AI Bosses Urge Caution, Bonds Climb

The 10-year Treasury yield breached 5%, increasing pressure on the Federal Open Market Committee ahead of its meeting and posing a broader risk to rate-sensitive assets. Bank of America shares fell after CEO Brian Moynihan said third-quarter trading revenue would be relatively flat year over year. The program also highlights escalating friction between AI developers, including Anthropic and OpenAI, and the Trump administration over calls for an industrywide AI-development slowdown, alongside renewed Trump-Carney tensions.

Analysis

The actionable signal is the rates/financial-conditions combination rather than the AI-policy rhetoric. A sustained 10-year yield above 5% raises the discount rate applied to long-duration growth and pre-revenue aerospace, while also increasing corporate refinancing friction; this is incrementally negative for EVTL, whose valuation depends heavily on future capital access rather than near-term cash generation. For BAC, higher long rates are not unambiguously bullish: asset yields reprice faster only if deposit costs remain contained, while securities-mark losses, weaker loan demand and rising credit normalization can offset NII benefits.

BAC's near-term earnings risk is a lower-quality revenue mix: muted trading activity leaves less capacity to absorb expense inflation or credit provisioning. The key 1-3 month catalyst is whether management’s next NII and expense outlook shows deposit beta stabilizing; a benign outcome could support a rerating versus money-center peers, but deterioration in commercial real estate losses or card delinquencies would matter more than a modest move in Treasury yields. AI-development restrictions remain primarily a multiple and capex-timing issue for MSFT, GOOGL, AMZN, ORCL and NVDA, not yet an earnings impairment absent concrete deployment, export-control, liability, or power-permitting rules.

Contrarian view: a 5% long yield need not imply persistently tight policy if it reflects term premium rather than inflation expectations. In that case, a rapid yield reversal would disproportionately benefit duration-sensitive software and EVTL, while removing a perceived NII tailwind for banks. The AI-policy dispute is also potentially constructive for scaled incumbents: compliance, safety testing and regulatory engagement costs can become barriers to entry, consolidating enterprise AI demand among hyperscalers rather than slowing industry revenue broadly.

Key Decisions for Investors

• Maintain a cautious tactical underweight in BAC into the next earnings/guidance update; use a 1-3 month BAC put spread rather than an outright short if implied volatility is reasonable. Thesis is strongest if deposit costs rise or credit provisions exceed consensus; exit on a clear NII-guide increase with stable deposit beta and no credit deterioration.
• Avoid adding to EVTL while long rates remain above 5%; treat any rally as an opportunity to reduce exposure unless the company discloses financing that fully funds its certification and commercialization runway. A move in the 10-year yield back below 4.5% without financing progress is a trading bounce, not a structural thesis change.
• For AI exposure, favor a 6-12 month long MSFT or GOOGL versus short a higher-multiple, capex-dependent AI basket proxy such as ARKQ or a selective semiconductor ETF hedge. Scaled cloud platforms can absorb compliance costs and monetize enterprise distribution; invalidate the pair if enforceable rules materially restrict model deployment or hyperscaler capex guidance is cut.
• Set a rates trigger: if the 10-year yield holds above 5% for two weeks and real yields rise concurrently, increase the duration hedge through QQQ puts or a long TLT put position. If yields retrace below 4.5% on easing inflation expectations, cover that hedge and reassess EVTL and long-duration software for a tactical rebound.