Investors are piling into “heavy assets, low obsolescence” stocks they hope can’t be clobbered by AI

In 2021, Intel said it would sink $20 billion into two new US chip factories in Arizona — while simultaneously boosting capital spending by almost 50%. The goal, Intel said, was to take on Taiwan Semiconductor in the Asian giant’s key contract manufacturing, or foundry, business. 

It did not go well.

Wall Street was scathing in what it said about the company’s chances. 

“We believe there is almost no chance of the foundry business succeeding,” Citigroup analysts wrote at the time. And Intel shares soon began to slide, losing around 70% of their value over the next few years.

For the better part of two decades, investors have given a skeptical eye to companies making big bets on long-term assets like the kinds of factories and equipment Intel was binging on.

The market wanted “asset-light” companies — firms whose costs consisted largely of employees, intellectual property, software, and other intangibles — that boasted consistently higher profit margins over time.

Not anymore. 

In 2026, large, lumbering, old-economy companies are all the rage, as the data center investment boom upends assumptions on how companies will survive in, and benefit from, the AI era.

The dynamic is obvious in the atrocious market performance of highly profitable asset-light companies like ServiceNow, Salesforce, Workday, and Intuit — software stocks that have been battered by worries they could soon be disrupted by AI.

Josh Brown, the CEO of Ritholtz Wealth Management and cohost of well-known markets podcast “The Compound,” watched the software beatdown happen in January and February — and the concurrent rise of stocks like oil field services firm Baker Hughes, tractor maker John Deere, and shipping giant FedEx, all of which surged more than 30% during that period. Other asset-heavy outperformers this year include freight rail giant CSX, natural gas driller Diamondback Energy, and Verizon.

“The more I was researching what was going on, the more obvious it became to me that this was the trade of the year,” Brown said. “All of the stocks that are susceptible to AI are shedding market cap and portfolio managers are allocating instead to companies that can’t be messed with by Anthropic and Gemini and ChatGPT.”

He minted a mnemonic that tries to capture the essence of what the market seems to be rewarding: HALO, meaning “heavy assets, low obsolescence.”

Whatever you call it, Wall Street has largely come to the same conclusion about the kind of stocks that are working right now.

Revenge of the old economy

Goldman Sachs analysts earlier this month wrote that “investors have re-rated many companies in the ‘old economy’ that have long been neglected and seen underinvestment,” citing sectors like energy, resources, chemicals, and industrials as spots with rising valuations.

“Market moves this year have reflected a shift away from AI disruption risk toward businesses perceived to be insulated from that risk,” they continued.

To be sure, some of that reassessment of asset-heavy companies is because of a sharp rally in energy prices and related stocks as a result of the Iran war.

But such businesses — energy, materials, and telecommunications, for example — also require large amounts of expensive machinery, equipment, and other physical capital that can insulate them from competition. It also sets them apart from software and other purely digital industries more at risk of being made obsolete by AI.

Likewise, Morgan Stanley analysts earlier this year said they expected to see “rising values for assets that cannot be ‘replicated’ by AI.”

“This prediction is unfolding much more rapidly and violently than we anticipated,” they wrote in a note this month, following up on that prediction.

One of the authors of that report, Global Head of Thematic Research Stephen C. Byrd, suggested that electrical equipment companies such as GE Vernova — which is powering the AI boom and has more than tripled in the past 12 months — are good examples of the dynamic at play.

The company’s roots are in General Electric’s old power system manufacturing business, which produced turbines for power plants. It was a solidly profitable, asset-heavy manufacturer that, nevertheless, had fallen out of fashion in recent decades. The arrival of AI as a major force in the US economy changes all that.

“In the future, all those heavy assets will have very different destinies from a profitability point of view,” Byrd said in an interview with Sherwood News. “Anybody who has a power solution that can give data center developers access to power quickly, we’re seeing it. Their margins are skyrocketing.”

GE Vernova is a case in point: the shares have soared on its plans to help supply the turbines to the on-site power plants that data centers increasingly need. 

Byrd also pointed out less obvious providers of power for AI. For instance, cryptocurrency miners that have pivoted toward data center development — such as IREN and TeraWulf— are benefiting from similar trends.

“Some of the bitcoin guys are doing much better than people appreciate turning their bitcoin sites into data center sites,” Byrd said. “And those stocks aren’t well understood, which is why we like them so much.”

Byrd’s comments touch on a potential irony of the HALO trade, however.

While many of the asset-heavy companies that are faring well recently don’t have to worry about being replaced by AI, they are often highly exposed to the AI boom itself.

Deere and Caterpillar, for example, have seen surging sales of their construction equipment, in part because of the rush to build data centers as quickly as possible. 

Utilities like Entergy — up 25% year to date — have likewise been lifted by AI-related deals, such as its plan to provide power to Meta’s big Louisiana data center. If the AI investment surge falters, these stocks will too.

But Brown, the godfather of HALO, thinks that the heavy asset/low obsolescence paradigm will continue to be a good guide to how the market will be looking at stocks this year, as long as the boom continues. 

“The entire market is reorganizing itself,” he said.

Intel might be as good an example as any. Its long-abused US manufacturing operations now seem to be getting a second look from the market, amid seemingly relentless demand for chips tied to AI.

The stock has popped in recent weeks as Intel announced deals aimed at supplying Elon Musk’s Terafab chip fab project as well as a collaboration with Google on building custom chips for Google Cloud centers and AI.

In fact, Intel is up more than 40% in the first half of April alone — making it the best performer in the S&P 500. It’s also having its best 12-month stretch in nearly three decades.

“I think it’s the stock of the year right now,” Brown said of Intel. “Those foundries on US soil, strategically, make it incredibly important, and as a result, incredibly HALO.”