Bloomberg Talks: Rebecca Patterson (Podcast)

Rebecca Patterson warned that excessive economic growth, AI capital expenditure and fiscal stimulus could create macroeconomic imbalances. She highlighted the risk that AI-related capex crowds out demand for US Treasuries, potentially affecting bond-market demand and yields. The article is an interview promotion and provides no specific forecasts, policy actions, or market-moving figures.

Analysis

The actionable issue is not AI demand itself but the financing mix: sustained hyperscaler capex alongside large fiscal issuance can raise the term premium even if the policy-rate path eases. That regime penalizes long-duration equities most exposed to distant cash flows and elevated multiples—particularly unprofitable software, smaller AI infrastructure vendors, and leveraged real estate—while favoring businesses with near-term free-cash-flow conversion and pricing power.

Over the next 1-3 months, the relevant catalyst is whether long-end yields rise independently of front-end expectations, visible in a bear-steepening of the 2s/10s and weaker Treasury auction tails. Such a move would compress AI-related valuation multiples before it necessarily impairs reported revenue; the second-order risk is that higher discount rates increase customers' return hurdles, extending enterprise AI payback requirements and slowing software seat expansion in 2027 budgets.

The consensus may be too focused on whether AI capex is productive and insufficiently focused on crowding-out: Treasury supply can absorb liquidity that would otherwise fund speculative growth and private-credit issuance. Conversely, this is not yet a standalone bearish signal—if productivity gains broaden into earnings and nominal growth remains firm, cash-generative AI leaders can sustain investment without a material multiple reset. Falsify the duration-risk thesis if the 10-year yield declines while long-bond auction demand improves and hyperscaler capex guidance is funded from operating cash flow rather than incremental debt.

Key Decisions for Investors

• Maintain a 1-3 month defensive duration overlay: long TLT puts or a 2s/10s steepener, sized as protection against a 25-40 bp long-end yield backup. Exit if the 10-year yield falls below the pre-auction range and consecutive long-bond auctions show materially stronger bid-to-cover.
• Express relative valuation risk rather than outright AI skepticism: long cash-generative mega-cap platforms (MSFT, GOOGL) versus short a basket of high-multiple, negative-FCF software via IGV or selected names. Target 8-12% relative return over 3-6 months if term premium rises; stop if enterprise AI bookings accelerate while 10-year yields remain contained.
• Avoid adding to highly levered data-center and digital-infrastructure exposures until financing terms are clearer. Watch the spread between investment-grade corporate yields and Treasury yields; widening credit spreads alongside higher Treasury yields would signal that capex financing is becoming a balance-sheet, not merely valuation, constraint.
• No directional trade solely from this interview. Reassess after upcoming Treasury refunding/auction data and hyperscaler earnings guidance; a synchronized increase in 2027 capex plans without corresponding operating-cash-flow growth would upgrade the short-duration-growth thesis.