Listen to the audio version of this article (generated by AI).

Hello, Reader.

Only 3% of companies remained.

That’s what happened by 1934, little more than a decade after the birth of commercial radio.

More than 600 firms stampeded into radio manufacturing during the boom of the 1920s. Less than 15 years later, all but 18 were gone.

The companies that survived – and eventually thrived – were not the radio builders. They were the radio appliers:

• The advertisers who used the airwaves to reach millions of consumers for the first time…
• The retailers who built national brands through radio sponsorships…
• The entertainment companies that turned programming into profit.

The contrast says it all.

RCA – the radio builder, the Nvidia Corp. (NVDA) of its day – soared some 200-fold in the 1920s. Then it crashed 98% between 1929 and 1932 and didn’t reclaim its peak until the 1960s.

Procter & Gamble Co. (PG) – the radio applier – invented the soap opera to move more soap. While its stock fell in the crash like everything else, the company paid a dividend every single year straight through the Great Depression, a streak unbroken since 1890.

One took 30 years to get back to even. The other has been paying its shareholders for 136 years and counting. So, radio was only as valuable as what companies chose to do with it.

History suggests that the greatest returns of the AI era will go not to the companies that build the technology, but the companies that figure out what to do with it, profitably. All else being equal, I trust history.

So, in today’s Smart Money, let’s look beyond the companies building the AI machine – and toward the companies learning how to make money with it.

Then, I’ll show you where to find these unexpected winners.

Let’s jump in…

When Great Earnings Aren’t Good Enough

Consider what happened last week to Hewlett Packard Enterprise Co. (HPE), a leader in essential enterprise technology and one of the companies building the physical guts of the AI boom.

Last week, the Houston-based company reported the best quarter in its history. It reached record revenue of $12.2 billion, up 34% from a year earlier, as well as record earnings of $1.11 per share – the company’s first time surpassing a dollar in a single quarter. Plus, management was so confident that it raised its guidance not once but twice during the same call.

The results were great, yet the stock fell about 5%.

The reality is that when you buy one of the strongest builders of technology, you’re often buying into expectations that it must continually surpass expectations rather than just meet them. It’s Sisyphus’s boulder, AI edition.

And HPE isn’t alone. It’s a pattern that has defined this entire earnings season.

According to Morningstar, AI hardware stocks – the chipmakers, the memory firms, the networking firms – keep posting massive earnings, often beating Wall Street’s estimates outright. But their stocks keep falling anyway.

Despite a remarkable 750% year-to-date gain as of late June, SanDisk Corp. (SNDK) fell 13% over two sessions after reporting weak guidance, even though quarterly sales had increased by 50%. The stock now boasts about a 640% year-to-date gain. That, of course, isn’t nothing. But even with an extraordinary run, SanDiskhas already given back a meaningful chunk of its gains.

Why the punishment? Because after doubling, tripling, or quadrupling in value, technology builders often become priced for perfection.

The lesson here is that past performance tells us where the money has already gone, not where it’s going next. That’s an important revelation: The easy gains from simply owning the builders are getting harder to come by.

Like the hundreds of companies that rushed to build radios a century ago, HPE and SanDisk are helping build the technology everyone wants. But the next class of winners may be the companies using that technology to make their existing businesses more valuable.

That takes us down to the farm…

AI: The New Farmhand

Deere & Co. (DE), widely known as John Deere, is a manufacturer of farm machinery and industrial equipment. But it is also becoming a good example of how a traditional company can use AI to make its existing business more valuable.

Deere launched its AI assistant, “JD,” earlier this month at the Farm Progress Show in Boone, Iowa. JD is built into the company’s John Deere Operations Center, where farmers can ask questions about their farms. They can analyze years of their own data to help them make better decisions.

In other words, JD puts an AI layer on top of the data Deere has already collected. And it’s already helping customers save money and potentially produce more.

According to The Wall Street Journal, one Iowa farmer cut corn-seed use from 34,000 seeds per acre to 29,000, while using about 40% less fertilizer. Another said Deere’s AI-powered sprayer saves him about $75,000 a year on herbicides.

And Deere has a very good reason to make sure farmers see that value. Its farm-equipment business has been hurt by weak demand. But the downturn may be nearing its end. CEO John May said 2026 could be the bottom of the current farm-equipment slump. The company hit a record high this week.

This creates an interesting setup: Its old business may be recovering just as AI gives it a new way to make that business more valuable.

Deere isn’t an AI company. It’s an AI Applier – a traditional business using increasingly powerful technology to improve the business it already knows best.

And it’s not the only one…

Don’t Buy the Technology. Buy What Companies Do With It.

When investors poured money into the companies building radios a century ago, the technology was real. The revolution was real. But the fortunes went somewhere else entirely – to appliers like P&G – not the firms like RCA cranking out the radio receivers.

AI Appliers follow the same pattern: They are companies leveraging AI to expand margins, refine pricing, and gain a competitive edge, with profits appearing on their income statements rather than in construction backlogs.

So, the smart move is to reduce exposure to those building AI technology and increase exposure to the appliers of it. These are the companies I’m watching closely.

That is why in the September issue of the Fry’s Investment Report – which will be available tomorrow – I will explore profitable, modest companies quietly turning AI into an untapped advantage the market hasn’t yet recognized.

The important point is that none of these companies needs to build the underlying AI infrastructure. They just need to figure out how to use it better than their competitors.

I will also reveal a new AI Applier recommendation – a company using AI to accelerate the discovery, development, and delivery of the drugs that drive its business. And that AI advantage is only part of the reason I think the market has mispriced this stock.

To receive my latest monthly issue and recommendation as soon as they’re available, click here to join Fry’s Investment Report.

And be sure to keep an eye out in your inbox.

Regards,

Eric Fry