US AI spending peak may arrive this year, Jefferies warns By Investing.com - ca.investing.com

Jefferies warns 2026 could be the peak year for the AI capex boom, citing rising scrutiny of AI capex returns and increased financing via debt/private credit that heightens risk to private credit and private equity; last quarter's US software sell-off illustrates the pressure. The bank also flags the Iran conflict and a potentially closed Strait of Hormuz as stagflationary risks and notes rising Treasury volatility that could strain equities. Morgan Stanley says Asian power prices are up 10%–100% since Feb 2026, Asia ex-China relies on imported gas for ~15% of power, coal generation is rising, and energy storage plus renewables are 20%–25% more competitive vs LNG, favoring large power companies, coal producers and grid operators.

Analysis

The important inflection is not whether AI spending slows — it’s that the marginal dollar of AI capex is now financed increasingly with credit, which converts what was a technology ROI problem into a credit-cycle problem. When yield-sensitive capex projects fail to clear hurdle rates, losses concentrate in privately originated, covenant-lite structures where repricing and liquidity are slow; a 5–7% realized default rate in these pools can produce 15–25% NAV markdowns because recovery values and leverage are both compressed.

Energy-price arbitrage in Asia is creating asymmetric winners over two horizons. In the next 6–18 months, thermal fuel recipients (coal miners, merchant generators, regulated grids) capture immediate spread gains and free cash flow; over 2–5 years, storage-plus-renewables manufacturers and installers win as levelized cost parity with LNG becomes structural, compressing merchant gas economics and shifting capex away from liquefaction and shipping.

Rising Treasury volatility is a stealth amplifier: it increases funding costs for credit funds, spikes margin calls for levered private-market acquirers, and raises the option value of staying in cash for corporates — which can pause capex and accelerate impairment cycles. If volatility stays elevated for a quarter or more, expect forced deleveraging in levered credit pools and a re-pricing of long-duration equity multiples, particularly among names funding growth via debt.

The near-term trade framework should therefore hunt convexity: go where cash flows are improving now (merchant energy, regulated utilities) while buying asymmetric downside protection on long-duration, credit-funded growth exposures and selective credit managers with high private-credit AUM and mark-to-model risk.

Key Decisions for Investors

• Long BTU (Peabody Energy) 6–12 months: buy shares with a 20–40% upside target if thermal coal prices remain elevated; set a trailing stop at -15%. Rationale: direct exposure to Asian thermal demand and immediate FCF re-rating; tail risk is rapid policy-driven coal demand destruction.
• Long SO (Southern Co) 12–24 months: accumulate for 12–18% total return thesis (dividend + modest multiple expansion) as regulated cash flows re-rate versus levered private credit managers. Hedge with 1–2% notional buy-write or short-dated call to improve carry against rate volatility.
• Pair trade (3–6 months): short IGV (software ETF) / long XLE (energy ETF) equal notional — target 2:1 upside vs downside. This captures rotation out of long-duration, credit-funded software risk into commodity-backed cash flows; close if 10y Treasury yield falls >50bp from entry.
• Buy protective put spreads on select high-valuation software names (example: buy 3-month 5–7% OTM put spreads on CRM) sizing to cover 30–50% of notional long exposure to growth names. Cost-limited hedge that pays if credit-funded growth starts to re-price; acceptable premium ~0.8–1.5% of notional.