Hey, What Just Happened? And Some Model Portfolio Updates

“All Warfare is Based on Deception” - Sun Tzu

I try my best to keep these macro musing posts to the bare minimum, but today’s market action was worth discussing.

A Chinese hedge fund released a large language model that caused a bit of a market reaction yesterday. The quality of the LLM was high, and an enormous number of market participants rushed to download the app to verify this. The hedge fund released a white paper claiming that the cost of production was extremely low, this is unverifiable. NVIDIA sold off 16.97%, evaporating about $490 billion of market capitalization. By the end of the day, downloads of the new LLM were restricted to Chinese mobile numbers, because it is always prudent to leave an audience wanting more.

Not only did NVIDIA sell off, but stocks that are Artificial Intelligence beneficiaries sold off as well, including Vistra Corp, a utility, down -28.27%, Oracle, one of the hyperscalers, fell 13.79%, the biggest one day drop since 2002. The list goes on and on, but as for my own portfolio, both ProFrac Holdings (ACDC), and New Fortress Energy (NFE) were down substantially. New Fortress Energy has almost no exposure to AI, and ProFrac Holdings has yet to receive any fundamental benefit as frack fleet counts are still falling nationwide. But when people are over leveraged and get margin called, everything gets sold together.

Applied Optoelectronics (AAOI) was down 20%, and that is obviously connected to the data center buildout. Centrus Energy (LEU) was down 15% today, but again, not a single nuclear reactor has been built yet due to the artificial intelligence revolution.

What will happen next? Probably a bounce, but a bounce typically doesn’t retrace more than 50% of the drop, that’s typically how these things work. If we don’t get the bounce on Tuesday, Jerome Powell speaks on Wednesday, and the market is pricing in about a 25% chance of an interest rate hike in 2025. Jerome Powell is likely to dispel those rate hike fears, and the market should respond favorably to that at least. I could be wrong, predicting day to day is a fool’s errand, but that is the mostly likely path for the week.

Was the Chinese large language model really made cheaply? - Probably not, I don’t remember the part of Sun Tzu’s Art of War where it was imperative to tell the truth to one’s adversaries. I remember Sun Tzu advocating for the opposite. Last quarter, 22% of NVDA revenue came from Singapore, which was just a backdoor to get around the export restrictions to China from the Chips Act.

Even if the Chinese LLM was made cheaply, does it mean that the AI revolution is over? No, probably the opposite. Marc Andreessen called this AI’s Sputnik moment. We are probably about to engage in a new Manhattan Project, or a new Space Race, and that means the US will spend more on AI for national security reasons.

In Trump’s first term, China was a meme to poke fun at. I wonder if the American public is waking up to the reality that we have a new peer rival, and that two tigers can’t live on the same hill? But unlike the Soviet Union, whose economy was a tiny fraction of ours, China is a manufacturing powerhouse who has leapfrogged the US in automobiles, drones, nuclear energy, and moving forward probably robotics as well.

The United States isn’t accustomed to being the underdog, but perhaps this will give us some sense of National purpose and unity, as opposed to the last fifteen years of infighting over luxury beliefs. If this really is a new Cold War, people are really underestimating the seriousness of Trump’s targeting of the Panama Canal, Greenland, and even Canada. If that is what it takes to stay ahead of the Chinese, nothing is off the table. With a renewed Monroe Doctrine, Latin America becomes even more interesting to invest in.

Jevons Paradox:

One of my favorite economists, William Stanley Jevons, had his name dragged across all of Twitter today for his findings from “The Coal Question (1865).” In it, Jevons settled peak coal fears with the argument that as coal becomes more dear from mine depletion, more effort will be spent to increase the efficiency of its extraction, prices will fall rather than rise due to new innovations, and the volumes will increase rather than decrease. This proved to be entirely correct, and yet for my entire life I have not been able to escape arguments in favor of peak oil. Even today, the decline of the Permian Basin seems to be consensus. Just know that if innovation makes tier 3 rock profitable, in the same way as it has done to tier 2 rock, there is so much tier 3 rock that we will not run out of shale for many more decades to come. Also, of the rock already drilled, there are estimates that between 50% and 90% of the hydrocarbons are still trapped and waiting for new techniques to be extracted.

But Jevons wasn’t brought up to deal with the peak Permian theory, he was brought up to deal with the AI data center buildout. If it really is true that the Chinese LLM uses 2/3rds of the electricity as the American LLMs, does this destroy the natural gas and nuclear narrative? No, because as costs fall, we will use more AI, and likely need even more natural gas and nuclear energy than before. Jevons has been correct for 160 years, I doubt that I am lucky enough that this is the precise moment where he is proven wrong.

It’s a bit of a tragedy that the whole world is talking about Jevons without taking a moment to get to know the man. In an era of British ethnocentrism, he had such a genuine love of the science of economics and all of its practitioners, that he corresponded across continents just to share ideas with the French and the Austrians. And when sharing ideas, he gave credit to sources, and made introductions so that other scientists could network. Citing sources was not common at the time, not because everyone was a plagiarist, but because it was assumed that men of letters would know the sources anyway. Jevons was good natured, significantly improved the camaraderie and the culture of economics, and died well before his time.

Free Trade?:

Buried underneath the AI news story was the confirmation of Treasury Secretary Scott Bessant, who unveiled the policy of starting with a 2.5% import tariff to raise revenues, with the plan of increasing these broad import tariffs by 2.5% every month until the target of 10% - 20% is reached. This gradual approach might not goose the market to the same extent that a sudden tariff would. Trump confirmed the strategy along with floating the idea of lowering or repealing the income tax.

While the corporate media is claiming that tariffs are evil, inflationary, and a tax on consumers, if income taxes are replaced dollar for dollar with import tariffs, on the margin this would lead to higher productivity, lower consumption, greater investment, higher domestic wages, etc. It’s a policy that would have more popularity if it weren’t for the leftwing embrace of internationalism, and the academic economic dogma of free trade. This situation reminds me of the time that John Maynard Keynes advocated for tariffs, and he was chastised by Lionel Robbins because, while it is true that sometimes tariffs are beneficial, you never tell that to a politician who will not have the judgement or discretion to use them only when appropriate.

The Political Economy Club founded in London in 1821 by James Mill, David Ricardo, Thomas Malthus, and Robert Torrens was originally a place that was open to exploring Torrens’ idea of reciprocity among trade. But critics argued that over time, the intellectual contribution of the club waned until the members could mostly be replaced by parrots, provided one taught the parrot how to say, “free trade.” This transition from exploring the nuances of trade to dogmatically repeating that free trade is unambiguously good still haunts the Economist Magazine in particular.

Be prepared to see article after article presenting free trade as unambiguously good, and purporting to have the weight of scientific consensus behind it. But know that brilliant economists have spent hundreds of years debating the issue, there are clearly nuances, and we might be in a situation right now where free trade is not the answer.

Model Portfolio Category Updates:

Logistics and Industrials:

One sector that rallied today on the overall market selloff was Logistics. If we are switching from income taxes to revenue tariffs, then logistics will be a huge beneficiary as supply chains shift from international ports to domestic land routes. And the boost to logistics will come quickly as orders are placed to front run the new escalating tariffs. In the model portfolio I had a 5% weight on Logistics for 2025 due to the end of the Destocking cycle, which does appear to finally be upon us. I thought Logistics would be improving by the second half of 2025, but it appears to be happening faster than I had predicted.

It is still early to know if Trump will accomplish changing out the income tax for tariffs, but with support for domestic manufacturing, the retirement of Baby Boomers, and the deportation of cheap labor, we will be confronted with a very tight labor market. Robotics and automation will have a huge demand increase for a decade at least. I believe the only company in the model portfolio with exposure to robotics right now is Alta Equipment Group (ALTG), and not much of their inventory is robotic at the moment. I have been on the lookout for more robotics or automation exposure, but I either find outrageous valuations, or a limitation of my own competence.

Silver Tsunami:

Also previously in the model portfolio, I had a category called Financials, but upon reflection, it is much more specific than that. The US is confronted with a “Silver Tsunami” of Baby Boomers retiring with their $78 trillion of assets as of the last measurement. With higher sustained interest rates, this can provide a significant income. About $40 trillion of that wealth is in financial instruments, which at a conservative 4%, is $1.6 trillion of spending annually which will go toward the Baby Boomers’ favorite diversions, whether that be Cracker Barrel or Carnival Cruise Lines.

Historically, retirees have been so afraid of outliving their money, that they would spend the gains on the account only. For example, for someone who retires at 65 with a 401k worth $500,000, if you check the balance of the account ten years later at age 75, the most common balance would be $500,000. The Baby Boomers are known as the selfish generation, but time will tell if their behavior is much different than their parents.

I had previously exited Jackson Financial (JXN) after it ran up from $35 to $90. But on further reflection, I have reopened the position. Jackson Financial, Finance of American Companies (FOA), and now Abacus Life (ABL), are a very interesting group of companies to take advantage of the Silver Tsunami. You have the market leader in variable annuities, the market leader in reverse mortgages, and the market leader in life insurance buyouts.

Expect a more complete updated model portfolio soon, it has been requested. And expect an industrials roundup, there are just so many companies that I have covered, it’s worthwhile to compare their current valuations. That includes Custom Truck One Source (CTOS), Alta Equipment Group (ALTG), Titan International (TWI), Titan Machinery (TITN), Mattr Corp (MATR.TO), Forum Energy Technologies (FET), and possibly the logistics names as well.