AI Slowdown Is More Macro Than Tied to One Company: T.Rowe Price’s Wang

T. Rowe Price portfolio manager Tony Wang says investors expect AI-related spending to boost revenue in companies’ core businesses and cloud operations while improving margins. He describes demand for AI chips as unprecedented, but cautions that rapid shipments may strain supply chains, create bottlenecks and raise execution expectations.

Analysis

The key risk is a timing and incidence mismatch: chip shipments can book before customers have enough power, networking, or deployment capacity to turn them into billable AI workloads. In that case, near-term supplier revenue may look strong while cloud utilization, customer ROI, and margins lag—raising the odds of a later capex pause rather than validating a durable earnings cycle. Bottlenecks can also shift value toward suppliers of constrained infrastructure, but only if they can deliver; delays may instead strand spending and defer monetization.

Over days, the commentary is not a standalone catalyst; avoid treating AI demand language as new earnings evidence. Over 1–3 months, focus on cloud revenue growth, AI workload utilization, capex and depreciation commentary, and delivery lead times. Over 6–18 months, the winners should be platforms demonstrating incremental revenue and margin contribution, not merely higher infrastructure spend. The contrarian risk is that investor expectations already price in both demand and successful monetization, leaving shipment strength insufficient to support further multiple expansion. This view weakens if cloud operators show accelerating AI-linked revenue alongside stable or improving margins; it strengthens if capex rises while growth, utilization, or margin guidance disappoints.

Key Decisions for Investors

• Do not add broad AI-chip exposure solely on demand rhetoric. Treat strong shipments without evidence of downstream utilization or customer monetization as a potential pull-forward signal.
• Set an earnings-season alert for cloud revenue growth versus capex and depreciation trends. A widening gap—spending accelerating while growth or margins soften—would favor reducing AI-infrastructure beta.
• Consider a conditional relative-value position: favor diversified cloud platforms over the highest-beta AI hardware exposure only if reported workload monetization improves and capex efficiency holds; otherwise, avoid the pair rather than assume platforms are insulated from spending risk.
• Track power, networking, and delivery constraints as potential bottlenecks that can defer workload deployment. Reassess if companies report easing lead times without a corresponding increase in utilization or revenue conversion.