〔Overnight Brief〕Treasury at Least Doubles Long-End Buybacks to $4B; Citi Targets 20Y at 4.9% From ~5.2% as 30Y Nears 20-Year High - Hawkish FOMC Minutes Test the Bid

0. Weekly Arc

The overnight session traded the long end. The US Treasury at least doubled the size of its long-end coupon buyback program from $2B to $4B for 10-30yr maturities, and bonds rallied ^[1][2]. Citi's Jason Williams team now recommends buying 20Y USTs, projecting a yield drop from ~5.2% to 4.9% as a 'yield-top' signal ^[3][4]. Yet the 30Y had touched a near 20-year high the prior session on debt supply, persistent above-target inflation, and AI-driven corporate bond issuance crowding out long-duration buyers ^[5], and the July 28-29 FOMC minutes released overnight lean hawkish even as ING does not expect a hike ^[6][7]. Net: a buyback-driven technical bid against a structural supply premium.

1. Long-End Mechanics: Buybacks, Not QE

• **[NEW] Treasury at least doubles long-end buybacks:** the US Treasury announced an expansion of the buyback program for 10-30yr nominal coupon securities, lifting the size from $2B to $4B ^[1][2]. Mortgage News Daily flags that the program has run for over 2 years and is primarily a liquidity-support tool, not yield control ^[1].
• **[NEW] Citi recommendation (Jason Williams team):** tells clients to buy 20Y USTs, calling the buyback expansion 'a very strong signal' to the market, with target 4.9% from ~5.2% ^[3][4]. If yields do not stabilize, Citi expects Treasury to cut 20Y and/or 30Y auction sizes at the November refunding; the move also implies Treasury is unlikely to expand coupon auction sizes for 7yr+ maturities for the remainder of the Trump administration ^[3][4].
• **[ESCALATED] Supply backdrop:** the 30Y had touched a near 20-year high the prior session on US government spending, long-end debt issuance, and AI investment-driven corporate bond supply pressuring long-duration buyers ^[5].

2. Policy Narrative

• **[NEW] FOMC minutes (July 28-29), hawkish lean:** minutes were released overnight ^[7]; ING reads the tone as hawkish but maintains a no-hike call ^[6].
• **[NEW] Morgan Stanley (Xing Ziqiang, Chief China Economist):** still expects no Fed hike this year, with possible cuts next year if needed; frames expected balance-sheet shrinkage (per Kevin Warsh) as a tool adjustment rather than a liquidity tightening ^[8]. On AI capex, Morgan Stanley flags Nvidia's $500B standalone compute-financing platform with six major financial institutions, citing a 2027 hyperscaler free-cash-flow gap of around -$140B and favoring ABS/CMBS structures collateralized by high-residual compute assets over unsecured IG corporate credit ^[8].
• **[NEW] (single source / unverified) Goldman's space call:** projects $1.8T global space economy by 2035; Elon Musk on X says the figure is 'much bigger' ^[9]. Goldman cites launch-cost collapse from $65,400/kg to LEO in 1981 to current levels as the structural driver ^[9].

3. Single-Stock Calls

• **[NEW] BofA on Adobe (ADBE):** target raised from $190 to $220 ^[10].
• **[NEW] BofA on Nvidia:** characterized as a 'compelling opportunity' at the current deep discount ^[11].
• **[NEW] JPMorgan on Tesla:** Neutral maintained, $445 target vs $339.30 share price ^[12]. Field research at Fremont ties Robotaxi scaling to FSD v15 launch this year, with ~40% of v15 core modules already in the Robotaxi test fleet; Optimus Gen 3 design frozen, supply chain locked, SoP targeted for 2H 2026 entry into 'Optimus Academy' training, with earliest commercial sales 2H 2027 ^[12].

4. China: Reusable Rockets, Humanoids, Housing

• **[NEW] Zhuque-3 reusable rocket recovery:** first successful landing-leg recovery of an orbital-class first stage in China; the developer, Lanjian Aerospace, had its IPO status resumed from 'suspended' to 'inquired' on the SSE in late June, targeting a 7.5B yuan raise (2.77B for reusable-rocket capacity, 4.73B for reusable-rocket technology) ^[13].
• **[NEW] Yushu Technology IPO debut:** opened +629.44% at 1,100 yuan/share, market cap 444.9B yuan, becoming the second A-share stock to break 1,000 yuan on debut; per-lot gain of 474,600 yuan at 500 shares/lot ^[13].
• **[NEW] Housing divergence (July, 70 cities):** Tier-1 new home prices flat MoM (Shanghai +0.2%, Guangzhou +0.1%, Shenzhen +0.2%, Beijing -0.3%); Tier-2 -0.1% MoM; Tier-3 -0.3% MoM; 23 cities saw flat-or-higher new-home prices, up 2 from June ^[14]. January-July retail sales +1.1% YoY, unchanged from 1H ^[14].

5. Tail Risks and Cross-Asset

• **[NEW] US-Iran MOU expired:** the 60-day negotiation window ended August 17 without substantive progress on Hormuz Strait issues; Trump says the US will not seek extension ^[5]. Zhongshan flags rising geopolitical risk premium as a near-term risk to global equities ^[5].
• **[NEW] China crude/refining (July):** crude imports 8.445 million b/d, +18% MoM, YoY contraction narrowed to -24% from -41%; refinery throughput 12.17 million b/d, -19.5% YoY; gasoline-diesel-kerosene production 7.39 million b/d, -18.1% YoY ^[15]. Demand: total oil consumption 7.576 million b/d, +2.1% MoM on summer driving; gasoline 3.52 million b/d, +4.6% MoM ^[15].
• **[ONGOING] Mortgage/lending backdrop:** IMB per-loan production cost remains at $11k; AI-lending quality-control scrutiny flagged as a regulatory risk vector ^[16].

**What would falsify the bid:** a hotter-than-expected print on the next inflation release, a Treasury refunding announcement that does NOT cut 20Y/30Y auction sizes, or a renewed risk-off from a Hormuz Strait incident. The buyback is a liquidity tool, not QE, and the structural supply pressure from AI-related corporate issuance remains the dominant overhang on the long end ^[1][5].