I know we are past Halloween, but everyone seems to be scared of an AI bubble. It seems to be the only thing people can talk about in investing circles. So, are we in an AI bubble?

Of course we are, but that is not the right question. If one wants a useful answer, then she must ask useful questions. In the late 1990s I was part of an investment team that was running an asset allocation strategy, and the lead manager was convinced that we were in a technology stock bubble. Based on our work we moved away from large growth technology and went heavily into real estate; if we could have then time traveled forward a decade, he would have looked like a genius. Unfortunately, he made that call several years before the bubble burst.

In the meantime, people remember famous quotes from the era. Fed Chair Alan Greenspan famously warned of “irrational exuberance.” We all remember that, but do we remember that he said that in 1996, four years before the tech bubble burst? There was lots of talk about bubbles in the 1990s. NASDAQ, the home of most of the high-flying technology companies at that time, had more than a dozen corrections during the 1990s. (A correction is a drop of 10 percent or more.) Every single time there were people out there saying the bubble was bursting, and they were wrong every time.

What they were really saying then, and what many are saying now is, “I am not invested in the internet (AI today) and I’m relieved there is a correction because I have missed out on all those returns.” They will continue to do that for as long as they can stand it. Then, the very same people will ultimately give in to their fear of missing out (FOMO) and change their tune.

One of my colleagues at Invesco during that era had great success investing in technology firms. He declared that valuations no longer mattered, and once told me that “small value” was an oxymoron. His fund then dropped 49 percent in value in 2001. The small value category was the best place to be in the market for the next decade. He lost his job, but not before demonstrating how people talk before a bubble bursts.

The right question is not whether we are in an AI bubble, but where are we in the inflation process of that bubble? The answer to that question is that we are far closer to the beginning than we are to the end. How do we know? Because everyone keeps talking about being in a bubble. When that stops and they start saying things like “this is a new world,” “valuations don’t matter,” or “every other investment makes no sense,” that is when the bubble is about to burst.

Warren Buffett did not go on the internet bubble ride, and you don’t have to go on the AI ride. However, if you choose that path, you had better be sure that you are as psychologically strong as Buffett. During that time articles were written describing how he was washed up and could no longer succeed in the new world. He didn’t care, and he came out just fine. Likewise, had my colleague not been fired and been allowed to ride out the decade of the 2000s, he would have been fine as those tech stocks did ultimately pay off. He was not allowed to be out of sync with the market that long; few investors can. That is really Buffett’s secret: After a good ten-year run in what today would be called a hedge fund, he bought Berkshire Hathaway and fired his clients. He can ride through long periods of being out of sync.

Most of us have to live somewhere in the middle. AI is a bubble and eventually it will burst, but that is most likely years away. Instead of guessing, we believe that one should remember that it is better to be an owner of companies than a trader of stocks, and diversification is our friend. There is no law that says one can own only AI or energy but not both.

Finally, at the end of 2002 I left Invesco to start Iron Capital. Shortly thereafter my girlfriend (now wife) and I were at a dinner party where one of the other couples asked, “Do you think we are in a housing bubble?” That was approximately 2004, and we were, but that was the wrong question…

Warm regards,

Chuck Osborne, CFA

~Bubble, Bubble Toil and Trouble

Sir Rotheram has been murdered, found dead in his bathtub. Scotland Yard is completely baffled so they call the world’s most famous consulting detective, Sherlock Holmes. Holmes immediately notices the bath salts and requests that the police find where they are stored, and while at it look for footprints at all the windows. As Holmes dismisses the police he utters, “Data, data, data, I cannot make bricks without clay.”

After the police leave the room to find this supposedly valuable data, it becomes obvious to the audience that the bath salts hunt was a simple rouse to get the police out of the way while Holmes searches for more meaningful data, which turns out to be Sir Rotheram’s secret study in which he practiced his magical rituals. This data eventually leads to the capture of the movie’s primary villain, Lord Blackwood.

This movie scene has come to mind lately as one of my daily rituals has been disrupted by the government shutdown. We have our investment committee meeting every workday morning, and one of the agenda items is to review the economic data released by our government. We also review the economic data from around the world, but most of what we look at comes from right here at home. With the government shutdown, that data is not flowing.

What is the unemployment rate? We don’t know. How many filed for unemployment last week? No clue. Data, data, data…how are we to make bricks without clay? I will freely admit that my name is Chuck and I am an economic-data-aholic. It has been 20 days since my last economic data update. Having said that, this has actually been a nice reprieve, since much of the data flow from places like the Bureau of Labor Statistics is noise, not news, and we may be better off without it.

Of course, we are investors and not traders, so the daily flow of data is far less concerning to us. I can’t imagine the stress this blindness presents for those who program computers to trade on every utterance from government officials. Meanwhile, we get to focus on the data that actually counts: Corporate earnings. It never ceases to amaze me how many intelligent people lose focus on what we are actually doing when we invest in stock. They forget that stock is ownership in a company, and that in the long run, it is the business results of that company that matter.

This quarter’s corporate earnings reports have just begun so it is too early to tell, but thus far they have been good. Eventually this lack of economic data will be a problem as the economic environment does impact the business results of companies, but in the interim we are reminded that not all data is created equal. What really matters to investors are the business results of the companies in which they are invested. That will be our focus.

There are far more critical functions of government that need doing and we hope this shutdown will be over soon. In the meantime, it might be good to know that the missing data is closer to the location of bath salts than to the critical clues that lead to the capture of Lord Blackwood. A good analyst, like a good detective, knows that while data is the clay with which the bricks are made, not all data is created equal.

Warm regards,

Chuck Osborne, CFA

~Government Shutdown: What Do We Miss?

“The market is wrong.” A senior portfolio manager at my old firm made this adamant statement in response to learning that he was being moved off of the portfolio he had been managing for more than a decade. His portfolio was not keeping pace with the market, which, in this particular case, was defined as the S&P 500. He needed to add more technology stocks in his portfolio to boost the return and be more aligned with his benchmark. He refused because he believed those stocks were in a bubble and it would be ugly when that bubble burst. Senior management’s response was to do what they had done before with experienced portfolio managers, “promote” him so that he no longer had day-to-day investment decision-making responsibility.

I know what you are thinking:  In hindsight that proved to be a huge mistake, and this gentleman proved to be correct. Yes, but unfortunately for him, this conversation took place in early 1998 – two full years before that dot-com bubble finally burst. In the investment business there is little difference between being early and being wrong.

Are we in an AI bubble? Late last week I took an Uber home from the Atlanta airport after being gone several days. The very friendly driver asked what I did for a living. I hate that question, because it usually leads to a second question that isn’t really a question but a statement that the questioner wishes for me to reaffirm.

In this case, my Uber driver asked me what I thought about the valuations of AI companies. Fortunately for me, this gentleman didn’t even wait for my response before telling me the answer. According to my Uber driver, AI is wonderful technology but it is not the end game, they want to create actual human-level intelligence, which they will never do, so these stocks are all in a bubble that is going to burst and cause huge financial damage.

I never responded to anything the gentleman said, which turned out to be fine because he didn’t stop talking long enough for me to get a word in even if I wanted to. He was just as certain about us being in a bubble as that senior portfolio manager was all those years ago. He is even more wrong.

As Mark Twain supposedly said, “History doesn’t repeat itself, but it often rhymes.” We are not in the same place today that we were in the 1990s, but there are certainly similarities. AI is an exciting new technology with huge potential, much as the internet was 30 years ago.

However, the internet came at a time when value investing was king. Many of the internet startups were staffed by IT professionals who had found themselves unemployed when the tech giants of the day had massive layoffs in the early 1990s, while the AI movement came after a decade of large technology firms already dominating stock market returns. Most internet companies went public with little more than the promise of having “.com” at the end of their names, whereas AI has largely been driven by already established firms such as Nvidia and Microsoft.

AI has immediate productivity-boosting capabilities while the internet had potential. As a result, AI revenues have grown as fast, if not faster than, most of the stocks. This means that thus far there has been far less speculation than at the same moment in the 1990s. This is not a simple repeat of history, although there are lessons to be learned.

The one thing that drives all bubbles is human psychology. That has not changed, which means we likely will have an AI bubble. It is our nature to overdo; we overreact to news both good and bad, and we do it on repeat. However, if that does happen with AI, we are much more likely in 1997 or 1998 as opposed to 2000. Bubbles don’t burst when Uber drivers are talking about being in a bubble; they burst when everyone believes it is a brave new world.

Two years after that uncomfortable meeting with the senior portfolio manager, I had another one. This time it was me stepping out on a limb and suggesting that our current star portfolio manager was about to blow up. This portfolio manager had just won a “Manager of The Year” award from Morningstar and had appeared on Louis Rukeyser’s “Wall Street Week,” where he infamously suggested that valuations no longer mattered. That was February of 2000; the bubble burst in March, and thankfully for my career, I was proven correct. There is little difference between being early and being wrong, and it is too early for bubble warnings.

Warm regards,

Chuck Osborne, CFA

~Artificial Bubble?

All the best stuff happens on Fridays. After a rather quiet summer, potential market-moving news hit us two Fridays in a row. First, on Friday, Aug. 29, the U.S. Court of Appeals for the Federal Circuit ruled that the president does not have the authority to issue to so-called reciprocal tariffs. Then, on Friday, Sept. 5, we received a weak jobs report.

The jobs report is a dark cloud. For most of my career, the rule of thumb on jobs has been that the U.S. needed to create 200,000 jobs every month to maintain the same level of unemployment. That may sound like a large number, but the U.S. is a very large country. There are always new people entering the workforce, and in any free society there will always be some level of turnover. People lose old jobs, so new jobs are needed.

Many economists now believe that number is much smaller, and there has been a lot of discussion on this lately as unemployment rates have stayed steady while we have created 200,000 jobs or more on only six occasions over the last 25 months. This means 19 of the last 25 months have fallen short, yet the unemployment rate has barely moved. In my opinion this conversation, while interesting in an academic sense, is simply missing the point: Our economy created only 22,000 jobs in August, and June was revised to be negative.

If economists are correct and our working-age population is shrinking, therefore we do not need to create as many jobs to maintain a low unemployment rate, that is great – as far as not having too many unemployed. However, an economy that produces only 22,000 jobs is a weak economy. We are talking about much weaker growth than we are accustomed to in the U.S. That is bad for investors, so as far as I am concerned, the entire discussion on the unemployment rate misses the larger point.

It gets worse. If we look at where jobs were gained and lost, we see that manufacturing lost 12,000 jobs and wholesale trade lost 11,700 jobs. All the gains were in services. The tariff policy is intended to do the opposite, but this is exactly what the laws of economics would suggest. Tariffs raise costs on select goods, mostly manufactured goods. To the extent that those costs are passed on to consumers and then demanded, those goods will fall. To the extent that those costs are eaten by manufacturers, they reduce the incentive to create supply, and supply is therefore reduced. Best case scenario is that they simply lower growth rates from what they otherwise would be; worst case, tariffs plunge us into a recession, just as they have done historically.

So, what is the silver lining? The U.S. Constitution puts the authority to create trade policy in the hands of Congress. The president does not have the authority to do this, and our courts are there to keep each branch of the government in its own lane. The administration has lost in court and now has lost on appeal. This is going to the Supreme Court, and based on the history of this court, the most likely outcome is that these tariffs are going away.  This may take time, but the president has asked the court to expedite the case.

How things get done matters; we have lost sight of this in recent years. Trade agreements are often negotiated by administrations, but they only go into effect upon approval by Congress. That is how our system is supposed to work. That can be frustratingly slow for people who feel strongly on any given issue, but that slow, thoughtful pace is what has provided stability to our system even though we have elections every two years. Stability is good for investors.

In the meantime, we will continue to do what we do best: Looking at investments from the bottom-up. Some companies will grow even if the economy doesn’t. Tariffs are a loser for the economy, but there are both winners and losers among specific companies. As always, this too shall pass, and the prudent investor will get through it.

Warm regards,

Chuck Osborne, CFA

~A Dark Cloud and a Silver Lining

In the past week I have seen two articles in The Wall Street Journal, one pro-tariff and one anti-tariff, which both ask a basic question: If tariffs are so bad, why have we not fallen into a recession?

Never mind that the rule-of-thumb definition of a recession has historically been “two quarters in a row of negative GDP growth,” and that the initial tariff announcements were made only one quarter ago, then delayed for 90 days. I realize that, in these days of our 24-hour news and social media noise storm, three months can seem like three years – but in reality, this past April was just a little more than 90 days ago.

Last week we received a surprisingly high first estimate of second-quarter GDP growth of 3 percent, which some touted as proof positive of the wonderful impact of tariff policy. Then we received a very poor jobs report, which others touted as proof positive that government data is rigged and can’t be trusted (except the GDP number we liked, of course).

There are plenty of people now explaining how government data is compiled:  first releases are always estimates, and data continues to be collected. Revisions happen all the time, and in some cases, years later. What people are not talking about as much is important: How does the economy grow at 3 percent and yet create so few jobs?

The answer to both comes with context, something we seem to lack the patience for these days. The economy growing at 3 percent was driven by an extremely low import number. The context here is that this 3 percent reading was preceded by a reading of negative 0.5 percent. In other words, when one simply looks at the data, then the economy shrunk by 0.5 percent in the first quarter, then grew by 3 percent in the second quarter. If all one did was average those two quarters, then he could see that growth in the first half of 2025 was an anemic 1.25 percent. That is not a recession, but it isn’t good either.

The wild swing in GDP is driven by international trade. Imports are subtracted from the math of the GDP calculations for reasons we have discussed before. There was an enormous influx of imports in the first quarter as many consumers and companies stocked up on international goods in fear of potential tariffs making them more expensive. There was then an enormous decline in imports the second quarter as those previously imported goods were used instead. Less discussed in the second quarter estimate has been the negative from a drawdown in inventories.

All of this simply leads to the conclusion that all data must be viewed in context and no single report tells the whole story. This is doubly true for the employment situation report. Even prior to the revisions, which have been attacked, the economy was on a downward trajectory in terms of new jobs being created. For most of my career, the thought has been that we needed the economy to add 200,000 jobs every month to keep the unemployment rate steady; this amount simply kept up with growth in population and turnover. In the last 24 months, the economy has added that many jobs only six times. Put differently, the economy failed to add 200,000 jobs in 18 of the last 24 months.

The economy is slowly losing steam, and we can kill the messenger if we wish, but that does not change reality. We are not in recession, but we are not growing fast enough to add sufficient jobs.

What does this mean for investors? The relationship between the stock market and the economy is loose at best. In the short run stocks move on headlines and emotions, and in the long run they reflect the earnings of the companies. While the economy has been slowing, corporate earnings are still growing nicely. It is too early this earnings season for any meaningful conclusions, but so far so good. AI is real and the demand for AI investment is just going to grow. It is already helping some companies grow productivity, which will help produce earnings growth even if revenues slow. We are cautiously optimistic, but more cautious than we were a few months ago.

It could all improve if the tariff situation settles down. It would not hurt for the Fed to lower rates (more to come on that), but in my view they are late to act. Meanwhile, we can avoid the madness by keeping reports in context, accepting reality for what it is even if we don’t like it, and focusing on what really matters, which are the companies we own and their actual businesses.

Warm regards,

Chuck Osborne. CFA

~Context Matters