HALO Stocks

The Setup

In early February 2026, Josh Brown, CEO of Ritholtz Wealth Management, coined the term HALO on his blog and on CNBC's Fast Money Halftime show. The acronym stands for Heavy Assets, Low Obsolescence. The thesis: in a world where AI can write code, replace call center agents, and automate financial analysis, the companies that own physical things, power plants, pipelines, railroads, factories, copper mines, are the ones AI cannot disrupt. These are the stocks you want to own when the SaaS-pocalypse arrives.

The timing was impeccable. On February 23, Anthropic announced that Claude Code could modernize COBOL, sending IBM down 13% in a single session, its worst day since 2000. Software stocks cratered across the board. Investors panicked and rotated into anything with a physical asset base. Generac surged 67% year to date. FedEx climbed 24%. Caterpillar gained 17%. The S&P 500 Equal Weight ETF outperformed the cap-weighted index by 6.4 percentage points in the first two months of the year, one of the widest gaps on record.

The HALO framework captured something real. But somewhere between the insight and the trade, the analysis got lost. A label that correctly identifies what AI cannot disrupt is being used as a substitute for the work of evaluating whether the stocks carrying that label are actually worth buying at today's prices.

What HALO Gets Right

The core observation behind HALO is not wrong. It is, in fact, one of the more useful mental models to emerge from the AI anxiety cycle.

For 15 years after the financial crisis, the market rewarded asset-light businesses above all else. High margins, low capital requirements, recurring subscription revenue, and scalable software economics were the blueprint for every IPO and every analyst upgrade. Physical assets were considered burdens. Capital intensity was a negative signal. The ideal business owned nothing and charged for everything.

AI is inverting that logic. The companies most vulnerable to disruption are precisely the ones the post-2008 market loved most: software companies with high margins, low switching costs, and products that an AI agent could replicate or replace. Meanwhile, the companies the market ignored, businesses that own pipelines, generators, rail networks, factories, and mines, suddenly look like the durable moats they always were. No large language model can build a nuclear reactor, operate a copper mine, or deliver a package to your door.

Brown's insight is that this inversion is not a trade. It is a regime change. The assets that AI cannot replicate are becoming the new premium, and the assets that AI can replicate (code, reports, analysis, customer service) are being repriced downward. That observation is correct and valuable.

The problem is what happens next.

A Label Is Not a Valuation

When an acronym gets adopted by Goldman Sachs, printed on Nasdaq.com, and repeated on every financial podcast in America within six weeks, something important shifts. The label stops being an analytical tool and becomes a buying signal. People stop asking "is this company a good investment at this price?" and start asking "is this company HALO?" Those are fundamentally different questions.

HALO tells you one thing: this business has heavy physical assets that AI is unlikely to render obsolete. That is useful information. It tells you something about the durability of the revenue stream and the competitive moat.

But it tells you nothing about whether the stock is cheap or expensive. It tells you nothing about whether earnings are growing or declining. It tells you nothing about whether the current price already reflects the HALO premium. And it tells you nothing about whether the company is well-managed, appropriately capitalized, or generating returns above its cost of capital.

In other words, HALO is a filter. It is not a thesis. And treating a filter as a thesis is exactly how investors overpay for narratives.

The Exxon Problem

Exxon Mobil (NYSE: XOM) is the poster child for HALO. Josh Brown cited it as his leading example. It is a massive company with physical assets that no AI model can replicate: refineries, pipelines, drilling rigs, chemical plants. It is the definition of Heavy Assets, Low Obsolescence. And the stock is at all-time highs.

Here is what the HALO label does not tell you.

Exxon is trading at 24 times trailing earnings, a 60% premium to its own 10-year average. The Iran war has boosted near-term earnings (Exxon signaled up to $2.9 billion in additional Q1 upstream profit from higher oil prices), but the stock has already rallied 30% to price that in. Wall Street consensus targets of $140 to $144 actually imply downside from the current price, meaning even the analysts who cover Exxon think the rally has overshot. The stock is at all-time highs on earnings driven by a war premium that is, by definition, temporary and unpredictable.

This is the value trap in action. The company is real. The assets are real. The immunity from AI disruption is real. But the price has detached from the earnings power. Investors who buy Exxon here because it is HALO are making the exact same mistake as investors who bought Palantir at 186x earnings because it was "at the epicenter of the AI revolution." The label is different. The analytical error is identical: substituting a narrative for a valuation.

Exxon is not the only example. Multiple names on the popular HALO lists are trading at stretched multiples relative to their own histories, pushed there by the rotation rather than by fundamental improvement. When the entire market buys the same label simultaneously, the label gets priced in. And once it is priced in, the margin of safety disappears.

Where HALO Actually Works

The HALO framework becomes genuinely useful when it is combined with fundamental analysis, not used as a replacement for it. The question is not "is this stock HALO?" The question is "is this stock HALO and does the data support the price?"

There are sectors where both conditions are met.

Utilities with contracted demand growth. Companies like NextEra Energy, Southern Company, and Constellation Energy own the physical power generation infrastructure that AI data centers need, and they have multi-decade contracts with hyperscalers to prove the demand is not speculative. These are HALO by definition (you cannot replace a nuclear reactor with a language model), but the investment case is built on specific financial data: contracted revenue, rate base growth, and earnings visibility. The HALO label is additive. It is not the thesis.

Defense primes with record backlogs. Lockheed Martin, RTX, and General Dynamics own the manufacturing capacity for weapons systems that no AI model can produce. They also have $580 billion in combined backlog, a proposed 188% increase in missile procurement, and multi-year revenue visibility backed by sovereign credit. These are HALO because the assets are physical and undisruptable. They are investable because the earnings are growing, the dividends are covered, and the valuations reflect the business, not just the label.

Infrastructure beneficiaries of the physical buildout. Companies like Caterpillar (construction equipment for data center sites), Deere (agricultural equipment), and Vulcan Materials (aggregates for construction) are HALO because their products are physical and essential. But the quality of the thesis depends on whether the earnings support the multiple, and that varies company by company.

The pattern is clear: HALO works when it identifies a durable moat, and the investor then does the additional work of verifying that the earnings, cash flow, and valuation support the price. HALO fails when the label substitutes for that work.

The Missing Half of the Framework

What HALO provides is a disruption filter: which companies are AI-proof? What HALO does not provide is a valuation filter: which of those AI-proof companies are worth buying at today's price?

The missing half of the framework is the part that actually determines whether you make money. Being undisruptable does not make a stock a good investment. Being undisruptable and underpriced relative to earnings power makes a stock a good investment. Those are not the same thing, and conflating them is how the HALO rotation eventually becomes its own bubble.

Consider the parallel. In 2020 and 2021, the market decided that "digital transformation" was the only theme that mattered. Anything SaaS, cloud, or e-commerce got a premium multiple. The framework was correct: digital was the future. But the valuations detached from the earnings, and when they reverted, investors lost years of returns. HALO could follow the same path if investors treat the label as sufficient and stop looking at the financials underneath.

The cure is the same discipline that has always worked in investing. Look at what the company actually earns. Look at what you are paying for those earnings. Look at whether the earnings are growing or declining. Look at the cash flow, the balance sheet, and the competitive position. Then decide if the price is justified.

The HALO label might be what gets a stock on your watchlist. But the financial data should be what gets it into your portfolio.

The Bottom Line

The HALO framework is a useful contribution to how investors think about portfolio construction in the AI era. Josh Brown identified something real: the market is repricing what counts as a moat, and physical assets that AI cannot replicate are becoming more valuable relative to software assets that AI can.

But a framework is a starting point, not an endpoint. The moment a Wall Street acronym becomes a buying signal, it stops being analysis and starts being narrative. And narrative, however compelling, is not a substitute for reading the financial statements.

Some HALO stocks are genuinely well-positioned with strong earnings, growing backlogs, and reasonable valuations. Others are riding the label to decade-high multiples on flat or declining earnings. The difference between the two is not the label. It is the data.

At Wealth Engine Pro, the philosophy is to evaluate companies based on what they are, not what a label says they should be. HALO is the latest version of an old pattern: a correct observation about the market gets turned into an acronym, the acronym gets turned into a trade, and the trade gets crowded before most investors have looked at the numbers underneath. The investors who will do well in HALO stocks are not the ones who bought the label first. They are the ones who bought the fundamentals first and happened to find HALO characteristics along the way.

Data over narrative. It applies to AI hype. It applies to sell-side buzzwords. And it applies to catchy acronyms printed on trading desks across Wall Street.