AI Is In A State Of Panic: Start Buying It

The selloff looks more like forced de-grossing than a clean reassessment of terminal demand. In that setup, the first money back in usually goes to the names with the best balance-sheet quality, the clearest order visibility, and the tightest supply constraints, because those are the ones institutions can defend on earnings season and on margin. That argues for a faster rebound in the infrastructure layer than in the broader “AI beta” basket, especially where pricing power can still outrun incremental capital intensity.

The second-order read-through is that this is a relative-value event as much as an absolute one. If AI revenue is indeed scaling faster than the cost base, then the market should eventually re-rate the suppliers that monetize each incremental dollar of customer spend, while punishing adjacent names whose AI exposure is mostly narrative. The risk is that capex sustainability is being inferred from a narrow window; if hyperscaler budgets pause after this digestion phase, the rebound in semis will stall before the next guidance cycle.

Contrarian view: the crowd may be too quick to declare the flush a healthy reset. When positioning is crowded and leverage is high, prices can overshoot fundamentals in both directions; a sharp bounce can happen even if the medium-term multiple stays capped. The key falsifier is not today’s tape but whether the next round of cloud and enterprise spending commentary confirms that AI monetization is broadening faster than depreciation, inventory buildup, and financing costs.

Over a 1-3 month horizon, the cleanest setup is a rebound trade, but over 6-18 months the winners should be the vendors with the strongest installed base and the least dependence on perpetual narrative expansion. If AI spending growth merely normalizes instead of re-accelerating, the high-multiple second-tier names will likely lag even as the index recovers.