A New Framework for a Lopsided Market
Today's equity markets are unusually concentrated around a single thesis. The hyperscalers, the so-called Magnificent Seven, and a rapidly expanding universe of memory and semiconductor names have powered most of the gains investors have enjoyed in recent years. Even passive investors who simply hold an S&P 500 index fund find themselves heavily tethered to the artificial intelligence trade — nearly 45% of the index's constituents have meaningful AI exposure. While that concentration has been broadly positive, it also leaves portfolios vulnerable to the inevitable fits and starts that come when markets begin to question the magnitude of AI-related spending and the timing of its eventual payoff.
The traditional answer to this kind of imbalance would be to tilt toward "value." But a closer look at conventional value and growth indices reveals that they are largely built on simple price-multiple screens such as price-to-earnings ratios. The result is that you can find surprising overlap and unexpected names in both buckets — and neither lens really speaks to the central risk of our moment, which is the risk of being disrupted by AI itself.
Introducing HALO: Hard Assets, Low Obsolescence
A more useful framework has emerged under the acronym HALO, which stands for Hard Assets, Low Obsolescence. The concept, first coined by Josh Brown, focuses on a simple but powerful question: which companies are genuinely resilient against a future in which AI continues to disrupt one industry after another? Rather than sorting stocks by valuation multiples, the HALO approach drills into what a business actually does. Does it own meaningful hard assets? Does it operate within embedded networks that cannot be easily replicated? Is its business model genuinely durable in the face of large language models, no matter how capable those models become?
The companies that satisfy these criteria look very different from the names that dominate today's growth indices. Top holdings in this category include TFI International, Cummins, AutoZone, Philip Morris, and Matson. These are firms anchored in shipping networks, freight infrastructure, durable consumer goods, and physical equipment. A copper or gold miner sits comfortably in this category. So does a maker of diesel engines and backup power generators — the kind of equipment that simply cannot be replicated or replaced by software, no matter how sophisticated. The need for HVAC equipment, freight railcars, and the physical movement of goods around the world is not going away because of a large language model.
This is, in a sense, an "anti-AI" basket — not because the companies oppose AI or fail to benefit from productivity gains, but because they sit in parts of the economy that AI cannot meaningfully cannibalize.
The Counterweight to a Memory-Driven Market
Importantly, the HALO thesis is not a bearish call on AI. The same investment philosophy that produces HALO-focused vehicles can coexist with deep conviction in AI itself. Memory chips, for example, have emerged as one of the most critical bottlenecks in the entire AI buildout. The demand from data center expansion has made memory one of the most consequential segments of the semiconductor industry, and dedicated memory-focused ETFs have grown at a remarkable pace — among the fastest-growing fund launches in history — precisely because investors recognize how essential memory is to the broader AI infrastructure trade.
Generative AI funds, Magnificent Seven baskets, and memory-focused vehicles all express genuine bullishness about the future of artificial intelligence. The HALO approach is not their opposite; it is their complement. The point is portfolio construction. If an investor is already heavily exposed to AI through their core holdings — and most investors are, whether they realize it or not — they need a tool to hedge that exposure intelligently, and the existing options have largely been inadequate to the task.
Rethinking the Growth-Value Divide
Perhaps the most important implication of this framework is that the old growth-versus-value debate is becoming less useful for thinking about portfolio balance. The more relevant question for the years ahead is a different one: how much AI exposure do I have, and how much HALO exposure do I have? That single reframing changes how an investor approaches diversification. It moves the conversation away from accounting metrics and toward the underlying economic reality of what each business does, what assets it owns, and what risks it faces from a wave of automation that is still in its early innings.
For investors who believe AI will continue to reshape industries — and there is plenty of reason to believe it will — the strongest hedge is not a cheaper version of the same trade. It is exposure to businesses that operate on a different axis entirely: businesses rooted in physical infrastructure, embedded networks, and durable goods. In a market that has become extraordinarily concentrated around a single thesis, owning the counterweight may prove just as important as owning the thesis itself.