Treasuries' Worst Quarter Since '94

The 10-year Treasury yield rose 87.1bps during the September quarter—the largest quarterly increase since 1994—and briefly reached 5.34%, its highest level since 2002; the 30-year yield rose to 5.67%, a 24-year high. Strong U.S. growth, persistent energy-driven inflation concerns, mounting fiscal debt and heavy debt issuance are reinforcing a higher-for-longer rate outlook, with markets pricing at least three additional Fed hikes by mid-2027. Unwinding Japanese-yen carry trades may add selling pressure to global government bonds, although tight credit spreads and an S&P 500 near record highs indicate risk assets have not yet fully absorbed the bond-market repricing.

Analysis

The key mispricing is the coexistence of elevated real discount rates with unusually tight credit spreads and equity multiples that still assume frictionless refinancing. If term premium—not merely policy-rate expectations—is driving the move, lower-quality borrowers and long-duration equities should underperform before headline growth data weaken; the transmission typically appears through weaker issuance, reduced buybacks, and downward 2027-28 earnings revisions over the next 1-3 months.

AI infrastructure is not uniformly exposed. CEG’s contracted power scarcity value can reprice faster than its utility-style duration sensitivity, while AMZN, ORCL and AVGO face a more difficult capital-allocation test: incremental data-center returns must clear a materially higher hurdle rate. MU is comparatively better positioned near term because memory pricing and shipment momentum can offset discount-rate pressure, but its multiple remains vulnerable if hyperscaler capex budgets are deferred rather than cancelled.

The contrarian case is that markets are too quick to treat higher yields as an inflation signal rather than a fiscal/liquidity shock. A disorderly Treasury-auction or funding-market episode would widen spreads and damage cyclicals, small caps, REITs and levered digital-asset proxies such as MSTR disproportionately; conversely, a 10-year yield retreat below 4.75% alongside stable spreads would invalidate the near-term de-risking thesis and favor a rapid duration rebound.

Key Decisions for Investors

• Initiate a 1-3 month pair: long CEG / short ORCL in equal dollar amounts. The trade expresses power-scarcity monetization against hyperscaler capital-intensity and financing sensitivity; reassess if CEG’s forward power-price realization weakens or the 10-year yield closes below 4.75%.
• Add a tactical long XLE / short IWM position over the next 4-8 weeks. Higher nominal yields combined with energy-price pressure favor cash-generative energy over refinancing-dependent small caps; target 8-12% relative return, with a stop if Brent falls below $80 and high-yield spreads remain contained.
• Maintain MU as the preferred AI-semiconductor exposure but use a 10-15% trailing stop or funded downside puts through the next earnings cycle. Upside depends on independently verified DRAM/NAND pricing and hyperscaler order trends, not management guidance alone.
• Reduce or hedge MSTR exposure via 3-month puts rather than treating it as a clean Bitcoin proxy. It carries compounded liquidity, premium-to-NAV and convertible-financing risk if real yields remain elevated; cover the hedge if Bitcoin holds above its prior high while MSTR’s NAV premium compresses.
• Set a risk alert at 10-year yields above 5.50% or a 50bp widening in high-yield spreads. Either condition would warrant broad reduction in long-duration software, unprofitable growth and REIT exposure, as equity markets would likely begin pricing the credit transmission currently absent from valuations.