The Thomas Jefferson treatment... AI can't be ignored... But not every 'slam dunk' AI stock will make money... Magnificent Seven stocks are underperforming... Worrying about 'the bezzle'... Big booms and busts... There's always something to do...
AI gets the Thomas Jefferson treatment...
Yesterday morning, I (Dan Ferris) attended a breakfast modeled after Thomas Jefferson's famous dinner parties at Monticello. Jefferson invited his guests to engage in spirited, moderated conversations on specific topics. And every guest participated.
The breakfast was part of my friend and well-known value investor/author Vitaliy Katsenelson's Intellectual Investor Conference, held every summer in Vail, Colorado.
The topic at hand? How we as investors use AI.
The comments were wide-ranging, but one common theme caught my attention...
Several attendees mentioned they've regularly found data-reporting errors in various financial-data services, including Bloomberg and S&P Capital IQ. One attendee captured screenshots of data errors and forwarded them to Bloomberg, which did nothing about the underlying issues.
When you pay nearly $32,000 a year for a Bloomberg terminal, you expect few – if any – errors and swift service in fixing them.
The attendee finally got fed up and built his own suite of tools using Claude Code. The tools go directly to the U.S. Securities and Exchange Commission ("SEC") website to gather up the original source data, build spreadsheets just the way he likes them, and even audit the resulting documents to make sure all his instructions were followed to the letter. Bloomberg no longer sits as an intermediary between him and the SEC filings he needs to analyze companies.
The humans at Bloomberg can't seem to fix the underlying problems with the data they publish... so investors are using AI to eliminate Bloomberg entirely – for $200 a year instead of $32,000.
The conference started Wednesday night, with presentations on six different stocks...
It continued Thursday night with eight more pitches and concludes tonight with seven more (including my own).
What happens at the Vail conference stays in Vail, so I can't give you many details. But I can tell you that virtually every speaker felt compelled to address the impacts of AI on the stocks they were presenting.
And that's not because they're all irrationally caught up in the AI frenzy. Far from it. These are some of the most rational and brilliant folks I know. They're simply demonstrating the same insight I've offered before...
I expect AI to turn out just like the Internet, for both investors and users. Just like the Internet, AI will impact every business and individual connected to it (and likely many who aren't directly connected).
AI can't be ignored, and if you're not using it today, you'll be using it soon. One investor at breakfast yesterday said, "I love it. If I could inject it into my veins, I would." We chuckled, but everybody in the room recognized that they simply can't operate without AI.
Unfortunately, also like the dot-com boom, investors will buy "slam dunk" AI companies... and lose their money.
Just consider... since the start of the year, only one of the Magnificent Seven companies – Alphabet (GOOGL) – has (just barely) outperformed the S&P 500 Index.
As of yesterday's close, the index was up 7.5% since December 31. Meta Platforms (META) and Tesla (TSLA) were both down roughly 17% to 18%, and Microsoft (MSFT) lost 27% during the same period.
Perhaps investors are starting to worry about 'the bezzle'...
John Kenneth Galbraith coined the term "bezzle" in his 1955 book The Great Crash, 1929. He described it as the "undiscovered embezzlement in... the country's businesses and banks." It's all the fraud that builds up during a bull market.
When times are good and people are "relaxed, trusting, and money is plentiful," the bezzle grows. When times are bad, "all this is reversed," and the bezzle shrinks as financial mistakes and crimes are brought to light. In short, the bezzle becomes exaggerated in a stock market mania, amplifying the fallout in the ensuing bust.
In a June 19 Reuters essay, economic historian and author Edward Chancellor took account of the current bezzle. Most prominently, he noted the roughly $800 billion invested in OpenAI by Microsoft, Oracle (ORCL), Nvidia (NVDA), Advanced Micro Devices (AMD), Amazon (AMZN), and CoreWeave (CRWV). OpenAI, in turn, is using those investments to buy those companies' products and cloud-computing services.
Chancellor notes how well it rhymes with the rampant vendor financing of the dot-com era, when telecom-equipment providers made billions in ultimately unrecoverable loans to risky Internet startup companies. That enabled the startups to buy expensive networking gear, boosting short-term telecom revenues. Of course, in the end, companies like Lucent Technologies, Global Crossing, and NorthPoint Communications couldn't collect on the loans and went bankrupt.
Chancellor catalogs more bezzle items, including how the hyperscalers have adopted accounting practices similar to the ones that first inflated Enron's performance, then ruined the company and sent its executives to jail. Today, Morgan Stanley estimates, "around $1.8 trillion worth of liabilities in the AI domain have been kept off hyperscalers' balance sheets."
Since the AI boom is still on, the trillion-dollar AI bezzle will likely keep growing. Who knows what heights it'll reach by the time it all unravels.
So maybe Mag Seven underperformance doesn't mean investors are worried...
Maybe it's just a normal short-term fluctuation of the stock market. Maybe it's a sign that investors are simply rotating into other AI-related stocks. Or maybe it's something else.
However, as Bloomberg Macro Strategist Simon White reported on June 12:
Net equity supply in the US has gone positive for the first time since 2021, ahead of a flood of expected [initial public offerings ("IPOs")], secondary issuances, and [capital expenditure] challenged buybacks.
White cited Fed equity supply data, which was released the day SpaceX (SPCX) went public. It was the largest IPO in history at $75 billion.
Unlimited Funds founder and Stansberry Investor Hour podcast guest Bob Elliott has noticed the same trend and recently said:
Today's equity market faces one of the biggest shifts in supply we have seen in some time. Buybacks led by cash producing megacap tech are slowing and issuance is surging with the new IPOs and secondary issuance from existing names as well like Google.
I wrote about the flood of new equity in the most recent issue of The Ferris Report. It's one of three simultaneous trends that preceded the dot-com bust. All three of those trends are solidly in place today and have been underway for months.
I can't predict when the AI frenzy will peak. Nobody knows that. And recent market weakness could just be another buyable dip in the bull market.
So, as usual, I'm not telling you to sell everything and head for the hills. That's never the right advice. The world has a way of not ending... and every market catastrophe winds up being an opportunity to seed generational wealth.
I doubt the AI saga will be any different. It'll make a few prudent speculators rich on the way up, a lot of folks poorer on the way down, and create a massive opportunity for outsized long-term compounding once the inevitable bust bottoms out.
One breakfast participant yesterday exhibited this same sense of patience in his view toward using AI. He said he doesn't need to get into the weeds of using Claude Code to build what would likely be primitive investment tools when he knows those capabilities will soon be "delivered to him on a silver platter" by a business that can do that sort of thing far better than he ever could.
Patience is how investors should have approached the best dot-com companies in the late 1990s. Cisco Systems (CSCO) plunged from $80 at its peak in March 2000 to $8.60 at its trough in October 2002. From that bottom, it compounded at more than 12% a year through its all-time high of $130 on June 4, 2026.
I expect many of the best AI companies to behave roughly the same way. They'll all collapse in the bust... and prudent, patient investors will get to buy them at cheaper valuations than they'd believe. Markets can go higher than you'd ever dream and fall further than you'd ever imagine. Huge risks and opportunities emanate from that one well-established historical pattern.
It's hard to see a huge bull run like the one we're in as containing the seeds of big losses. And once those losses come to pass, it's even harder to see the years of compounding that lie ahead from those far more attractive valuations. But that's just how markets work.
I'm not saying that investing is simply about waiting for big booms and busts to happen...
Far from it. As the late global value-investing guru Peter Cundill used to say, there's always something to do. Right now, for example, there's a growing opportunity in stocks that will likely benefit from AI, but otherwise remain largely untouched by it.
I'm talking about railroad, energy, chemical, and other types of industrial companies that make things for which there is no viable substitute.
AI will likely help many of these companies become more efficient, but is unlikely to directly impact demand for the basic commodity products they provide, without which a modern standard of living is impossible.
Mike Barrett and I recommended one of these names in Extreme Value recently. I'm presenting it today at the conference in Vail. I've also recommended several such stocks in The Ferris Report.
In the end...
AI is a massively powerful new technology that is impossible to ignore.
The folks here in Vail aren't speculating wildly on AI stocks. They're thoughtfully assessing its impact on the businesses they're considering investing their own and their clients' money in.
More than anything else, they're using it every day, learning how to build tools and become more efficient. And they're using AI to more quickly and effectively assess potential new investments. (One breakfast attendee said he uses AI for idea generation – finding new stocks.)
I recommend you do the same. I'd be willing to bet that you'll make more money on AI by forgetting about chasing short-term speculative profits and focusing more on how to use it in your own business and other daily activities.
Take your time. Learn about it. Figure out what it means for your life and work. There's no reason to feel like you're missing out. I promise you won't feel that way when the AI mania is over and stocks like Nvidia are down 60%.
But you might feel left behind if you haven't learned how to use AI to make your life more efficient in five years.
I'll leave you with a quote from one of yesterday's presentations in Vail. It's by a brilliant investor and Stansberry Investor Hour podcast guest, Gary Mishuris. As an MIT-graduate, Gary brings a very thoughtful and methodical process to investing and how he uses AI. He is perhaps the most sophisticated AI user I've met among professional investors.
He said:
We try to predict the future too much and we underprocess what is happening right now.
We underprocess what is happening right now...
I got a real taste of that when I got back to my room around 9:30 last night. I noticed that, with six of the Mag Seven underperforming the S&P 500 so far this year, my October 2024 Ferris Report recommendation of the State Street Industrial Select Sector SPDR Fund (XLI) made a new all-time high yesterday.
It's loaded with several industrial companies, many of which are already benefiting from AI-related energy demand. The stock is up roughly 15% this year, double the S&P 500 and well ahead of every Mag Seven stock.
Gary might tell you that being bullish on AI hyperscalers is essentially a prediction that their massive capital spending will pay off, even though they're only burning cash today. He might also tell you not to "underprocess" the outperformance of industrial companies that supply essential infrastructure components.
I would say simply: Prepare, don't predict.
New 52-week highs (as of 6/25/26): Applied Materials (AMAT), Alpha Architect 1-3 Month Box Fund (BOXX), Healthpeak Properties (DOC), W.W. Grainger (GWW), iShares Biotechnology Fund (IBB), Illumina (ILMN), LXP Industrial Trust (LXP), Roivant Sciences (ROIV), Snap-on (SNA), Twist Bioscience (TWST), UnitedHealth (UNH), and State Street Industrial Select Sector SPDR Fund (XLI).
In today's mailbag, feedback on yesterday's Digest, including thoughts on Micron Technology (MU) shares... and a take on why the Federal Reserve won't want to raise interest rates... Do you have a comment or question? As always, e-mail us at feedback@stansberryresearch.com.
"If the Fed raises rates it will prick the AI bubble, which will reduce US tax income and it will crimp consumer demand and reduce employment, further reducing tax income but at the same time it will increase the US Treasury's spending on interest (since much of the US debt is short term debt). So all of that will cause an economic crisis, ultimately leading to much lower rates. So I am not convinced the Fed will raise rates." – Subscriber S.J.I.
"I wish I had purchased MU a year ago. I live in central New York, where Micron is building their new plant. Whitney wrote [earlier this year], that memory chips are commodities and their returns will follow the commodities cycle of undersupply/oversupply. The time to buy them is when they are not in demand..." – Subscriber Kent M.
Good investing,
Dan Ferris
Vail, Colorado
June 26, 2026
