The main implication is not that AI is suddenly changing retail economics; it is that the spend is still flowing upstream to the infrastructure layer, where NVDA captures budget before retailers or cloud vendors can prove ROI. For WFAFY, the market is likely underpricing how much of the current resilience is defensive share-gain versus true incremental demand, while overpricing the durability of free cash flow once finance costs and capex step up.

Second-order, the AI/marketplace push looks more like a margin-defense tool than a growth accelerator: useful for retention and labor productivity, but not enough to lift valuation unless it translates into sustained ROIC expansion. That leaves GOOGL as only a marginal beneficiary from cloud/AI adoption, while NVDA remains the cleaner expression of enterprise AI capex. The broader retail lesson is that value operators can keep taking share in a weak consumer, but the upside is capped if baskets stay small and promotional intensity normalizes.

Risks sit over 1-3 months in consumer data, rates, and management commentary on FY27 capex/interest expense; over 6-18 months the issue is whether the mix shift toward higher-investment businesses lowers the group’s capital-light profile. The contrarian miss is that this is not just a defensive compounder story: if capital intensity stays elevated, investors may eventually re-rate it more like a slow-growth industrial than a cash-return retail platform. For NVDA, the falsifier is any sign that guidance growth is decelerating faster than expected or that gross margin is being competed away.