〔Day Digest〕October Pause Priced Yet Hike Odds Sit ~50-50 as Barclays Adds a January Move, AI Capex at 3.6% of GDP Reshapes the Long End - Bessent Sees Post-War Inflation Relief

0. Today's Setup

Markets enter the Asian session with a pause priced for the October Fed meeting, yet economists are splitting closer to 50-50, and Barclays has now added a January 2027 25bp hike to a December 25bp move ^[1][2][3]. Dow Jones futures rise as rate-hike fears ease ahead of big-tech earnings ^[4]. Long-end Treasury yields eased in Asian trade as oil declined, sitting comfortably below recent multidecade highs ^[5], and the curve has steepened visibly ^[6].

1. Policy Narrative

• **[ESCALATED] Fed Chair Kevin Warsh:** a "3 words" framing that the worst is yet to come for Wall Street ^[7]. The September decision restarted hikes after a three-year pause; the dot plot still shows one more move this year, with Warsh citing strong activity, above-target inflation and Middle East tensions, and adding that financial conditions are not restrictive ^[8].
• **[NEW] Former Fed Vice Chair Donald Kohn (Brookings):** Warsh shows less continuity than prior chairs; he expects reform to push the Fed toward better data-driven assessment and communication, and praised the pivot workgroups ^[8].
• **[NEW] Barclays:** Now projects 25bp hikes in both December 2026 and January 2027, versus a prior single December move ^[3].
• **[NEW] Treasury Secretary Scott Bessent:** The US energy crisis persists, but inflation will slow once the war ends ^[3].
• **[ONGOING] October meeting:** consensus is a pause ^[1][2], but a minority sees a hike and some flag back-to-back October and December moves ^[2].

2. Key Data and Market Read

• **[ONGOING] US industrial production** rebounded, but manufacturing remains weak; import-led price increases are still passing through to retail ^[9]. The December path now hinges on whether the consumer-pressure pressure point inflects ^[9].
• **[NEW] JPMorgan CEO Jamie Dimon:** US inflation has persisted for five years and does not appear to be fading ^[10] - a direct contradiction of Bessent's post-war relief thesis ^[3][10].
• **[NEW] Airfare, delivery and trucking** prices jumped last month on war-driven oil, per the same US data round ^[3].
• **[NEW] BoJ Governor Kazuo Ueda:** potential inflation is "very close" to the price target; PM Takaichi denies driving yen weakness or elevated rates ^[3].
• **[ONGOING] ECB:** President Christine Lagarde said the bond selloff is dragging growth and will suppress inflation; Governor Joachim Nagel said the job is not done; Governor Olli Rehn warned of second-round effects from energy ^[3].

3. Curve Mechanics and the AI Anchor

• **[NEW] AI capex, the cause; long-end yields, the effect:** a Brookings-style study frames the current AI build as the largest capital-spending wave in 200+ years of US history, with annual capex/GDP projected at 3.6% before 2032, above the 19th-century railroad era's 2.2% ^[11]. Implication: long-end yields are repricing away from a pure inflation/rate anchor toward an AI-capex expectation, and low US savings amplify the sensitivity ^[11].
• **[NEW] Safe-haven bid returning:** the simultaneous tech selloff and Treasury rebound is read as the bond's defensive function healing, though the durability of the rebound is unproven ^[12][13].
• **[ONGOING] Curve steepening:** the long end has eased while low-rated credit is under pressure - a classic quality bid ^[6][14].
• **[ONGOING] USD at 100:** long positions have rebuilt for four straight weeks and are getting crowded; 101-103 is framed as the high-end battle zone ^[6][15].

4. Contrarian and Tail Risks

• Two clean contradictions run through the packet: Bessent says war-end will cool inflation ^[3], Dimon says five years of inflation is not fading ^[10]; the market prices a pause ^[1][2], Barclays calls two more hikes ^[3]. The falsifiable test is the October FOMC decision and the next consumer-pressure print ^[1][9].
• Source quality control: items ^[1] and ^[2] are aggregator headlines, ^[13] is a headline-only piece, and the 3.6% capex/GDP figure is one research note ^[11] - treat the capex-to-yield channel as a single-source thesis until a second dataset confirms it.