A Wall Street Journal Report on $3 Trillion in Off-Balance-Sheet AI Commitments Recently Tanked Vertiv and GE Vernova. Here's What Actually Changed.

Wall Street Journal analysis of Big Tech’s AI disclosures found roughly $3 trillion of off-balance-sheet AI-related liabilities (e.g., data center leases and technology purchase commitments) not reflected on balance sheets. The report initially rattled related AI infrastructure beneficiaries, including GE Vernova, whose orders rose 88% last quarter, and Vertiv, with Q2 sales up 24% YoY, as investors feared potential overspending could be canceled. The article argues the commitments are largely consistent with broader AI infrastructure plans and that McKinsey’s view of enterprises moving toward ROI may reduce the risk that demand evaporates.

Analysis

The market’s first instinct was to treat disclosed AI commitments as a demand cliff, but that’s the wrong mechanism. For hyperscalers, these obligations are better viewed as a stretched call option on future compute demand: they pressure free cash flow timing, not solvency, and they are hard to walk away from without penalties. That means the near-term risk is multiple compression at the spender level, while the suppliers only face real downside if enterprise AI monetization stalls enough to slow new bookings in the next 1-2 quarters.

For GEV and VRT, the selloff looks more like a liquidity/positioning event than a fundamental reset. Their businesses are levered to order visibility and backlog conversion, so the key question is not whether commitments exist, but whether they translate into executable capex and power provisioning over the next 6-18 months. The second-order winner is likely the rest of the AI infrastructure stack — electrical gear, thermal management, gas turbines, and grid components — because any pause in one layer typically shifts spend to another, rather than eliminating it outright.

The contrarian miss is that investors are focusing on the size of the liabilities instead of the counterparty quality. These are largely the best-capitalized firms in the market; if anything, the longer they defend AI share, the more durable the supplier revenue stream becomes. The real falsifier is a drop in hyperscaler capex guidance or a clear slowing in GEV/VRT order growth, not the existence of commitments themselves.

Key Decisions for Investors

• Buy GEV and VRT on further weakness over the next 1-3 weeks; target a 3-6 month hold with ~15-20% upside if backlog/orders remain intact, and cut if quarterly order growth decelerates sharply or management flags cancellations.
• Use XLK or QQQ weakness to hedge a long GEV/VRT basket if you want beta control; this isolates the AI infrastructure trade from broader mega-cap multiple compression.
• Avoid shorting AMZN/MSFT/GOOG/META purely on the size of disclosed commitments; the better short trigger is any evidence of capex-to-revenue efficiency worsening or free cash flow guidance coming down in the next earnings cycle.
• Set a watch item on next hyperscaler earnings: if capex stays elevated but AI revenue monetization improves, add to GEV/VRT; if capex is cut or delayed, de-risk immediately because the market will re-rate the whole AI power chain.