All Ships Can Sink

The script has (mostly) flipped... Retailers and their memes... The Titanic market... How folks like me 'miss out'... Arithmetic, not value... 95% of AI is losing money...

Mom and Pop were the first to panic... 

As I (Dan Ferris) have pointed out before, during the COVID-19 crash of March 2020, nearly 1% of individual "retail" investors abandoned stocks entirely. That's one factor in the market dropping 34% in about a month.

As Mark Hackett, chief market strategist at Nationwide Mutual Insurance Company, recently told the New York Times, retail investors have always been the first to panic. They're not doing that anymore.

During the tariff tantrum of early April, Mom and Pop put $30 billion into mutual funds and exchange-traded funds ("ETFs"), a common way of gauging individual investor activity. During the same period, Bank of America reported that professional asset managers had logged their biggest two-month drop in U.S. stock investments, as the pros panicked and went to cash and Treasury bills.

I'm not surprised...

Regular Digest readers might remember a chart that I've shared repeatedly since 2022. I made the chart after my friend, investor/author Vitaliy Katsenelson, quipped to me that you could take the past 20 years and invert it to arrive at a reasonable set of expectations for the next 20. Here's how that would look...

The first item is still true. The 10-year Treasury yields around 4.2%, while it was only around 3.1% when I first published this chart on July 8, 2022.

The second, third, and fourth items have clearly not played out yet. Equity valuations are higher than ever, with investors aggressively buying dips and fully embracing a risk-on mentality.

The next few items are more nuanced. U.S. stocks, foreign stocks, and many commodities have done well recently. Growth is still trouncing value. And we've seen political ideology drive the markets...

First, the Donald Trump administration took a 10% stake in chipmaker Intel (INTC)... Then investors dumped shares of Cracker Barrel Old Country Store (CBRL) in the belief that the restaurant's new logo was too "woke."

Today, though, we're most interested in 'Institutions Move Markets' becoming 'Retail Traders Move Markets'...

And we've seen plenty of evidence of this point.

Besides the fund flows we cited above, there's the recent meme-like action in heavily shorted stocks.

For example, Kohl's (KSS) doubled quickly off its April 8 bottom. Cracker Barrel more than doubled and is still trading roughly 70% above its low. Opendoor Technologies (OPEN) hit bottom at $0.53 on June 30 and was up roughly 850% by August 22. Even after a correction, it's still more than 700% above its lows.

All three still have large outstanding short positions, according to data compiled by Bloomberg, so who knows? Maybe they're all good for another pop or two before they fizzle out?

So to the retail crowd... not only are all dips buyable, but every stock is just waiting for its moment to shine, no matter how badly the business is performing. (And institutions don't heavily short a stock because it's doing well...)

Confidence in an unsinkable market...  

In a recent LinkedIn post, researcher/author Peter Atwater reflected on the hype surrounding the Titanic... the ship that was famously considered unsinkable until about four days into its maiden voyage in April 1912.

Atwater says that overconfidence creates what economists call "moral hazard." It means that you feel you're so safe already, you have no incentive to protect yourself further.

Moral hazard might lead to a ship setting out without enough lifeboats... or a retail investor buying stocks as if they can never go down.

It's not that retail investors think stocks aren't risky. It's that the idea hasn't even crossed their minds. Taking on risk is as automatic as breathing.

Atwater concludes:

Applied to an unsinkable market, the Titanic experience suggests that investors will underreact to declines in prices ahead. They will dismiss the naysayers and will hold on far longer during the decline than they might have otherwise, had they not believed the market was unsinkable... investors won't let go of it until it is too late, if at all.

Psychologically speaking, investors are finally back to where they were in 1999. That's the year after the Federal Reserve bailed out the highly leveraged hedge fund Long Term Capital Management. The market dipped when the fund was threatening the broader financial system. Then the Fed saved the day, and stocks continued to roar as they'd been doing solidly for the previous two years.

Retail investors have come to believe their financial ship will never sink. Even worse, they believe the seas will never be stormy again... the way they were in 1929 to 1954, 1966 to 1982, 2000 to 2002, and 2008 to 2009, plus other episodes in markets around the world.

And yes, I realize this is the moment where folks like me seem to go wrong... 

We value investors start to get worried that valuations are getting extreme, with the S&P 500 Index's price-to-sales ratio recently hitting a new all-time high of nearly 3.3. Every time that has happened in the past, it has spelled big trouble for the next several years' returns – and heralded a long, deep bear market.

That has never happened without an accompanying credit crisis, which doesn't seem imminent at the moment. But that doesn't mean equities will generate great returns... It doesn't mean they won't decline and go sideways for the next several years (as I've warned more than once).

Risk is there, whether you see it or not. Risk exists even when the hazard isn't playing out in a given moment. When you take risks, you must be ready for the best and the worst.

Maybe my caution will cause me to miss out on some of the apparently abundant upside for as long as the market continues to soar to unprecedented heights. Folks will say I've lost my touch or that value is dead, or both. I'll cite how folks said the same thing in 2000.

But it's not just about value investing. It's also about arithmetic.

Right now, investors in the S&P 500 are paying 39 times cyclically adjusted earnings, while the long-term median is 16 times.

When you pay $16 for $1 worth of earnings, you will earn a higher return than if you paid $39. Paying $16 gives you a 6.25% current yield on your money. Paying $39 gives you a 2.6% yield on your money.

A simple return to the median implies 59% downside from current levels. And you know how folks really behave in a steep decline. They stay too long, sell out well below the median value, and exit stocks right at the moment they should be going all-in and buying the ubiquitous bargains.

Maybe those old dynamics will return... maybe they won't. But folks paying $39 for every $1 of earnings will not make as much as those paying $16. That won't change.

That kind of arithmetic has one drawback...

It has no predictive value whatsoever.

For example, investors might well get excited enough to pay even more than $39 for $1 of earnings. They paid more than $44 near the peak of the dot-com boom, in December 1999. That provides you with a paltry 2.3% yield on your money.

But people don't think in terms of valuations and yields. They see stocks going up and they keep buying. They look back through near-term history and see that stocks have always recovered and made new highs. They see that the Fed and the government have always stepped in to fix things in a crisis.

Of course, this arithmetic doesn't apply to every individual stock. It's possible to make a fortune on a stock you pay 39 times earnings for. And you can lose everything on one that cost you 16 times earnings.

The problem is that when you pay 39 times earnings – as you are for the entire S&P 500 today – the investment must absolutely continue to grow its earnings power at a high rate for a long time. Anything less, and you won't even make a decent return. And if you want a higher return, it has to shoot the lights out for many years.

There's no margin for any kind of error...

At 59 times earnings and 30 times sales, Nvidia (NVDA) must grow earnings and revenues at high rates simply to justify its valuation, and it must grow even more to make investors any money over the long term.

On Wednesday, it reported a 56% year-over-year increase in revenue in the second quarter, plus a 59% annualized increase in net income during the period. But the stock market didn't seem impressed, as the stock fell 0.8% yesterday.

So if the stock can fall when revenues and profits grow like well-fed weeds, it doesn't take a genius to consider what could happen if and when the cycle inevitably turns against Nvidia's highly cyclical business. But in the meantime, all anybody can think about is the glorious future.

In May, Nvidia CEO Jensen Huang predicted that AI data-center spending would hit $1 trillion annually by 2028. It's an AI arms race among companies like Alphabet, Microsoft, and Amazon to build data centers and not be left behind by the competition.

To an AI bull, it probably sounds like these companies are minting money, which certainly contributes to the idea that the stock market has absolutely nowhere to go but up.

Problem is, AI isn't making any money...

Researchers at MIT interviewed 150 business leaders, surveyed 350 of their employees, and analyzed 300 public deployments of AI technologies. They concluded that 95% of AI pilot projects stall out and fail without generating any benefit to corporate profit and loss statements. (There's a Fortune article about it if you care why they failed.)

That seems about right to me. It's all brand new, after all. We have no idea what the long-term economics will look like.

Yes, as a new and potentially highly disruptive technology, it makes sense to be optimistic about possible productivity gains. That would be a typical outcome and has happened many times throughout centuries of history.

But the financial outcomes for specific industries and businesses are much harder to think about... and virtually impossible for investors to do anything but speculate on. In other words, AI is largely uninvestable. It's still a speculation at this point – and it will remain that way until MIT or someplace like it reports that more folks are making money with it than not.

I keep hearing even the savviest analysts telling me things like, "Entire organizations are using AI more and more every day." But nobody says, "We're minting money with AI." Growth in the technology absolutely does not mean growth in revenues or earnings.

Mike Green of Simplify Asset Management recently pointed out in a Substack essay that while Internet usage grew 1,000-fold from 2000 to 2022, telecom-services revenues were cut in half.

That makes sense to anybody who understands economics. Investing massive amounts of capital to make a good or service hyperabundant – like fiber-optic cable in 2000 and AI-focused data centers now – tends to send the price of that good or service plummeting. That decimates margins and bankrupts the most leveraged players. Why should it be any different this time?

So in the end, no matter how many new highs the S&P 500 makes in the next 10 minutes, folks right now are paying $39 for $1 of earnings that aren't nearly as likely to grow to the moon as they absolutely must in order to justify paying that much.

I know, I know...

I can't tell you the market will fall next week or next month. I can't predict who will be the biggest AI winners and the biggest AI losers.

If you care about that stuff, I understand, but it has nothing to do with being a great investor.

Because so many people feel otherwise, I know sentiment around AI investments will shift.

When it does, stock prices will suffer and investors will regret going all-in on AI.

And they'll get a reminder that there is no such thing as an unsinkable ship.

New 52-week highs (as of 8/28/25): ABB (ABBNY), Atour Lifestyle (ATAT), Alpha Architect 1-3 Month Box Fund (BOXX), BP (BP), Equinox Gold (EQX), Comfort Systems USA (FIX), Alphabet (GOOGL), Hawaiian Electric Industries (HE), Houlihan Lokey (HLI), iShares Convertible Bond Fund (ICVT), iRhythm Technologies (IRTC), iShares U.S. Aerospace & Defense Fund (ITA), JPMorgan Chase (JPM), Lumentum (LITE), Newmont (NEM), OR Royalties (OR), and Vanguard S&P 500 Fund (VOO).

In today's mailbag, feedback on yesterday's edition, which also covered Nvidia's latest quarterly earnings report... Do you have a comment or question? As always, e-mail us at feedback@stansberryresearch.com.

"A stock that is priced for perfection closing down less than one percent after posting earnings strikes me as a non-event. I just wish I had been an Nvidia officer for the last 10 years or so, racking up stock or stock options as part of my compensation package." – Subscriber Sherwin R.

Good investing,

Dan Ferris
Medford, Oregon
August 29, 2025