The AI rotation stole the spotlight from a strong start to earnings season

IBM’s warning is less a company-specific miss than a signal that enterprise AI budgets are being re-allocated toward control layers and infrastructure, not broad application software. That tends to help cybersecurity and select hardware/memory names first, because they sit closer to the compliance and capacity bottlenecks in AI rollouts; it pressures CRM/NOW-style spend where customers can defer implementation without killing the strategic AI story. The second-order effect is that vendors with usage-based pricing and security attach rates should see better durability than large-suite software vendors over the next 1-3 quarters.

The semis selloff looks more like a crowded-position unwind than a genuine demand break, because the capex guidance from foundry/equipment names still argues the buildout is alive. The market is starting to distinguish between beneficiaries of AI demand and beneficiaries of AI spending, which means names with the cleanest conversion of capex into near-term cash flow should outperform while higher-beta design/IP names remain vulnerable to multiple compression. If oil stays elevated, that also complicates duration-heavy software valuations by keeping inflation expectations sticky and the Fed optionality less supportive.

Contrarianly, the consensus may be underestimating how much of the current AI spend is defensive plumbing versus incremental revenue generation. That favors CRWD/PANW and hardware enablers over application-layer optimism, while the biggest near-term risk to the rotation is a broad risk-off move if energy-driven inflation reaccelerates. On the upside, if CPI stays tame and hyperscaler capex prints remain strong, this pullback in semis could reverse quickly as investors are forced back into the highest operating leverage names.