The AI Boom Will Pull Back, Warns Former SEC Chair Gary Gensler

Former SEC Chair Gary Gensler expects the AI investment boom to slow or reverse within 18 to 30 months, citing historical precedent. He says mounting debt could shape how disruptive a correction would be and raises the possibility that the U.S. has overbuilt AI infrastructure; he also discusses interest rates, monetary policy and financial stability.

Analysis

This is a warning about the financing and utilization cycle, not a near-term catalyst: a former regulator’s timing estimate does not establish that spending is peaking. The more important transmission is whether data-center capacity is funded against durable customer demand or against expectations of continued asset appreciation. If utilization or monetization disappoints, order cuts can travel quickly from accelerator and networking vendors into data-center construction, power equipment, and related credit; leveraged projects and lenders would be more exposed than cash-rich buyers. Conversely, slower buildout could ease power and capacity constraints for users and lower future input costs, so an infrastructure correction need not imply an equivalent collapse in AI adoption.

Over days, treat this as a sentiment/positioning input, not a standalone short signal. Over 1–3 months, monitor hyperscaler capex guidance, supplier lead times and order commentary, data-center financing spreads, and evidence that AI revenue or productivity gains are absorbing depreciation. Over 6–18 months, a sustained gap between capacity additions and paid utilization is the bearish structural case. The contrarian risk is that aggregate “overbuild” masks regional, power-connected capacity scarcity and that lower costs accelerate deployment. Thesis is weakened by stable capex plans alongside improving utilization and monetization; it is strengthened by guidance cuts, order deferrals, or widening financing spreads.

Key Decisions for Investors

• Do not trade the 18–30 month forecast in isolation. Keep broad AI exposure sized to tolerate a capex-cycle drawdown; use supplier order trends and buyer guidance as confirmation before adding a directional short.
• Watch for a relative-value setup: favor diversified platforms with demonstrable AI monetization over capital-intensive infrastructure suppliers if capex decelerates but adoption remains intact. Require evidence in earnings disclosures; otherwise keep this as a watchlist, not a position.
• Track data-center credit spreads and project financing alongside hyperscaler capex. A sharp spread widening combined with order cancellations would make infrastructure-linked credit and high-beta semiconductor exposure more vulnerable; stable spreads and firm orders would argue against the bearish case.
• Use a three-part falsification checklist for any underweight: capex guidance remains intact, supplier lead times/orders stay firm, and disclosed utilization or AI revenue improves. Reassess if all three hold through the next reporting cycle.