We got more defensive last week as Wall Street raised the bar for AI stocks

The key regime shift is not AI demand decay; it is the market demanding shorter payback periods and cleaner funding paths. Higher front-end yields raise the discount rate on every long-duration capex story, so the names with the most visible near-term cash conversion and the least balance-sheet ambiguity should keep winning relative to the rest of tech. That argues for continued rotation out of the broad “AI buildout” complex and into more defensive, fee-driven, or dividend-supported exposures.

Within semis, the differentiation is becoming ecosystem control versus pure infrastructure monetization. NVDA still has the best shot at sustaining a premium because software, developer lock-in, and platform effects make its demand stickier than hardware-only exposure, while AVGO is increasingly hostage to customer concentration and questions about how much of the AI spend pool is really financeable at current rates. If yields stay elevated, the next leg of underperformance should hit the lower-quality AI beneficiaries first, then ripple into adjacent data-center suppliers that rely on stretched hyperscaler budgets.

The contrarian point is that the market may be over-penalizing strong execution in cybersecurity and underestimating how defensive IT spend becomes when macro volatility rises. PANW should still be treated as a structural beneficiary, but near-term price action can remain choppy because investors are using strength to de-risk. The catalyst path is clear: if yields back off or CPI cools over the next 1-3 months, the AI complex can re-rate quickly; if they do not, factor rotation into cash-flow stability can persist for 6-18 months.