Even ‘Bond King’ Bill Gross warns ‘don’t own bonds’ as long-term debt enters a new ear of volatility

PIMCO cofounder Bill Gross warned that roughly $84 trillion of government, mortgage and corporate debt has made the credit system unbalanced, with US federal debt at 100% of GDP and AI-sector borrowing adding to risks. He advised avoiding bonds other than one-year Treasury bills yielding 4.55% and urged caution on record-high equities, arguing higher yields will compress profit margins and increase Treasury-market volatility. The 10-year Treasury yield has risen more than 100bps since the Iran war began, while hedge funds' Treasury ownership has nearly doubled since 2023 to 8.5%, amplifying liquidity and basis-trade unwind risks.

Analysis

The actionable signal is not simply higher yields, but a more asymmetric Treasury liquidity regime: leveraged relative-value capital can support tight cash/futures spreads in calm markets while becoming a forced seller when repo financing, equity collateral, or volatility shocks tighten simultaneously. That raises the probability of episodic 10-year yield overshoots and wider swap/credit spreads over days to weeks, even if the medium-term macro path ultimately turns disinflationary. Long-duration assets—TLT, rate-sensitive REITs, utilities, and expensive software—therefore carry a liquidity premium risk beyond their conventional duration exposure.

AI capex is the more important second-order issue. If hyperscaler investment remains debt-financed while monetization lags, the market will differentiate cash-rich platforms from levered infrastructure vendors and power/data-center beneficiaries whose valuations already discount sustained spending. Over 6-18 months, a higher real-rate regime compresses the terminal-value component of AI multiples; the falsifier is clear evidence that incremental AI revenue and operating margins are scaling faster than depreciation, interest expense, and power costs.

T and VZ face a double duration problem: large refinancing needs make equity cash flows more rate-sensitive, while satellite and fixed-wireless competition can cap pricing and raise retention spending. Their dividend yields may limit near-term downside only if free-cash-flow coverage remains intact; a renewed rise in long yields or weaker postpaid/ broadband net additions would turn yield support into an equity-value trap. The bearish credit/liquidity thesis is vulnerable to a growth scare that drives a durable flight to Treasuries, rapid Fed easing, or successful Treasury-market liquidity reforms.

This is not yet a broad risk-off call. The article's claims do not establish imminent deleveraging, and elevated yields can remain attractive carry for well-capitalized financials and short-duration credit. The better expression is to own liquidity and convexity rather than mechanically shorting all bonds after a large yield move.

Key Decisions for Investors

• Maintain a duration barbell for the next 1-3 months: long SGOV/SHY against a modest TLT underweight or put spread. Use this as protection against a further term-premium shock; reassess if the 10-year yield declines 40-50bp while credit spreads remain contained, which would improve long-duration entry economics.
• Pair trade over 3-6 months: short T or VZ versus long TMUS, sized beta-neutral. TMUS has relatively greater pricing and execution optionality, while T/VZ remain more exposed to refinancing and competitive capex; cover if T or VZ demonstrate sustained service-revenue acceleration and dividend FCF coverage improves, or if long yields fall materially.
• Reduce exposure to unprofitable AI infrastructure and high-multiple software; retain or favor cash-rich hyperscalers such as MSFT and GOOGL only where AI revenue disclosure supports capex returns. A better catalyst window is the next two earnings cycles, when capex guidance, depreciation, and cloud margin trends can validate or refute the investment cycle.
• Establish an alert rather than a trade on Treasury-market stress: if repo rates spike, MOVE rises sharply, and investment-grade spreads widen concurrently, add TLT downside hedges and cut levered credit. Absent that three-part confirmation, a basis-trade unwind remains a tail risk rather than a base-case catalyst.