〔Overnight Brief〕Goldman Raises Brent Forecast to $105 as All-A Ex-Financial Q3 Profit Grows 20.94%, but an 8.54% Drop and 49.3% PMI Keep the Reboot Conditional

0. Weekly Arc: Strong Earnings Meet Thin Liquidity and a Two-Sided Oil Call ^{[1][2][3][4][5][6][7]}

Source note: the packet mixes detailed brokerage reports with newswire summaries, headline-only items and truncated teasers, so the shorter items should carry less evidentiary weight than figures disclosed in full reports ^{[8][9][10][11][12][13][7]}.

The central arc is a disagreement between improving earnings and stressed prices: Chinese bank and new-economy results strengthened, Goldman Sachs split on the AI return cycle, and A-share strategists argued that an oversold rebound is possible but not yet fully confirmed ^[1][2][4][6].

Goldman Sachs lifted its December 2026 Brent forecast to $105 a barrel from $90, but a separate weekly strategy report saw room for a high-level oil price to converge lower as geopolitical premia unwind while OPEC+ supply flexibility, released U.S. supply and weak global manufacturing demand constrain prices ^[3][5].

Morgan Stanley's Mike Wilson described several possible stock-market clearing events that could push prices lower temporarily and create buying opportunities regardless of who wins the U.S. midterms ^[14].

1. China Earnings and Corporate Themes: Banks Improve as New-Economy Profit Broadens ^{[1][15][2][16][17][18][13]}

• **[NEW] Zhongtai Securities — banks:** First-three-quarter bank revenue rose 8.1% year on year and net profit rose 3.8%, with both improving at the margin as net interest margins stabilized and other non-interest income strengthened; asset quality remained stable overall, but dispersion widened as some lenders increased nonperforming-loan disposal ^[1]. The house favored large banks with a funding-cost advantage and selected regional city and rural commercial banks, while seeing medium-term value anchored to more certain ROE, stable dividends and attractive yields despite a short-term technology seesaw ^[1].
• **[NEW] Guotai Haitong — brokers:** Q3 proprietary trading was the central drag as volatility in primary and secondary equities widened the performance gap between leading and smaller brokers; Guotai Haitong favors leading brokers with international and wealth-management strengths because of their superior ROE, growth and valuation appeal ^[15].
• **[NEW] AI and new energy:** The completed A-share Q3 reporting season showed earnings focus shifting toward new tracks, with chips, storage, optical modules and printed circuit boards highlighted as especially strong, while power batteries, energy storage and lithium materials showed structural recovery ^[2].
• **[NEW] Growth-board split:** Wind All A excluding financials and petrochemicals posted Q3 cumulative attributable profit growth of 16.08%, the innovation-and-venture boards delivered 83.66%, main-board growth slowed from 13.89% to 2.86%, and electronics reached 183.92% ^[18].
• **[NEW] Beijing Stock Exchange:** Revenue across all listed companies reached RMB223.711 billion in the first three quarters of 2026, up 15.30% year on year, while attributable net profit reached RMB16.880 billion, up 2.27%; information technology and chemical new materials led five major industries with revenue growth of 21.67% and 22.76% and were the only two with positive profit growth, at 44.61% and 7.02% ^[17].
• **[NEW] ESG governance:** The China Securities Index's new ESG V4.0 methodology placed sustainability-related governance in a core evaluation dimension, and from October multiple A-share companies updated their ESG rules or expanded Strategy Committees into Strategy and ESG Committees ^[16].
• **[NEW] Innovation drugs — headline-level source:** Eli Lilly beat expectations and raised full-year revenue guidance for a third time in 2026, while domestic innovative-drug companies entered earnings validation and commentary pointed to the spread of the dividend-style rotation ^[13].

2. AI Crosscurrents: $1.14 Trillion of Backlog Meets a 2-3x Storage Valuation ^[4][19]

• **[ONGOING] Debate shift:** The central market question has moved from whether AI demand exists to how AI investment returns will be realized ^[4].
• **[ESCALATED] Goldman split — Tony Pasquariello:** Technology giants' Q3 results made a peak in S&P 500 earnings growth plausible, with market momentum weakening; Pasquariello argued that cloud strength can mask weakness elsewhere, larger capital expenditure has not demonstrated additional earnings room, and forward EPS revisions have slowed ^[4].
• **[ESCALATED] Goldman split — Mark Wilson:** Google Cloud revenue rose 103% year on year and its backlog increased 256% to $561 billion, while AWS revenue rose 39% and backlog increased 189% to $578 billion; the resulting $1.14 trillion total was cited as clear evidence that AI returns are being realized ^[4].
• **[NEW] Nomura storage thesis:** Samsung Electronics and SK Hynix were valued at only 2-3x expected 2027 P/E, an 85-90% discount to technology giants, and Nomura said supply shortages should persist through 2028 as long-term supply agreements increase, making storage a relative safe haven whose shareholder-return cycle is only beginning ^[19].
• **[NEW] Cash strain:** The same report showed AI investment weighing on cash generation at large technology companies: Meta's quarterly free cash flow turned negative for the first time since its IPO at -$2.1 billion, Alphabet posted a second consecutive negative quarter at -$9.1 billion, and Amazon's rolling 12-month free-cash-flow gap widened to -$42.4 billion ^[19].

3. China Tape: An 8.54% October Drop Creates a Reboot Test, Not a Confirmation ^{[20][9][10][6][21][22][7]}

• **[NEW] Drawdown:** The CSI 300 fell 8.54% in October, the worst October performance since 2011; the innovation-and-venture board complex lost more than 10% for the week, trading heat stayed low, all three major U.S. indexes closed lower, and international gold and oil each fell about 3% ^[7].
• **[ONGOING] Leverage and style:** A separate monthly strategy report put the CSI TMT index down 13.8% year to date as of October 30 and said October financing purchases averaged 7.93% of A-share turnover, below the 8.50% historical average since 2019 ^[9].
• **[ONGOING] Earnings versus price:** Reported all-A ex-financial single-quarter attributable profit rose 20.94% year on year but fell 1.41% quarter on quarter, with Q3 year-on-year growth versus Q2 turning negative; historically, the CSI 300 rose in 12 of 15 years after Q3 reporting for an average 2.29% and median 1.45%, although the incremental earnings information typically decayed within 1-4 weeks ^[10].
• **[NEW] Reboot case:** One strategy note says the adjustment has been sufficiently deep and funding conditions for an oversold repair are maturing, while CICC domestic-strategy chief analyst Li Qiusuo says valuations are at historical mid-low levels and policy must quickly stabilize growth, the market and expectations to close the gap with fundamentals ^[20][22].
• **[ONGOING] Falsification:** Another bottom-signal report says short-term price sentiment and volume may already be at historical extremes and an oversold rebound has a foundation, but leverage, institutional positions and sector crowding need a longer clearing period; it keeps a barbell in dividends, free cash flow, small caps and themes including physical AI, commercial space and Middle East friction ^[6].
• **[NEW] CITIC Securities:** Four allocation frameworks were proposed: stable returns around high dividends or industries with supply-clearing potential, expectation-driven valuation repair in universally bearish and lightly held sectors, non-AI exporters and resources conditional on falling overseas long rates, and fast-growing new-demand sectors once institutional and active-money positioning clears ^[21].
• **[NEW] Citi:** Sungrow was downgraded to Neutral with a RMB70 target price ^[23].

4. Macro and Next Tests: Oil, Rates and Policy Set the Reboot Clock ^{[3][5][12][18][7]}

• **[ESCALATED] Oil remains the macro fork:** Goldman Sachs raised its December 2026 Brent forecast to $105 from $90, while the opposing weekly view says the prior rise was geopolitical rather than a sustained supply squeeze, sees room for high-level convergence and argues lower oil would ease energy CPI, PPI pass-through and Federal Reserve anti-inflation pressure without independently resetting medium-term core inflation ^[3][5].
• **[ONGOING] U.S. duration:** U.S. long-term capital inflows may no longer keep pace with financing needs, foreign buyers have shifted from official Treasury purchases to private stocks, and reduced appetite for leverage leaves long-end yields difficult to push lower; the stress signal is higher volatility, rapid leveraged exits and rising short-term funding rates occurring together ^[12].
• **[NEW] China policy:** By October 28, all RMB800 billion of new policy financial instruments had been deployed, new special-bond issuance for project construction was accelerating and RMB550 billion of local-government debt capacity remained available, including RMB50 billion more for major economic provinces than a year earlier; the strategy view expects Q4 GDP growth to continue rebounding ^[18].
• **[NEW] Macro test:** October official manufacturing PMI fell to 49.3% from 50.1% and missed the 50.0% expectation, while September industrial profits rose 14.4% and first-three-quarter profits rose 15.5%, with electronics making a significant contribution ^[7].
• **[ONGOING] Event calendar:** The Federal Reserve held rates after its October meeting, the strategy calendar highlights the U.S. midterms and October nonfarm payrolls, and the report notes that President Donald Trump might resume strikes on Iran after the election as major technology shares remain volatile following results ^[7].
• **[ONGOING] Falsification test:** The A-share rebound case weakens if manufacturing stays below 50%, leverage and sector crowding fail to clear, or renewed Middle East conflict keeps the oil and inflation premium elevated; the immediate tests are the U.S. midterms, October nonfarm payrolls, evidence of AI return realization and Chinese policy execution ^[5][6][7].