AI Momentum Drives 2Q Earnings: Markets Snapshot

The key takeaway is not that AI demand is fading; it is that the market is moving from a scarcity narrative to a capacity-allocation narrative. That shifts the economic rent toward the infrastructure layer that can actually ship product on time: semicap equipment, networking, power management, cooling, and data-center buildout names. The large platform companies may still spend aggressively, but once the market believes capex is durable rather than accelerating, the multiple expansion upside for the obvious AI winners gets harder to justify.

Near term, the bigger risk is not an earnings miss but a valuation air pocket. A geopolitical shock that lifts oil and rates at the same time is especially toxic for long-duration growth, because it raises the discount rate precisely when investors are paying peak multiples for future AI monetization. In that setup, the first order loser is broad tech beta, while beneficiaries are more defensive cash-flow sectors and energy-linked hedges; even within AI, names with weak free-cash-flow conversion or dependency on aggressive 2025-26 assumptions will underperform.

The consensus is too binary: either AI is real or it is a bubble. The more likely outcome is dispersion—capex stays strong, but returns concentrate in the picks-and-shovels rather than the crowded application layer. What would falsify the constructive infrastructure view is a sequential slowdown in hyperscaler capex commentary, a drop in AI backlog/order intensity, or a sustained rise in Brent and yields that forces budget discipline over the next 1-3 months.