〔Day Digest〕AI Shock Sends Nasdaq-100 Down 6.4% and SOX Down 12.2% as 2027 Risks Rise — Yet Morgan Stanley Calls the Correction Near Its End

0. Weekly Arc ^[1][2][3][4]

• **[ESCALATED] AI liquidation:** US technology stocks suffered their worst weekly selloff since the pandemic, with the Nasdaq-100 down 6.4% in Thursday trading and the Philadelphia Semiconductor Index down 12.2% ^[3].
• **[NEW] Positioning reset:** Morgan Stanley’s technology, media and telecom momentum factor fell 70% from its June high, while hedge-fund net leverage dropped to 42% in one week, the eighth percentile since 2010 and the zeroth percentile over the trailing year ^[3].
• **[NEW] Cross-asset response:** The Nasdaq-100 lost $3.2 trillion of market value in the session after the election, the VIX rose, Treasuries advanced, the dollar and gold strengthened, and oil retreated as US-Iran talks showed signs of progress ^[4].
• **[ONGOING] Index damage:** The S&P 500 fell 2.7% last week to 7,376.86 and stood 5.7% below its October 6 record, leaving Michael Wilson’s 7,100-7,300 downside zone in play ^[1].
• **[ONGOING] Evidence quality:** The packet contains research summaries rather than primary market-tape data; its strongest AI claims are second-hand and several entries are visibly truncated, so positioning conclusions should be treated as hypotheses rather than confirmed execution ^{[5][6][2][3][7][8]}.

1. AI Regime and Election Shock ^{[9][10][2][3][4]}

• **[ESCALATED] Final-surge warning:** Goldman Sachs hedge-fund business global head Tony Pasquariello says the equity bull market’s main trend remains intact but is entering its countdown ^[2]. He expects a last S&P 500 advance followed by the largest net-stock-supply shock since the 2000 internet-bubble break, with earnings growth, fund flows and the AI narrative deteriorating together and volatility’s baseline rising ^[2].
• **[NEW] Internal contradiction:** Goldman Sachs strategist Rich Privorotsky’s report was titled “1998 or 2000?”, creating two references to the dot-com peak within one week as the Nasdaq-100 posted a 6.36% one-day decline and hedge-fund annual returns were almost entirely erased ^[2].
• **[ESCALATED] Election channel:** The Morgan Stanley selloff report attributes the immediate trigger to Democratic state wins and calls to pause data-center expansion in Texas, Georgia, Ohio and Michigan, while strong macro data leaves limited scope for Federal Reserve support ^[3].
• **[NEW] Investment path:** Capital Economics chief economic adviser Vicky Redwood expects AI investment to keep growing but slow next year and then fall about 10% in 2028 ^[10]. A few hyperscale cloud providers are on course to spend $800 billion this year, twice last year’s amount, and $1 trillion or more in 2027; she says external-customer AI revenue must increase tenfold by 2030 to justify that path, implying substantial substitution of existing businesses and labor ^[10].
• **[ONGOING] Counterview:** Columnist Tao Dong says the long-term AI trend remains intact but computing expansion may slow and profit realization may arrive later or be smaller, characterizing the selloff as a stock reshuffling rather than the end of an era ^[4].

2. Strategist Crosscurrents ^{[1][9][6][2][7]}

• **[ONGOING] Near-term rebound case:** Morgan Stanley strategist Michael Wilson says the S&P 500 adjustment may be close to ending after approaching the 7,100-7,300 downside target, and recommends adding exposure to high-quality cyclicals whose valuations have fallen while earnings remain strong ^[1]. His team sees limited risk to data-center deployment after the midterm election ^[1].
• **[NEW] Expectations reset:** The UBS CEO describes the AI-sector pullback as a correction of previously excessive expectations, while acknowledging that the election result may slow AI infrastructure investment and delay the industry’s path to monetizing value ^[9]. He advises clients not to overweight short-term political issues in asset allocation ^[9].
• **[NEW] Thin evidence:** Citi downgraded US equities to Neutral in a one-line item that names no strategist, making it too thin to establish a fully specified house view ^[6].
• **[ONGOING] Chip-cycle clock:** Bank of America’s Vivek Arya and team calculate that the SOX index fell 38% from 14,634 to 9,011 over 136 days through November 5, compared with an average 253-day duration and 32% peak-to-trough decline across four previous major semiconductor downturns ^[7].
• **[NEW] 2027 timing:** The Bank of America team believes the indiscriminate semiconductor selloff may finish in early 2027, with the index historically leading actual semiconductor-sales declines by six to nine months and averaging a 42% rebound after earnings expectations were absorbed ^[7]. Conceptually, EPS cuts could begin in the first quarter of 2027, with capital-spending reductions or slowdowns appearing in the second half ^[7].
• **[ONGOING] Read-through:** The packet spans “correction near the end,” an expectations reset, a neutral rating and a final-surge/poor-2027 warning, supporting a live tactical debate rather than a single consensus ^[1][9][6][2].

3. Macro, Rates and FX ^{[4][8][11][12]}

• **[NEW] Bond-market buffer:** The 10-year US Treasury yield fell 8 bp to 4.86% as the same weekly described recovering bonds, a rebound in gold, a firmer dollar and the removal of oil’s geopolitical premium following the election ^[11].
• **[NEW] Payroll consensus conflict:** October payrolls rose by 165,000 and unemployment fell to 4.1%, but the excerpts give conflicting expectations of 65,000 and 70,000; neither excerpt establishes which consensus figure is authoritative ^[11][12].
• **[ONGOING] Wage pressure:** October hourly earnings rose 0.4% month over month and 3.1% year over year, above the respective 0.3% and 2.9% expectations, while the prior monthly reading was 0.1% ^[12].
• **[NEW] Cost signals:** US nonfarm-business labor productivity grew at a revised 3.6% annualized quarter-on-quarter rate, while revised unit labor costs rose 2.4% year over year from 2.6%, indicating easing but still persistent cost pressure ^[11].
• **[ONGOING] Policy interpretation:** The macro weekly argues that employment is no longer the core variable determining Federal Reserve policy, with the focus remaining anchored to medium-term inflation risks ^[12].
• **[ONGOING] Dollar resilience:** A secondary summary of Barclays’ FX research says resilient US data and limited scope for a near-term sharp reduction in AI capital spending support buying dollar dips; it puts the dollar risk premium near 1%, the lowest in more than a year, and says EUR/USD could briefly reach 1.14 on crowded euro shorts but should not sustain that move, while safe-haven demand supports the yen and French fiscal risk weighs on the euro ^[8].
• **[NEW] Commodity split:** The oil decline after the election and progress in US-Iran talks cooled inflation expectations slightly and helped base metals and gold rebound, while ferrous commodities remained in a low range and most agricultural products weakened amid soft fundamentals ^[12].

4. China Allocation and Equity Calls ^{[13][14][15][16][17][18][19][20][21][22][23]}

• **[NEW] A-share rebound:** The ChiNext Index rose 2.14%, the Shenzhen Component Index gained 3.16% and the Shanghai Composite added 1.39% last week; photovoltaic equipment, film and cinema, and precious metals led, while optical chips, copper-clad laminate and oil and gas extraction lagged ^[20].
• **[NEW] Weekly catalysts:** The decision calendar includes China and US October CPI and PPI data, Russia-US-Ukraine trilateral talks in Abu Dhabi, and earnings from SMIC and CoreWeave ^[20].
• **[ONGOING] Structural case:** China Securities Co. sees A-shares entering a structural revaluation phase supported by lower relending costs for buybacks and stake increases and improved capital-market rules; it advises buying semiconductor pullbacks rather than chasing the rally, identifies photovoltaic as a midterm-election beneficiary and treats AI applications as a short-term trade ^[20].
• **[ONGOING] Consumer rotation:** CITIC Securities says food-and-beverage fund holdings and portfolio weights recovered in the third quarter as money returned from technology, with baijiu holdings rising most and dairy gaining the most among mass-market products; it recommends increasing leading baijiu positions near the sector bottom and focusing on competitive improvement, high dividends and high-growth demand ^[19].
• **[NEW] Construction outlook:** China Securities Co.’s 2027 construction strategy highlights policy support, the “six networks” buildout, infrastructure renewal and the shift in new-home sales toward completed units, while continued energy and AI investment creates additional specialist-engineering demand ^[21].
• **[NEW] High-dividend screen:** East Wu Securities finds that 334 of 3,737 stocks with available yields offer more than 4%, or 8.9%; 220 companies with market capitalizations below CNY20 billion represent 65.9% of that high-yield group, with a median yield of 5.05% and median market capitalization of CNY5.86 billion ^[22].
• **[NEW] Yield and valuation:** The above-4% cohort has median price-to-book of 1.46 times and trailing price-to-earnings of 15.8 times, versus 2.88 times and 59.5 times for yields below 1%, but East Wu warns that dividend sustainability still requires separate verification ^[22].
• **[NEW] OLED materials:** Kaiyuan Securities sees OLED organic materials benefiting from display demand and accelerating domestic substitution, citing an estimated CNY1.4124 trillion global display-panel market in 2025 and identifying organic materials as a high-barrier component that directly affects panel efficiency and lifespan ^[23].
• **[NEW] Split target-price moves:** Citi initiated Muyuan Foods Group at Buy with a CNY50.90 target implying 14% upside; Royal Bank of Canada cut Cloudflare’s target from $405 to $350 while retaining Outperform; Barclays raised Tapestry’s target from $185 to $195 with Overweight; Citi raised Coherent’s target from $420 to $440 with Buy; and Wells Fargo raised AppLovin’s target from $325 to $347 with Market Perform ^{[13][14][15][16][17]}.
• **[ESCALATED] Copper caution:** UBS cut Jiangxi Copper’s target from HK$65 to HK$38 while retaining Buy, reduced 2026 and 2027 earnings forecasts by 5% and 15%, and lowered its target copper-business multiple from 12 times to eight times on 2027 estimates because of more conservative sulfuric-acid pricing, contract-processing-fee assumptions and weaker earnings visibility ^[18]. UBS remains constructive on copper because of electrification, grid investment and AI-related power infrastructure, but says Jiangxi Copper’s mining, smelting and hedging mix prevents higher copper prices from converting directly into sustained earnings ^[18].

5. Falsifiers and Next Tests ^{[9][10][3][7][20][11][12]}

• **[NEW] First test:** Hyperscaler and AI-infrastructure earnings, including SMIC and CoreWeave, will test whether the 2027-2028 investment warnings are developing; sustained capital spending and progress toward external AI revenue would challenge the bearish path ^[10][20].
• **[NEW] Second test:** The key policy test is whether state-level demands become actual data-center project delays; the UBS CEO warns that the election result may slow infrastructure and monetization, while the Morgan Stanley report identifies concrete state-level pressure ^[9][3].
• **[ONGOING] Third test:** Inflation and labor data must resolve the two-sided macro picture; the 165,000 payroll gain, 4.1% unemployment rate and firm wages suggest less scope for easing, while the October CPI and PPI releases will test whether cost pressure is easing ^[20][11][12].
• **[NEW] Fourth test:** Bank of America’s early-2027 endpoint and six-to-nine-month lead for semiconductor sales imply that earnings deterioration could still arrive in 2027 even if the stock-market selloff rebounds first ^[7].
• **[ONGOING] Bottom line:** The packet supports a split regime rather than a unified risk-on or risk-off call: Morgan Stanley sees a near-term rebound setup, while Goldman Sachs and Capital Economics warn that AI investment, equity supply and earnings could deteriorate into 2027-2028 ^[1][10][2].
• **[ONGOING] Evidence control:** Citi’s one-line Neutral rating, the unnamed-company research note and truncated secondary summaries should not be merged into a quantified house-consensus statistic ^{[5][6][2][3][7][8]}.