2-5-2026
I feel like it needs to at least be mentioned at the outset, even though this post is supposedly about the financial markets, that we are in the camp of wanting some warmer weather. For those of us who are “wintering in place” in the Fort, it’s been a rough 3 weeks. We know, it hasn’t exactly been summerlike in Florida either…it was 10 degrees warmer in Seattle than it was in Naples on Monday. Even Ketchikan, Alaska had a higher average temperature Monday than Naples. If anyone reading this has a line on how some relief might come to pass, we will help you lobby for it.
Overall, the stock market has had a nice, smooth run since the Liberation Day crash in March and April, 2025, triggered by the Trump tariffs. Any volatility in the S&P 500 since then has looked pretty tepid compared to last spring:
Beneath the surface, though, a sudden chill has descended onto some parts of the stock market this week, namely software stocks, but also tech overall. Before we continue, let us say unequivocally that this article has a shelf life of 1 day, given how quickly things are moving in the market. If you’re reading this and it’s not February 5, 2026, it’s past the use-by date, so make sure it still smells okay before you ingest it.
Tech has led the market upward for the last 3 years, in part thanks to the boom in capital spending on AI data centers. There is still much debate over how much corporate America is getting out of new AI endeavors, with many surveys showing most are still not getting a return for their investment (in terms of dollars and time and attention) that the companies themselves deem adequate. It appears that companies will continue to try to automate more and more workflows, as evidenced by the large number of large layoff announcements in the last 2 quarters. In other words, they think they’ll see more payoff in terms of productivity in the future, as AI tools and software gets more useful and (hopefully) less buggy, so existing workers can work more efficiently.
There is no debate that tech companies have been spending like drunken bandits to build out data centers as fast as they can. These data centers are ostensibly needed to fulfill the data processing-heavy needs of more and more AI use cases…at least that’s the belief of the “hyperscalers” like Meta (Facebook), Google, Amazon, OpenAI and Microsoft. The numbers are staggering: Meta spent $70 billion in capital expenditures in 2025, up from $37 billion in 2024. They plan to spend $135 billion in 2026. Not to be outbid, Alphabet (Google) announced last night their capital spending will rise from $91 billion to $185 billion; it was $52 billion in 2024. Where does that spending go? Lots of new buildings, computer servers, high-powered semiconductors, electrical equipment to run it all, cooling equipment to cool it, and generators to provide backup power. Given that total U.S. business spending runs around $4 trillion, the growth in capital spending from just these 2 companies is 4% of total business spending. In other words, if no other businesses increased their capital spending this year, business spending would still grow 4%. That’s quite a buoy for the economy.
There is obvious consternation from investors that the hyperscalers will ultimately be throwing some portion of this down the drain. The hyperscalers say, “we know, but we think the greater risk is not arming up and fighting in this war right now,” as if the war is being fought for a huge new previously undiscovered gigantic territory with no inhabitants and unending opportunity. When the base business of each company continues to perform well, with strong growth and guidance for strong growth to continue, investors seem resigned to assume that the ends justify the means, so both stocks have held up well since they reported earnings.
As a result, companies providing equipment to the data centers are presently printing money, and this looks set to continue in 2026. At some point, however, growth will slow, most likely through some realization that there is enough compute capacity, whether through a slowdown in the adoption of new use cases, or greater productivity in generating the compute. This exact situation happened in 2000 with fiber optics. Our contention is you do not want to be around any stock that is entirely dependent on AI or data centers when that happens, and we just don’t know when that will happen, but it could be sooner than some of the Kool Aid-drinking “experts” believe.
The concern du jour is on software stocks. When you think of software, you might think of Microsoft, and rightly so, as they are the king of enterprise software. There are many durable software vendors that survived the dot.com bust, and they have pretty much all shifted their software to the cloud from corporate-managed data centers. There is a whole ‘nother industry of “cloud-native” software companies that have been created in the last 25 years which have grown to become very successful, even large, companies. The investment case for them has been that more and more companies have moved to the cloud, and this has opened up a lot of new market share for them. As that move to the cloud is mostly in the rear-view mirror, stickiness becomes a key selling point for them. Once they become ingrained within a large corporate client, it becomes painful for that client to ever get rid of them. Then they cross-sell other products to that client, generating additional revenue growth. Then they raise prices annually, because they can. So, it seems to be very durable, reliable revenue growth. And investors have priced software companies accordingly, with very expensive valuations, because they extrapolate this growth out a long time into the future.
One thing they forgot along the way: if there’s any constant in technology, it’s that technology constantly disrupts technology. Lo and behold, investors are suddenly concerned that some form of AI might be a chink in the software armor. The tip of the iceberg this year, particularly this week, is the release of a new AI tool by Anthropic, one of the new successful AI companies which always has its latest LLM (large language model) ranked near the top of LLM rankings. LLMs are the base of AI; applications that mine AI for data use these LLMs. Anthropic has created tools on top of its Claude LLM that have formed a dominant market position among data scientists who are creating a lot of the “use cases” for AI. In other words, whereas OpenAI’s ChatGPT has focused more on consumers, Claude has focused on the creators. The Anthropic tool in question is called CoWork, which claims to make it easy for corporate IT folks to write their own software.
So the trillion(s) dollar question is: will companies start to replace their enterprise software with internally created software? Among the reasons companies would be interested in doing this is cost…software costs a lot of money. And now that companies are allocating some greater portion of their IT budgets to AI endeavors, in theory there’s less money for other things. In practice, however, the average company is more likely allocating the growth in their IT spending to AI, with flattish spending on everything else IT. But still, when every software vendor you have is raising prices every year, that makes it tough for you to maintain flattish spending. One fear that probably has the most validity is that companies will pay for fewer “seats,” or users of the software, whether that’s corporate-wide or within the IT department.
The greater fear is large companies replacing the software that employees have grown accustomed to using, which works well for them, and has security protection, and doesn’t “hallucinate” (give results that are really wrong or highly offensive), in favor of Coder Joe’s home-cooked creation. It’s far-fetched to consider many companies doing this, but will a few do it? Who knows? Jensen Huang, founder and CEO of Nvidia, the epicenter of AI spending, believes it is a ridiculous proposition that companies would even consider doing this, for what it’s worth.
Wall Street, however, loves a good story. If it’s a negative story, the side of the Street that really likes it are the hedge funds who short stocks, meaning that they sell the stocks without owning them…a way to bet on their stock prices going down. The speed and magnitude of this drop implies that nearly every hedge fund is participating. Here’s the price chart for the most widely-followed software industry index, called IGV:
This is a 5-year chart, showing that software had a great run from late 2022 (when ChatGPT debuted) to last fall, storming +130%. Since the peak, though, the index has lost nearly one-third of its value. Note that a lot of companies have lost a lot more than that, generally those with the highest valuations.
The more we think about it, the more ingenious this concocted story is. To wit, there has been virtually zero proof that companies are going to rip out their software, though the market is on high alert looking for said proof. What we have seen this earnings season is software companies continuing to beat earnings and sales estimates and saying that the expected level of growth will continue into 2026. The nitpicking on companies like Microsoft, SAP, and ServiceNow has centered mostly on the fact that the companies didn’t raise guidance, or call for an acceleration of their growth rate. What software company in their right mind would stick their neck out and call for an accelerating growth rate right now? We are looking closely at some suddenly cheap software stocks, but we are mindful that they are falling knives right now, and it’s hard to envision what might pump up their stocks, when the fear that is gripping them is something that might not even start to play out for years, making it hard to refute the story with proof.
Meanwhile, AI data center plays like Nvidia and AMD continue to raise their guidance, and with capital spending announcements by the likes of Meta and Google, that shouldn’t be a surprise. And yet their stocks have been hit because they didn’t raise guidance enough. This might be the 3rd derivative of growth. Sales will continue to grow strongly this year on the back of that data center spending growth. Valuation has long been a concern for these stocks, but now that the stocks have staggered, and earnings have grown sharply, valuations are suddenly lower. Our contention, however, has been and continues to be that these companies are grossly over-earning. Chip companies have essentially been auctioning their chips to buyers because they’ve been in short supply, so chips are obviously being priced like something that is in short supply. Nvidia’s net profit margin, after all expenses including income taxes, is 56%. This is unheard of in corporate America, let alone in semiconductors. We all know what happens when something goes from supply-constrained to not-supply-constrained… prices fall. This will happen at some point. Then consider that a lot of other chip companies are coming after them, even their own customers (Google especially).
As if this week’s move in tech stocks hasn’t been crazy enough, there has been an opposite move in safety stocks (consumer staples, health care), but also in other sectors like financials, industrials, and consumer stocks. Many have been up 3% every day this week, which is off the charts (bad joke). The core stocks, on average, are +5% year-to-date, despite a number of tech stocks on the list. The S&P 500, meanwhile, is -0.3% as we write this. Not that we ever view the Nasdaq as a viable benchmark, but it’s -2.8%.
And yet, the question we get most often is: “WHAT’S GOING ON WITH GOLD?” The correct answer might be out there somewhere, but it seems to vary depending on whom you trust. Probably every answer needs to include one justification: because of how crazy the world seems to be getting. We won’t be diving into the hornet’s nest of politics, but maybe we can agree that the pace of change and chaos coming from the White House is a bit unprecedented. In case you have a bone to pick with that, I invite you to explain why gold suddenly nose-dove from $5,600 on January 29 to $4,840 today, on the news that Trump has nominated Kevin Warsh to be Fed chair. Warsh is seen as perhaps more independent than some of the other folks that were under consideration, and also more likely to care about inflation. Gold had been on quite a ride until then, although silver’s ride has been even crazier. Silver was +69% since the beginning of 2026 when it hit $121 on January 29, and this after rising from $29 to $72 in 2025, +148%. It has now dropped from $121 to $75. We still believe that some of the boom in precious metals, which has been corroborated by booms in other industrial metals, can be traced to fears that the world is de-globalizing, and that more and more producers will hoard their natural resources rather than exporting them. What would happen next is countries around the world who are importers of essential commodities will start buying them hand over fist so that they at least have some. Some market participants believe this has been happening.
Meanwhile, on the subject of crazy, there’s crypto. Bitcoin has been plummeting since its peak in October, now down almost 50% from its high! Its price is down 12.5% today alone, as we write this. Crypto was supposed to be a bona fide play on the debasement of the U.S. dollar, and the financial system in general, which were thought to have played a role in the boom in the price of gold, which continued and even accelerated since October, as bitcoin was crashing. The crypto skeptic would take this to conclude that crypto’s value is tied to little more than as a speculative asset with no fundamental value.
This post is hopefully long on information; whether it’s useless or useful is in the eye of the beholder. We continue to try to keep on top of everything that drives the financial markets and are always interested in taking advantage of dislocations. We will always strive to buy into good companies that have durable, defensible, long-term businesses with a path to growth that is highly likely to occur. If we can buy at a good price, all the better.
If you have any concerns or questions, please don’t hesitate to reach out to us. And of course, obviously a shout-out to the National Champion Hoosiers and all members of Hoosier Nation!