Money is finding places to hide from the AI selloff. On Friday, six S&P 500 sectors rose even as the index fell 1.09%; energy and utilities led the gainers. That is rotation, not an indiscriminate exit from equities.

But Brent jumped 3.7% to $109.86. The market is trying to diversify away from expensive technology while the oil shock threatens the inflation relief that would make its new refuges more comfortable. Changing sectors reduces one risk. It does not remove the cost of money.

The encouraging interpretation is that investors are becoming more selective. The Dow finished the week higher while the Nasdaq lost roughly 4.5%. A market that can reward businesses outside technology has more resilience than one dependent on a handful of AI names. The danger is assuming that broader participation automatically means lower economic risk.

Energy and utilities illustrate the distinction. Higher oil can improve producers’ near-term cash generation, but it also taxes household spending and raises costs elsewhere. Utilities offer comparatively steady demand, yet new generation and grid investment require substantial financing. Their earnings may be defensive while their share prices remain sensitive to bond yields. Those are different exposures—not interchangeable shelters.

Nor is a collapse in current activity the clean explanation for this week’s anxiety. The Federal Reserve reported September industrial production up 0.7%, including a 0.3% increase in manufacturing. TSMC’s third-quarter revenue reached $46.72 billion, and it guided the fourth quarter to $52.3–54.3 billion. Real output and chip sales are still growing. Strong supplier revenues can coexist with doubts about the returns earned by the buyers funding all that capacity.

Government bonds have begun to absorb some of that anxiety. Thursday’s Treasury rally showed that they can still cushion a growth scare, but Friday’s oil surge challenges the idea that the inflation problem has passed. Long bonds offer more income at higher starting yields; they also carry substantial price risk if persistent inflation forces investors to demand still higher yields. Neither “bonds are broken” nor “bonds are back” is a sufficient portfolio thesis.

Our read: separate operating resilience from financial resilience. Favour businesses that can finance committed investment from cash generation, examine which costs can actually be passed to customers, and keep shorter-maturity high-quality bonds in the defensive conversation. For energy-importing Asian economies—and New Zealand—the same oil rise can squeeze income even when local portfolios own little US technology. A different postcode does not guarantee a different risk.