Does anyone remember the HALO trade? It was a big deal – for a while – because it marked a big change in market regime.

It stands for Heavy Asset, Low Obsolescence – and refers to the sudden popularity of a group of hitherto unfashionable companies that make expensive, real products (hence Heavy Asset), that last a long time and are not easily replaced (Low Obsolescence).

HALO (coined by Josh Brown at Ritholtz Wealth Management in early 2026) started after a combination of Donald Trump’s first term as president of the United States, the aftermath of the COVID-19 pandemic and Russia’s invasion of Ukraine led to:

·       a breakdown in just-in-time global supply chains

·       a surge in energy costs

·       panic about suddenly shaky American security guarantees.

Almost overnight, it seemed, every nation scrambled in response:

·       to build (or rebuild) domestic manufacturing capabilities

·       to secure its own energy supplies

·       to reverse decades of dwindling defence spending.

Investors clamoured to own Rheinmetall, a maker of tanks, having ignored it for most of this century, along with other businesses that employ lots of engineers who make things in factories. HALO was a big change from the first 20 years of the 21st century when the exact *opposite* type of company dominated markets (so-called “asset-light” software businesses with products that continually became obsolete, requiring rapid upgrades). Stock markets were dominated by software which was eating the world. Building anything physical was deeply unfashionable.

Today’s trillion-dollar tech titans seem to be in the worst quadrant of this asset vs obsolescence matrix. Microsoft, Amazon, Meta, Alphabet, OpenAI and Anthropic are spending trillions of dollars on capital equipment they expect to last for five or six years. On top of this, SpaceX also launches satellites into orbit which last perhaps five years, before burning up in the Earth’s atmosphere – or, as happened last week, crashing into the moon.

The payback necessary to justify this expenditure is unfathomably vast, in their own words: the SpaceX IPO filing claimed a total addressable market (TAM) of $28 trillion. Success for the AI companies would lead to an unprecedented degree of social upheaval: “extreme levels of inequality”, as predicted by Anthropic’s CEO, Dario Amodei, not to mention “that AI will disrupt 50% of white collar jobs over 1-5 years”.

Economies and societies are not mechanical. People won't simply stand by and allow such changes to take place without significant countermeasures, perhaps even revolutions.

In other words, these HAHO AI companies build real things that last a few years and have to be replaced all the time, at ever greater cost, with only one outcome, a complete transformation of society as we know it, to make the staggering costs worthwhile.

On the face of it this is a terrible prospect for investors. “Trust me, once we’ve spent the first two trillion dollars, we’ll make more money than any company in history.” What if most portfolios are hitched to the biggest AI hallucination of all?