Something interesting is happening beneath the surface of this market selloff, and it has nothing to do with Iran or tariffs. Institutional investors are quietly rotating out of the asset-light, knowledge-economy darlings that dominated the last decade and piling into something nobody’s talked about in years: hard assets.
We’re talking railways, commodity producers, industrial conglomerates, mining companies — the kind of boring, capital-heavy businesses that Wall Street spent the 2010s telling you were dinosaurs. Turns out dinosaurs are pretty hard to replace with a large language model.
The logic driving this rotation is surprisingly simple. If AI actually delivers on its biggest promises — automating knowledge work, replacing white-collar tasks, disrupting professional services — then the companies most at risk are the “capital-light” businesses investors have loved for their fat margins and low capex requirements. Consulting firms, software companies, financial services — these are exactly the sectors that AI threatens to commoditize.
Meanwhile, you can’t automate a copper mine. You can’t replace a power grid with a chatbot. You can’t build a data center without steel, concrete, and an enormous amount of electricity. The AI revolution, paradoxically, is making physical infrastructure more valuable, not less.
Goldman Sachs data shows a clear divergence: stocks with heavy physical assets and constrained supply are outperforming their asset-light peers by a widening margin. Defence stocks are up 11.7% year-to-date while the broader market sits flat. Energy infrastructure is surging. Even old-school industrials are catching bids.
There’s also a geopolitical overlay that’s reinforcing this trend. In a world defined by wars, supply chain disruptions, and energy insecurity, owning things that exist in the physical world — things that are hard to build, hard to replace, and essential to keep civilization running — suddenly looks a lot more attractive than owning things that live on a server somewhere.
This doesn’t mean AI stocks are dead. But it does suggest that the market’s next leadership cycle might look very different from the last one. After a decade of rewarding asset-light business models, the pendulum may be swinging back toward companies that actually build, mine, transport, and generate. If you’ve been ignoring the industrial sector, now might be a good time to dust off those old stock screeners.