CHAT: AI Story Derailed By Macro Factors, Growth Story Remains Strong

The most important second-order read-through is that AI infrastructure is increasingly behaving like a multi-year utility buildout rather than a discretionary tech capex cycle. That matters because the winners are no longer just the obvious hyperscalers; the durable beneficiaries are the picks-and-shovels names with pricing power and bottleneck exposure in power, networking, semis, and thermal management. A sustained 2026-2027 spend ramp should widen dispersion between AI-capex enablers and software/application names that are being valued on near-term monetization rather than throughput demand.

Near term, the market is likely underestimating the volatility created by capex concentration: when a handful of large buyers dominate demand, suppliers can get multiple expansion on order-flow visibility, but they also become vulnerable to a single customer delaying or re-phasing budgets. The biggest risk is not demand destruction, but timing slippage from permitting, power availability, and public pushback that can stretch delivery into the right tail of 2026/27. That suggests the trade is better expressed through names tied to physical constraints than through generic AI beta alone.

Consensus may be too relaxed about inflationary pressure because higher data-center build costs can squeeze returns before they hit revenue. If financing costs stay elevated, the market may start rewarding balance-sheet strength and recurring cash flow over pure growth exposure, especially if the AI theme enters a digestion phase after strong multiple expansion. The contrarian view is that the setup is still constructive, but the easy money is in infrastructure quality rather than thematic breadth; broad AI baskets may lag the concentrated winners once investors distinguish capacity creators from capacity consumers.