Warren Buffett is fond of saying, “This is such an interesting movie, I can’t wait to see how it turns out.” Warren is really good at understanding the interrelationships of apparently unrelated events and taking advantage on behalf of Berkshire (BRK.B). On the other hand, one of his superpowers is the ability to distinguish between the most important things and the interesting noise. We will try to be as discerning in delivering this quarterly “review” of the investment movie.

One of the loudest and most distracting subplots of today’s feature is the domestic political scene. (With the midterm elections looming in November, the volume will not go down from here.) Washington provides plenty of dark humor , but the polarization of Congress and the resultant abdication to the executive branch (which did not begin with this president) is not funny. No comedy here. The country needs solutions to a number of serious problems (debt and deficits, immigration, healthcare, education, poverty...), but the inability to compromise or to give the other side credit for anything makes this aspect of the movie a tragedy . It just does not have to be this hard.

A year ago at this time, the biggest economic story out of D.C. was tariffs. After months of turmoil, the process quieted down considerably when China briefly played the “rare earths” card, showing that it could shut down global manufacturing at will. Then, as businesses and trading partners were settling into the new reality, the Supreme Court ruled the tariffs illegal. Although courts at various levels, including the “big” one, have spent a lot of time and energy on this, it seems more like a legal slog than a thriller .

We entered 2026 with multiple wars in progress—Ukraine, Gaza, Sudan—and at the end of February, the president initiated a war with Iran. The process was unusual in that neither Congress nor our allies were consulted. It was expected to be short—a few weeks—and the shooting may have stopped by the time this letter is published. The immediate investment implication is disruption of oil and gas production and transport, along with a spike in energy prices. Interruptions to supplies should be temporary, but in an ironic twist, the administration is lifting sanctions on the sale of Russian and Iranian oil, even as they attack us and our allies, in order to obtain oil for China and others. The long-term consequences of this part of the movie are TBD.

The Federal Reserve makes periodic cameos. They wrestle with their “dual mandate” of maximum employment and stable prices, and their tools are limited. As we have written before, we think the attention paid to the short-term Fed Funds rate is mostly wasted effort. Long-term rates , on the other hand, are very important since they drive asset allocation decisions for businesses and the valuation of long-term assets. Chronic budget deficits and growing government debt should theoretically lead to higher long-term rates—bad news for stocks—but the “inevitable” continues to be postponed.

All of these elements of the investment movie have implications for investors. There is plenty of uncertainty about what will happen next, and some of the possibilities are a little scary. For the most part, though, these are recognizable types of events . What we find fascinating, and what makes this a mystery , are developments in artificial intelligence . Elements of science fiction and fantasy .

Balancing the Plot and the Noise

Software Apocalypse

Every generation or so, a new technology captures the imagination of investors. The emergence of the Internet in the late Nineties set off the last speculative tech bubble. Imperfect (at best) understanding of the technology and vivid imaginations made everyone's best hopes and worst fears seem possible. Many of the early business applications failed and much of the capital spent on infrastructure was not needed, at least in the near term.

But the "big idea" eventually turned out to be even bigger than expected. In the 10–15 years after the bubble burst in 2000, the Internet changed almost every aspect of life and the overbuilt fiber-optic capacity turned out to be valuable. Businesses whose profitability models depended on keeping their customers in the dark as to price or product alternatives were destroyed. Online retailing began taking share from "brick-and-mortar" stores and the shift continues. Email and other means of sharing information electronically has changed personal and business life. Being right, but early, was expensive. Denying the change was terminal.

Today's tech "miracle" is AI. In the three years since ChatGPT was introduced, new models have been released in rapid succession. AI capabilities have progressed from glorified web search to models and agents which can do coding and other tasks better and faster than the humans they "serve." Some can apparently train and evolve themselves (!). I am over my head describing these phenomena, much less explaining them, but the pace of development is astonishing.

The explosion in AI raises lots of questions for investors. Capital expenditure plans for the creators and providers have taken quantum leaps. Announced orders for chips and the data centers to house them have topped $1 trillion for 2026. The required permits, power supplies, water, etc., have created logistical and political complexities that seem destined to cause bottlenecks and delays. A much bigger issue is whether the customers will be willing to spend enough on AI to provide a good return on these investments.

Another question—one which has already had a major impact on stock prices—is whether the power of AI will make incumbent software and business service providers obsolete. Will businesses still call on consultants for advice and software implementation services or just "ask ChatGPT?" Will companies still need enterprise software, and if they do, will they need fewer "seat" licenses because AI has made a smaller workforce more productive? The fear that today's leading software could be eclipsed by an upstart evokes the image of explorers of Columbus' era sailing along under clear skies but fearing they might suddenly fall off the edge of an Earth that turned out not to be flat after all.

In response to this uncertainty, investors have opted to "sell first and get answers later." (Hence the catchy term, "software apocalypse.") There is also evidence that employers are deferring hiring decisions until they can assess whether AI can be used to replace some of their humans. Recent high-profile layoffs have been blamed on AI. (Sign of the times or convenient excuse for correcting previous hiring mistakes?)

Of course, the answer to all of these questions about returns on capital investment and competitive positioning is, "It depends." It seems inevitable that capacity will be overbuilt and that some weaker competitors will suffer. On the positive side, it also seems logical that established consultants and software providers have the advantages of incumbency—domain knowledge, technical and financial resources, and customer relationships—that should insulate them from potential disruptors. As for the future of the job market, the optimistic case is that AI will serve humans well—relieving them of tedious, unrewarding tasks and freeing them for more interesting and valuable work. Let's hope so.

It would not be an "apocalypse" without at least a little collateral damage. Our portfolios have been on the receiving end of some of the "precautionary" (panic) selling. We believe the fears are misplaced, or at least overdone. Nevertheless, when sellers abound, the stock doesn't argue—it just goes down. This has contributed to our investment results being disappointing in absolute terms and relative to the indexes driven by the "AI 'winners.'" The Portfolio Managers' Commentaries provide details on portfolio winners and losers. The only

Balancing the Plot and the Noise

consolation, but an important one, is that the problem stocks have suffered much more from investor sentiment than from fundamental business disappointments. Intact and growing business value should eventually be rewarded.

This part of the movie really is interesting, if not always fun. We are learning a great deal about AI, though as my son answered when asked if he had had fun at debate camp, “I learned a lot and it was good for me.” Fortunately, I have some younger colleagues who have real aptitude for understanding both the hardware and software. Having said that, I think we all realize that the more we learn, the more we know how much we don’t know—and can’t know yet. We are approaching this whole subject with as much humility as we can muster.

So, the movie keeps rolling. The plot of the war in Iran (and the surrounding neighborhood) has changed directions a couple of times between drafts of this letter, and who knows what will happen between submission and publication. A larger question is whether the stagflation monster will be awakened by tariff and oil price inflation and ongoing softening of the job market.

The good news is that while the movie continues, we do not need to react to every new scene. The pace of technological development does force us to tweak our 5-year valuation models more often these days, but our basic m.o. is to buy very good businesses at reasonable prices and let them grow and compound value. Even in a faster-paced world, we think that makes sense. Our core belief is (still) that “The (business value) truth will out,” even if the proliferation of passive, algorithmic, momentum and “pod shop” investors makes it take longer for the stock price to reflect the value.

Original Post

Editor's Note: The summary bullets for this article were chosen by Seeking Alpha editors.