‘The economy is increasingly reliant on AI gains’: U.S. GDP grew 2.2% amid ‘sudden reversal of optimism’ on AI tech

U.S. second-quarter GDP growth was revised up to a solid 2.2% annualized rate from an initial 1.5% estimate, although it slowed from 2.5% in Q1. Consumer spending accelerated to 3.8% from 0.7%, while nonresidential business investment rose 9%, driven by the AI investment boom; underlying domestic demand grew 4.6%. Imports surged 12.6%, subtracting nearly 1.7 percentage points from GDP, and the economy remains exposed to an AI-optimism reversal and elevated energy prices linked to the Iran conflict.

Analysis

The relevant equity signal is not broad “growth,” but an increasingly narrow profit pool: AI-linked capital expenditure and asset-price-sensitive consumption. Favor the picks-and-shovels layer—VRT, ETN, PWR and CEG—over the highest-duration semiconductor beneficiaries, because power, cooling and grid bottlenecks convert announced compute spending into multi-quarter backlog and pricing power. The near-term risk is that imported equipment represents a pull-forward of deployment rather than a durable acceleration in end-demand; hyperscaler capex guidance over the next two earnings cycles is the key validation point.

A resilient demand backdrop reduces the probability of an imminent policy easing cycle, which is more consequential for valuation than for current earnings. If the 10-year Treasury remains above 4.25% or moves higher on inflation-sensitive data, expensive software, REITs and long-duration small caps face multiple compression even while nominal activity remains firm. That creates a better setup for cash-generative payment networks and affluent-consumer exposures—V, MA, RCL, CCL—than for rate-dependent housing and broad lower-income retail.

Consensus may overread the apparent domestic resilience as uniformly bullish for cyclicals. A consumption mix concentrated in higher-income households typically benefits premium services and travel but is less supportive of mass merchants, subprime lenders and discretionary staples volumes; XRT is therefore a poor proxy for the consumer impulse. Over 6-18 months, continued AI infrastructure spending also raises the odds that power availability, not chip supply, becomes the binding constraint, supporting regulated utilities with credible data-center load pipelines while increasing regulatory and financing risk for those unable to earn returns on accelerated capex.

Key Decisions for Investors

• Initiate a 3-6 month pair: long VRT and ETN / short IGV. The long leg has backlog and physical-capacity support; the short offsets duration risk in premium software. Target 15-20% relative upside, with a stop if either company cuts data-center order commentary or the 10-year Treasury falls below 3.85%.
• Add V and MA on weakness ahead of the next monthly retail-sales and card-spend updates; prefer them to XRT for exposure to higher-income consumption. Expected 6-12 month risk/reward is roughly 2:1, but exit if cross-border volume and processed-volume growth both decelerate for two consecutive monthly data points.
• Maintain an underweight in IYR and selectively short rate-sensitive homebuilders such as DHI against long PWR. This is a 1-3 month rates-and-capex expression; cover the short if the 10-year Treasury breaks below 3.85% or mortgage rates decline enough to produce a sustained improvement in purchase applications.
• Watch CEG, VST and NRG for evidence that incremental data-center load is translating into contracted capacity and permitted generation. Do not chase on headline demand alone: initiate only after disclosed contracted-load growth or power-price hedging visibility improves, as regulatory intervention and project delays are the principal thesis risks.