Something strange is happening in the market, and it’s the kind of shift that only shows up clearly in hindsight — unless you’re paying attention right now.
Investors are rotating out of “asset-light” tech darlings and into companies that own hard, physical stuff. Railways. Mines. Pipelines. Defence contractors. The kinds of businesses that, for the past decade, Wall Street dismissed as too capital-intensive, too boring, too old-school. Suddenly, boring looks brilliant.
The logic is straightforward once you see it. If AI actually does what its biggest cheerleaders promise — replacing knowledge work, automating analysis, writing code — then the companies most at risk are exactly the ones the market has been bidding up: consulting firms, software companies, digital services businesses. Their moats evaporate if a large language model can do their job for pennies. But you can’t replace a copper mine with ChatGPT. You can’t automate a freight railroad. Physical assets become the ultimate hedge in an AI-disrupted world.
And there’s a delicious irony here. If AI doesn’t deliver on those grand promises — if it turns out to be useful but not transformative — then the hundreds of billions being pumped into AI infrastructure won’t generate adequate returns. Either way, the AI-everything trade has a problem. Hard assets win in both scenarios.
The team at Ruffer Investment Company points out that several megatrends are converging simultaneously: AI disruption fears, surging defence spending in a tense geopolitical environment, energy security concerns (hello, Strait of Hormuz), and rising healthcare demand. All of these favor companies with “constrained sources of supply” — physical resources and infrastructure that can’t be scaled up overnight or replicated by software.
Goldman Sachs data shows the rotation is already underway. Commodity equities, defence stocks, and infrastructure plays have been outperforming the broader market since late 2025. European defence names have surged. Energy stocks bounced hard. Meanwhile, some of the most beloved capital-light tech plays have been quietly bleeding out.
This doesn’t mean tech is dead — far from it. But the decade-long consensus that asset-light businesses are inherently superior is cracking. In a world defined by geopolitical instability, supply chain fragility, and real questions about AI’s economic payoff, owning things you can touch might be the smartest trade of the next cycle. As one analyst put it: it’s hard to replace a railway with a large language model. That’s the whole thesis in one sentence.