Guggenheim's Schwartz Concerned Private Credit Will Crack

Alan Schwartz said surging U.S. power demand from electrification and AI could lower long-term electricity costs if data centers and hyperscalers are treated as partners rather than adversaries. He also addressed recent concerns in the private credit market at the Milken Institute Global Conference, but no specific figures, policy changes, or transactions were disclosed.

Analysis

The important second-order effect is that hyperscalers are shifting from pure power consumers to quasi-anchoring customers for new generation and grid buildout. That changes the capital-allocation economics for utilities, IPPs, gas turbine suppliers, transmission equipment, and even nuclear/SMR developers: projects that were once merchant-risky can now be financeable on contracted cash flows if AI load forecasts are credible. The market may still be underestimating how much of the next 3-5 years of US power investment will be pre-sold by a handful of balance-sheet-rich buyers.

The near-term winner set is not just power producers; it is the bottlenecks around interconnection, transformers, switchgear, cables, and gas-fired peakers where lead times are already long. If data-center demand persists, the pricing power migrates to these constraint holders, while industrial users and residential customers face a slower but more persistent upward drift in delivered electricity costs. That creates a latent margin headwind for energy-intensive manufacturing and a potential political backlash if utility bills become visibly linked to AI buildouts.

On credit, the key risk is that private credit concerns are not about losses today but about refinancing conditions over the next 12-24 months. If spreads stay tight while underwriting loosens, the first stress will likely show up in sponsor-backed issuers with aggressive add-backs and stretched maturity walls, not in the headline platforms. A sharper-than-expected slowdown in funding growth or a surprise default cycle would quickly change the narrative from 'structural alternative capital' to 'illiquidity premium mispriced.'

The contrarian view is that the AI power story may be partially self-limiting: the economics improve only if utilization stays high enough to justify capex, and any slowdown in model deployment could leave utilities with overbuilt assets and lower regulated returns than the bull case assumes. Conversely, the private credit market may be less fragile than headlines imply because higher coupons and floating-rate structures have already reset borrowers' pain thresholds upward; the more likely issue is dispersion, not systemic collapse.

Key Decisions for Investors

• Long NEE / DUK on a 6-12 month horizon: benefit from regulated capex and incremental load growth; prefer names with transmission and data-center adjacency. Use a 10-15% trailing stop if regulatory pushback slows recovery of invested capital.
• Long ETN or HUBB versus broad industrials for a 3-6 month trade: transformer/switchgear scarcity should keep pricing power elevated as AI grid spend accelerates. Upside is driven by backlog conversion; risk is order normalization in 2026.
• Pair trade: long XLU basket / short energy-intensive industrials proxy (e.g., XLI or selected chemicals/metals exposure) over 6-9 months to express the electricity-cost pass-through divergence. This works if power prices rise faster than industrial pricing power.
• Short lower-quality private credit proxies through BB-rated leveraged loan ETFs or weakeners in sponsor-sensitive names for 6-12 months; focus on issuers with near-term maturities and weak FCF coverage. Risk/reward improves if refinancing windows close and dispersion widens before defaults do.
• Optionality idea: buy 12-18 month call spreads on GE Vernova (GEV) or similar grid/turbine beneficiaries into any AI capex pullback. The trade pays if hyperscaler demand converts into firm generation orders; the main risk is multiple compression if power demand headlines fade.