock-based compensation to its employees – who probably don’t care at all about how much money it makes.

Who knows how this ends?

My take: As I said last week, this is just one more sign of speculative juices running freely in the US stock market. Maybe this is just Schadenfreude on my part, as I don’t own the stock. Maybe a P/E of 1000x (yes, that’s three noughts) is the new value stock?

John Maynard Keynes famously said, “When the facts change, I change my mind. What do you do, sir?”

China is one place where I have changed my mind. I did so publicly in “Schrodingers China” in October 2024. Last week, a subscriber asked me why. So, to save you the time of reading the whole of that letter, here is the simple answer:

1. Valuations normalised. The P/E of the Chinese market went from over 25 times in 2021 to as low as 9 times in late 2022. So, although the Chinese markets present the same risks – VIE entities, communism, sanctions, and regulatory overreach – you are now being compensated for those risks through a lower price.
2. I read extensively about the market, especially from non-Western sources. This prompted me to revise my views on many issues – for example, I began to consider it less likely that China would invade Taiwan. Consequently, my tolerance for Chinese assets grew.

My take: This month’s letter from Cederberg Capital explains the valuation development in Chinese markets over the past 4 years quite neatly. I don’t often recommend other funds, but if you have an appetite for Chinese exposure, Cederberg should be at the top of your list. If regulations allowed me, I would consider taking Chinese exposure for the MWI Worldwide Flexible funds (also known as The Cockroach) through them, as I don’t have the time or inclination to conduct in-depth, stock-specific Chinese research.

Paranoid Android was the hit song off Radiohead’s stereotype-busting album “OK Computer”. At the time, the song was played everywhere. Today, 28 years later, the music remains relevant, and most of us who enjoy contemporary music are familiar with it. If you listen to the lyrics, it’s clear Radiohead were way ahead of their time – some would even call them prescient.

But some people have never heard it (!). In this YouTube video clip, a classically trained musician analyses the song, having never heard it before. She brings a fresh approach to the music, and her take is interesting: “It’s kind of weird. It reminds me of the feeling I had when I learned my first Bach minuets. Some of them were in a minor key…something just felt strange about the whole tonal setup. I get the same kind of feeling here.”

Spot on!

• Many find banking dull, and history too. I don’t – both history and the credit cycle fascinate me. For me, there’s only one thing better than a book that combines these two subjects, and that’s a book on the subject written by James Grant. For those who don’t know, James Grant is the editor of Grant’s Interest Rate Observer (to which I highly recommend subscribing), a fortnightly newsletter on financial markets. Specifically, credit markets, but with strong views on equity markets included for good measure.

  James Grant writes well. His writing is wry and witty and makes the world of interest rates, stock markets, and market psychology come to life. His explanation for the driving force behind market prices is a case in point:

  “People in markets are more suggestible than a layman might imagine. Try as they might, they can never know the one thing they really want to know – that is, the future. Not knowing, they compare notes with others. They work in units, not alone. They are brave together at the tops of markets and meek together at the bottoms.”

  The book is filled with insightful observations on how human behaviour and psychology drive markets, rather than rational analysis or conscious risk management. As he says, progress is cumulative in science, but cyclical in finance.

  This book tells the story of all the different credit cycles since the mid-19th century, illustrating how cyclical finance is. Each cycle was slightly different, but each had common characteristics:

  • • • Lending standards were gradually relaxed
      • The assets which banks were prepared to lend against expanded
      • Regulators averted their eyes from bad practices
      • The asset side of banks’ balance sheets grew rapidly
      • “Democratizing credit” became a watchword

  Each time the cycle ended in tears, after which regulators shut the stable door long after the horse had bolted. Each time, new management teams at the big banks (having replaced the previous management teams) said “Never Again”.

  Rinse and repeat.

  What makes Grant’s recounting of the credit story so interesting is that he writes with a biographical style. He brings the most prominent bankers of each cycle to life. It’s the human side of the cycle that interests Grant, with which he makes it so easy for us to read.

• I highly recommend this book.

I’ve made a few mistakes over the past few weeks.

First up, I neglected to mention Kim le Court Pienaar. In a week where I was in Mauritius, and we celebrated Women’s Day, I neglected to mention that Kim – who is from Mauritius – wore the yellow jersey in the Tour de France Femmes, becoming the first person from Africa to do so. Chapeau, Kim!

Two weeks ago, in “Trade Offs”, I wrote about the shocking remuneration practices at Woolworths, the South African retail business. I included this table:

I was trying to illustrate how the CEO (and, by extension, management in general) benefits while shareholders lose out. However, my math was poor, as several subscribers pointed out. Earnings per share declined by 18% p.a. (not 4.4%) and CEO pay grew at 25% p.a. (not 7.8%) – as anyone with even basic maths should have been able to see. Except me. At least my argument was directionally correct, if understated. The reality is much worse.

Finally, last week in “The Car Issue”, I quoted the SA used car market as trading in 4 million cars per year. From which hat did I pull that number? Just writing that number down should have triggered alarm bells in my head. The SA market is 400,000 cars, only 1% of the size of the US market, not 10% as I unthinkingly wrote.

Thanks to all of you who pointed out these errors!

I need to be much more careful. And I would ask you, too, to be very careful out there!

Lastly, my stepson Zac is celebrating his 25th birthday today. Time flies when you’re having fun. Ask him. Congrats, Zac!

Piet Viljoen
RECM