Editor's note: Today, Porter turns the Journal over to Matt Tuttle, CEO of Tuttle Capital Management and portfolio manager of the Porter & Company Porter Portfolio Index ETF (PCPP). Here, Matt explains why, in the age of AI, a stock that looks cheap may be cheap for a good reason.

Early on Monday, September 28, an email from JPMorgan's stock strategists arrived in clients' inboxes before the New York market opened. It was a research note on a group of stocks trading near record-low valuations.

For a value investor, that's Christmas morning. Traditionally, a low price relative to earnings has meant a margin of safety. But there was not going to be anything from Santa under this tree. JPMorgan wasn't recommending these stocks.

Instead, the note said that – despite their cheapness – these stocks could keep falling. The companies sat in software, business services, and media – those sectors where the investment bank saw the most risk from artificial intelligence (“AI”). They included Adobe (ADBE), Accenture (ACN), and Intuit (INTU), the maker of TurboTax. All three were down 30% or more for the year.

I believe that these stocks weren't cheap because the market had missed something. They were cheap because the market had seen something.

However, not everyone at JPMorgan agreed. Two months earlier, Paul Quinsee, global head of equities at J.P. Morgan Asset Management, looked at the same beaten-down stocks and wrote that fear of AI had left "some very attractive long-term returns on offer in the so-called 'AI losers.'"

Same bank. Same question. Opposite answers.

Quinsee himself put his finger on the problem. In the same note, he warned that some stocks "look quite expensive from a longer-term perspective, even when trading on single-digit multiples of near-term earnings."

That's the flaw in the price-to-earnings ratio (P/E). It tells you how much investors pay for each dollar of a company's profits. It can't tell you whether those profits will last.

For decades, there’s been little question about whether profits will rise. Recessions knocked earnings down, and then they came back. In 1990, the companies in the S&P 500 earned about $23 per share. By 2025, they earned about $272. But within that big picture, there are leaders – and the laggards.

Which companies may be laggards is one of the most important questions in investing. But the problem is, the old tool for determining that now often points the wrong way.

In 1992, mutual-fund rating service Morningstar drew a grid of nine squares to sort funds by investing style. Company size, measured by stock market value, runs down the side. Value, blend, and growth run across the top.

The grid works from the bottom up. Morningstar first places each stock in a square, based on its size and how its price compares with measures like earnings. Cheap stocks go on the left, into value. Pricey ones go on the right, into growth. A fund lands wherever the stocks it holds rate.

For most of value investing's history, a stock got cheap because something had gone wrong: a bad year, a product flopped, a strategy needed fixing. Investors priced a temporary problem as if it were permanent. If the business recovered, the share price eventually caught up with its earnings.

But the box sees only price and profits. It can't tell the difference between a business simply having a bad year and one that's permanently in decline. It doesn't know whether a rival, a new technology, or a free substitute is coming. And when profits do shrink, the stock can look like a bargain while the business erodes. Investors call that a value trap: a stock that looks cheaper the further it falls, until the profits disappear and it turns out it was never cheap at all.

Take two companies that look like twins on a P/E screen. Union Pacific (UNP), America's biggest freight railroad, trades at about 23x its earnings. FactSet Research (FDS), one of the largest financial-data platforms, trades at about 22x earnings. Investors are paying roughly the same price for each dollar of profit.

FactSet sells information. Analysts pay it for prices, estimates, and research, and for software that turns those numbers into models and slide decks. AI now does that work faster and more cheaply. In February, leading AI maker Anthropic added finance tools to its Claude Cowork assistant that can pull market data, update a spreadsheet model, and build the presentation in one session.

By contrast, Union Pacific runs freight trains over more than 32,000 miles of track through 23 states, along with the land, bridges, tunnels, and yards that go with them. To see what new track costs, look at Utah's Uinta Basin Railway, an 88-mile line to local oil fields. Five years after regulators approved it, construction still hasn't started. Its estimated cost has more than doubled, to about $3.4 billion, or nearly $39 million a mile. Even at a quarter of that rate, rebuilding Union Pacific's network would cost more than the whole company's $176 billion market value.

Now ask the management of Union Pacific and FactSet the same question: Could a well-funded startup, armed with the latest AI, rebuild your company in five years?

For FactSet: In my view, possibly. An AI-enabled competitor might be able to reproduce some of the analytical tools customers pay for at a lower price. That is a possibility, not an established conclusion: replicating FactSet's data access, customer relationships, and integrated workflows could present substantial obstacles.

For Union Pacific: In my view, that would be far less plausible. AI can help schedule trains and route cars. It can't lay track, buy land, or get a permit.

That's what the P/E can't see. With FactSet, investors are betting its earnings survive a technology that gets better every month. With Union Pacific, they're betting on something AI can't magically produce: 32,000 miles of steel.

The market has started to notice. Since the start of 2025, Union Pacific's shares have risen 28%, while FactSet's have fallen 32%. Two years ago, investors paid about 30x FactSet's earnings. Today they pay about 22x, the same as for a railroad.

Same price, very different risks.

Valuation metrics remain widely used and continue to provide important information regarding company fundamentals. However, investors often consider additional qualitative and quantitative factors when evaluating investments.

FactSet's future is still an open question. For one company, the answer has already arrived.

Chegg (CHGG) charges students a monthly subscription for step-by-step answers to homework and textbook questions. It had already been knocked down once. Remote learning during the pandemic sent its shares above $100 in early 2021. When students went back to campus, enrollments fell. On November 2, 2021, the stock dropped 49% in a single day. By April 2023, it was below $18.

That first fall looked like the kind of problem value investors love: a bad patch that should pass. The second one was different. On May 1, 2023, Chegg's CEO told investors that students were turning to ChatGPT, which did the same job for free. The next day, the shares fell another 48%, to just over $9.

The revenue told a slower story. In 2023, Chegg's sales slipped 7%. In 2024, they fell 14%. In 2025, they dropped 39%, to $377 million, about half of what the company brought in two years earlier.

In between, Chegg looked cheap on paper. At the end of 2022, investors paid about 21x its free cash flow, which is the cash a business has left after paying its bills and investing in itself. By the end of 2023, they paid about 7x. By the end of 2024, about 3x. But by the end of 2025, there was no free cash flow left to measure. That's the gap a value trap lives in: the price has already moved, but the numbers that make the stock look cheap haven't caught up yet.

The lesson isn't that every cheap stock is doomed. It's that investors should ask why it's cheap. Sometimes the market has overreacted. Other times, it's recognizing a problem before the rest of the world does.

The lesson isn't that every cheap stock is a value trap. It's that investors should understand why a stock is cheap before concluding it represents an opportunity.

Chegg isn't the only company that’s become a victim of AI. Research published on September 30 by index provider MSCI found that companies whose labor costs AI could most easily automate did worse than similar companies. Those with protections that are hard to copy – such as physical assets, regulation, and locked-in customers – held up better. Among technology stocks, the more protected companies beat their peers by 15% in stock returns.

This isn't the first time someone has buried value investing.

Warren Buffett held the funeral service himself. He learned from Benjamin Graham, the Columbia professor who gave us the phrase "margin of safety," and he started his career buying "cigar butts": struggling companies so cheap they had one last puff left in them. Years later, in his 2014 letter to shareholders, Buffett explained that the approach "worked very well while I was managing small sums" but was "scalable only to a point." Charlie Munger, Buffett’s business partner at Berkshire Hathaway, pushed him to stop buying mediocre businesses because they were cheap, and to buy great ones at fair prices instead. As Buffett summed up the lesson in 1989: "It's far better to buy a wonderful company at a fair price than a fair company at a wonderful price."

The second funeral came with the computer screen. Once every trading ratio sat on every screen for free, cheapness alone stopped being an edge. Value stocks trailed growth stocks for much of the 2010s, and "value is dead" became a common refrain. Defenders argued that the cheap companies weren't broken, just out of favor.

The third funeral is happening now, and this time, some of the businesses may be broken. AI can make the earnings behind a low P/E disappear. That doesn't just make bargains harder to find. It may undercut the idea that a low price gives you a margin of safety.

Porter’s longtime friend Dan Ferris has written a value newsletter since 2002. When he joined me on my podcast in May, he said, "P/E just doesn't mean what it used to." Value investors have long preferred businesses that need little capital, because a brand was harder to copy than a factory. Ferris points to a different kind of protection: chemical makers tied to their customers by pipelines and rail lines. "You're not gonna say, well, you're charging me too much, so I'm going to build another railroad."

When intelligence gets cheap, what lasts may be whatever cost the most to build.

Here are three questions to consider when evaluating a cheap stock.

1. What does this business sell? Information, analysis, and routine paperwork are getting cheaper to produce. Track, pipe, and power lines aren't.

2. Could a well-funded startup, armed with AI, rebuild it in five years? If so, a low price may be a warning, not a gift.

3. Is the share price falling faster than the revenue? If so, find out why. The market may have seen something the income statement hasn't shown yet.

None of this makes heavy-asset businesses safe. They need constant spending to keep running, and higher interest rates, rising costs, or a recession can hit them hard.

The second question is the idea behind the Tuttle Capital Heavy Assets Low Obsolescence ETF (HALX), which I manage. Its index scores large U.S. companies on how much they invest in physical assets, how steady their cash flows are, and how little they rely on intangible assets. Valuation isn't one of the measures.

Porter's own index asks a version of the same question. The Porter & Company Porter Portfolio Index ETF (PCPP), which I also manage, puts a quarter of its money in hard assets and a quarter in property-and-casualty (P&C) insurers. AI can help an insurer run its business, but it may not easily replace one. A startup can write the software, but it can't conjure the capital and licenses it needs to pay claims when a hurricane hits.

A new fund, Porter & Company Property & Casualty Index ETF (PCPC), holds just the insurance stocks. It weights them by underwriting discipline, meaning how little of each premium dollar goes out in claims and costs, rather than by market value.

Cheap used to mean the market had made a mistake. Today, it more often means the market has done its homework. Before you buy, make sure you've done yours.

Matt Tuttle
CEO, Tuttle Capital Management
Riverside, Connecticut

P.S. At the Porter & Co. conference, I closed on the T in my HEAT framework: Themes, the long-term shifts that decide which businesses will still matter. Every morning before the open, I apply that framework to the day's market in The Daily HEAT. It's free. Sign up here.

Tell us what you think of today’s Daily Journal: [email protected]

Straight From The Porter & Co. Annual Conference:

The Big Five Report

Our team just hand-picked the five highest potential opportunities from this year’s event, including Porter’s Trade Of The Decade.

>> Get your copy here <<