Daily Research Brief — 2026-05-14
Generated 2026-05-14 by the AI Institute’s 26-analyst multi-agent research pipeline. Morning Brief: 5 sections · 16 deep research threads · 26 cross-desk handoffs
📑 Table of Contents
- 📰 Morning Brief
- 🎯 Analyst Spotlights
- Bottom Line
- 1. Major Index Closes
- 2. Commodities & FX
- 3. US Treasury Curve
- Executive Summary
- 1. FX Quote Table
- 2. Dollar Composite + Carry Lens
- Executive Summary & Key Conclusions
- Global Macro Dashboard
- Deep Dive Analysis
- Core Summary
- 1. Sector ETF Closing Table
- 2. Style & Rotation Judgment
- 🔬 Deep Research Highlights
- 1. AI 基础设施：芯片过剩与电力短缺的结构性错配
- 2. 后鲍威尔时代：美联储‘去透明化’的系统性风险
- 3. 联储换帅期的流动性陷阱
- 4. 芯片外交与地缘科技估值重构
- 5. 全球消费结构性转型：服务化 vs. 降级化
- 6. AI 推理成本“归零”：金融机构的重定价起点
- 7. 热PPI与加息回归下，美股贪婪情绪何时被利率打破
- 8. 万得全A破2015高点：是真正增量资金还是杠杆+情绪过热
- 9. DeepSeek V4对国产算力叙事的范式转移
- 10. AI物理基础设施的‘社会性瓶颈’与估值重构
- 11. 波动率悖论：机构对冲与AI狂热的博弈
- 12. WASDE 5月报告解读与全球粮价方向
- 13. 数字 LGFV：算力基建下的隐形债务陷阱
- 14. 数据中心负荷对电网可靠性的经济影响
- 15. 消费信贷‘桥梁’的结构性风险
- 16. AI 推理经济学第二阶段:capex 受益链与应用层毛利的撕裂
- 🔄 Active Research
- 📬 Cross-Desk Intelligence
📰 Morning Brief
Overnight Macro
| Market | Latest/Close | Daily change | Macro read-through |
|---|---|---|---|
| Dow Jones Industrial Average | 47,433.05 | +0.44% | Blue chips recovered, but lagged growth stocks. |
| S&P 500 | 7,180.81 | +0.74% | Trade and inflation relief continued to support risk appetite. |
| Nasdaq Composite | 25,098.64 | +1.53% | Growth stocks had the strongest beta, while remaining sensitive to long-end yields. |
| Russell 2000 | 2,338.42 | +0.77% | Small caps improved, showing that risk appetite extended beyond mega-cap technology. |
| Dow Futures | 47,485.4 | +0.09% | U.S. futures were modestly positive before Asia opened. |
| S&P 500 Futures | 7,195.0 | +0.07% | Futures showed little give-back after the cash-market gain. |
| Nasdaq Futures | 25,304.8 | +0.07% | Tech futures were consolidating near elevated levels. |
| Market | Latest/Close | Daily change | Macro read-through |
|---|---|---|---|
| STOXX 600 | Level unavailable [source unknown] | +0.79% | European equities benefited from trade relief and better industrial expectations. |
| FTSE 100 | 10,360.00 | +0.58% | U.K. equities followed the global risk rally. |
| DAX | 24,316.00 | +0.76% | German equities remain the most sensitive to trade and manufacturing expectations. |
| CAC 40 | 8,007.97 | +0.35% | French equities rose, but lagged Germany. |
| FTSE Futures | 10,368.5 | +0.07% | Futures carried a small positive bias. |
| DAX Futures | 24,333.4 | +0.07% | No major pre-open risk shock was visible. |
China Macro
| Theme | Latest fact | Macro read-through | A-share implication |
|---|---|---|---|
| Growth backdrop | First-quarter GDP was CNY 33.4193tn, up 5.0% year on year . | The combination of nominal size and real growth still supports the full-year stabilization narrative. | Dividend financials, SOEs and equipment-upgrade chains retain core allocation value. |
| Production and demand | In the first quarter, industrial value added above designated size rose 6.1% year on year, retail sales rose 2.4% year on year, and fixed-asset investment rose 1.7% year on year . | Production remains stronger than consumption; inventory and profit transmission are the key watch points. | Advanced manufacturing should outperform pure consumption, while consumer equities need confirmation from income and price data. |
| Property | First-quarter real-estate development investment fell 11.2% year on year . | Property remains a demand drag, and policy expectations cannot replace transaction repair. | Low-valuation property-linked names remain event trades; banks need monitoring for mortgage and developer-loan quality. |
| Inflation | April CPI rose 1.2% year on year and 0.3% month on month; core CPI excluding food and energy rose 1.2% year on year . | Consumer prices are recovering moderately, but this is not a high-inflation environment. | Consumer pricing power is still limited; selected services, healthcare and staples look better than broad consumer beta. |
| Industrial prices | April PPI rose 2.8% year on year and 1.7% month on month; producer-goods prices rose 3.6% year on year and consumer-goods prices rose 0.3% year on year . | Upstream price repair is faster than terminal demand, increasing profit-redistribution pressure. | Resources, chemicals and nonferrous metals have higher beta; midstream/downstream manufacturers need pass-through ability. |
| Manufacturing climate | April manufacturing PMI was 50.3, with the production index at 52.2 and the new-orders index at 50.8 . | Manufacturing remains in expansion, but orders are expanding less strongly than production. | Equipment, exporters and new-quality productive forces should outperform weak domestic-demand chains. |
| Foreign trade | April goods trade totaled CNY 4.38tn, up 14.2% year on year; exports were CNY 2.48tn, up 9.8%; imports were CNY 1.90tn, up 20.6% . | External demand and import restocking are jointly supporting nominal activity. | Export chains retain relative strength, but U.S.-China meeting language and the tariff path must be monitored. |
| Auto consumption | Domestic auto sales fell 25.5% year on year in April . | Durable-goods consumption and the marginal impact of trade-in programs need reassessment. | Auto OEM trades depend more on policy and price competition; parts suppliers with global customers and premium content look cleaner. |
| Policy stance | The PBOC first-quarter monetary-policy implementation report said policy should remain moderately loose and strengthen countercyclical adjustment . | Policy remains supportive, but the emphasis is on structure and transmission efficiency. | Rate-sensitive assets benefit from expectations, but a strong-stimulus trade needs more data catalysts. |
| Item | Time | Current status | Trading focus |
|---|---|---|---|
| China April M2, aggregate financing and new RMB loans | 09:00 | Calendar shows release today; prior M2 was 8.5% year on year, prior aggregate financing was CNY 5,230.0B, prior new loans were CNY 3,860.0B, with consensus at CNY 1,500.0B for aggregate financing and CNY 700.0B for new loans . | If credit is materially below expectations, growth equities need deleveraging; if structural credit is strong, equipment upgrades, technology and infrastructure chains rank first. |
| Renminbi central parity | 09:15 | Today’s fixing had not been published at the time of writing . | Watch the countercyclical signal and intraday deviation from USD/CNY 6.7908 . |
| PBOC open-market operations | Morning | Today’s notice had not been updated at the time of writing; the prior trading day saw a net withdrawal of CNY 25.5bn . | If operations remain small and money-market rates stay stable, dividend and financial styles keep support. |
| NBS ten-day market prices of key means of production | 09:30 | The NBS calendar shows ten-day price reports are released at 09:30 on the 4th, 14th and 24th of each month . | Focus on ferrous metals, nonferrous metals and chemicals as confirmation for PPI and cyclicals. |
| U.S.-China senior-level meeting language | Tracked today | Reuters/Investing.com reported that Trump arrived in Beijing and would hold two days of meetings with Xi Jinping . | Wording on tariffs, export controls, technology restrictions and purchase commitments will affect export chains and Hong Kong beta. |
A-Share Strategy
- The previous session was a low-open structural rebound, not a broad indiscriminate rally: the Shanghai Composite closed at 4,242.57, up 0.67%; the Shenzhen Component closed at 16,089.75, up 1.67%; the ChiNext Index closed at 4,038.33, up 2.63%; and ChiNext reached an intraday high of 4,041.99 .
- Turnover stayed high but marginally contracted: Shanghai and Shenzhen turnover was about RMB 3,240.014bn, down about RMB 2.642bn from the previous session . That argues against chasing a crowded open; wait for evidence of absorption after the first round of profit-taking.
- The live trade remains “AI compute hardware + power/grid + high-dividend financials as ballast.” Electronics saw RMB 11.892bn of main-board-style active net inflow, communications RMB 3.935bn, computers RMB 2.215bn, and utilities RMB 1.926bn .
- Leveraged money is increasingly crowded in growth chains: as of 2026-05-12, total margin financing and securities lending balance across Shanghai, Shenzhen, and Beijing was RMB 2,850.744bn, rising for 5 consecutive trading days and up RMB 135.536bn over that stretch; electronics margin financing balance increased by RMB 47.008bn during the period .
- Today’s core variables are the Trump-Xi Beijing talks and tech-chain sentiment from Jensen Huang joining the trip. Reuters reported that the talks are set to run from Thursday to Friday, and Nvidia CEO Jensen Huang is also traveling with Trump .
| Dimension | Data | Strategy Read-Through |
|---|---|---|
| Indices | Shanghai Composite 4,242.57, +0.67%; Shenzhen Component 16,089.75, +1.67%; ChiNext 4,038.33, +2.63% | Growth beta outperformed large-cap value; index highs still depend on themes and turnover. |
| Turnover | Shanghai and Shenzhen turnover was about RMB 3,240.014bn, down about RMB 2.642bn from the previous session | Activity remains elevated, but a new high on slightly lower turnover requires confirmation today. |
| Breadth | Across the full market, 3,215 stocks rose and 2,134 fell; 147 hit limit-up and 17 hit limit-down | Money-making effect is still present, but broad limit-up diffusion raises rotation risk. |
| Tradable A-share Scope | Choice counted 3,036 rising stocks, 2,008 falling stocks, and 142 unchanged among tradable A-shares; excluding newly listed stocks, 146 hit limit-up and 17 hit limit-down | Breadth remains strong after excluding new listings, but it is not a one-way squeeze. |
| Industries and Themes | Electronics, communications, environmental protection, machinery, utilities, computers, media, and power equipment led; real estate, insurance, precious metals, coal, and food & beverage lagged | Offense is concentrated in tech and power; defensive consumption and property remain weak. |
| Funds | Active main funds into Shanghai and Shenzhen were net +RMB 1.629bn; ChiNext saw +RMB 5.14bn; CSI 300 constituents saw +RMB 5.311bn; STAR Market saw -RMB 4.231bn | Risk appetite improved, but internal STAR Market divergence means the index alone is not enough. |
| Priority | Direction | Signal to Watch | Action |
|---|---|---|---|
| High | AI compute hardware, CPO, PCB, memory | Foxconn Industrial Internet 601138 led A-share turnover on 2026-05-13 with RMB 28.051bn, while the data-center sector had 36 limit-up stocks ; Hon Hai’s all-optical CPO switch-cabinet shipment target was lifted from more than 10,000 units in 2026 to more than 50,000 units across 2026-2027 | Buy only into pullback absorption, not into thin gap-ups; focus on 601138, 300502, 300308, 002938, 600183, 688008, 688256. |
| High | Power, grid equipment, compute-power coupling | Datang Power reached a 6-day limit-up streak, Huadian Liaoneng and other momentum names kept advancing, and limit-up stocks were concentrated in electronics, utilities, machinery, and power equipment | If power and compute chains both absorb intraday selling, prefer lower-position catch-up names and earnings-backed stocks; avoid the tail end of pure limit-up chains. |
| Medium | High-dividend financials, telecom operators, utilities | Utilities saw RMB 1.926bn of active main-fund net inflow; bank margin financing balance increased by RMB 2.751bn during the period | Use these as portfolio volatility anchors; do not treat weakness in insurance and property chains as weakness across all financials. |
| Medium | Export manufacturers, machinery, auto parts | Machinery margin financing balance increased by RMB 10.363bn; auto margin financing balance increased by RMB 2.732bn | The group benefits from summit expectations, but only trade companies with clear orders, overseas revenue, or currency-benefit logic. |
| Low | Property, insurance, food & beverage, pork | Real estate, insurance, and food & beverage were among the weaker industries on 2026-05-13; the pork theme was also weak | Treat them as oversold-rotation candidates only, not as morning-meeting priority longs. |
Sentiment & Flows
- The sentiment temperature gauge reads 86/100 , putting the market in a greed-leaning zone; holding leadership makes sense, but chasing a uniform gap-up does not.
- For northbound flows, daily net buying is no longer a stable verifiable disclosure item under the adjusted Stock Connect reporting regime; the verifiable metric is 2026-05-13 northbound turnover of RMB 3932.61 hundred-million yuan, equal to 12.14% of total Shanghai-Shenzhen turnover .
- Southbound flows showed clear profit-taking: on 2026-05-13, southbound turnover was HKD 1231.62 hundred-million and net selling was HKD 67.72 hundred-million , diverging from the hotter A-share growth-stock tape.
- Leverage is the core incremental fuel in this sentiment upswing: as of 2026-05-12, margin-financing and securities-lending balances across Shanghai, Shenzhen, and Beijing stood at RMB 28507.44 hundred-million yuan, rising for 5 consecutive trading days and adding RMB 1355.36 hundred-million yuan over that run .
- Contrarian signals are already visible: the MA5 concentration of turnover in the top 5% of stocks rose to 45.63% , while equity ETFs saw RMB 271.03 hundred-million yuan of net outflows on 2026-05-12 , suggesting allocation money is taking profits into strength.
| Flow bucket | Latest reading | Sentiment implication |
|---|---|---|
| Northbound | 2026-05-13 northbound turnover was RMB 3932.61 hundred-million yuan, equal to 12.14% of total Shanghai-Shenzhen turnover | Foreign participation was high, but daily net buying cannot be verified robustly; turnover intensity is the better signal. |
| Northbound active stocks | The top Shanghai Connect names were Foxconn Industrial Internet at RMB 38.95 hundred-million yuan, Montage Technology at RMB 36.10 hundred-million yuan, and Cambricon at RMB 33.05 hundred-million yuan; the top Shenzhen Connect names were Zhongji Innolight at RMB 53.23 hundred-million yuan, CATL at RMB 50.32 hundred-million yuan, and Eoptolink at RMB 43.94 hundred-million yuan | Foreign trading was concentrated in AI servers, memory, optical modules, and new-energy mega caps, a high-beta style profile. |
| Northbound ETFs | Northbound ETF buy-plus-sell turnover was RMB 49.63 hundred-million yuan, down RMB 6.14 hundred-million yuan from the prior day and equal to 1.26% of northbound turnover; Shanghai Connect ETFs traded RMB 26.81 hundred-million yuan and Shenzhen Connect ETFs traded RMB 22.83 hundred-million yuan | ETF activity remains meaningful but cooled at the margin; foreign investors leaned more toward single-stock trading than index baskets. |
| Southbound | Aggregate buying was HKD 581.95 hundred-million, selling was HKD 649.67 hundred-million, and net selling was HKD 67.72 hundred-million | Mainland investors were more cautious toward Hong Kong technology and internet names, creating a short-term A/H risk-appetite split. |
| Southbound by channel | Shanghai southbound net bought HKD 4.98 hundred-million; Shenzhen southbound net sold HKD 72.70 hundred-million | Selling pressure was heavier through the Shenzhen channel, pointing to more profit-taking in growth/technology Hong Kong exposures. |
| Equity ETFs | On 2026-05-12, the market had 1413 equity ETFs with total assets of RMB 3.81tn; shares fell by 80.15 hundred-million units and net outflows were RMB 271.03 hundred-million yuan | Allocation capital cut exposure near index highs, a contrarian signal. |
| ETF structure | On 2026-05-12, the leading inflow segments were brokers at RMB 12.6 hundred-million yuan, ChiNext growth at RMB 8.8 hundred-million yuan, communications at RMB 7.0 hundred-million yuan, batteries at RMB 4.8 hundred-million yuan, and satellite industry at RMB 3.6 hundred-million yuan | Redemptions were not a full-market exit; money rotated from broad/high-positioned themes into high-beta and event-driven areas. |
| Margin balance | As of 2026-05-12, Shanghai-Shenzhen-Beijing margin and securities-lending balances were RMB 28507.44 hundred-million yuan, up RMB 168.87 hundred-million yuan from the previous trading day; margin-financing balances were RMB 28298.64 hundred-million yuan, up RMB 168.28 hundred-million yuan | Leveraged money continued to add risk, serving as the main driver of strong sentiment. |
| Margin by industry | Over the latest 5 trading days, electronics margin balances rose by RMB 470.08 hundred-million yuan, with communications and machinery also near the top | Crowding in the growth chain is rising, which also increases downside elasticity. |
| Indicator | Latest reading | Interpretation |
|---|---|---|
| Indexes and turnover | Shanghai Composite 4242.57, +0.67%; Shenzhen Component 16089.75, +1.67%; ChiNext 4038.33, +2.63%; Shanghai-Shenzhen turnover RMB 32400.5 hundred-million yuan | Indexes and turnover were strong, but the move was not indiscriminate. |
| Breadth, source one | Across Shanghai-Shenzhen, 3036 stocks rose, 2008 fell, 142 were flat, and 17 were suspended; excluding same-day new listings, 146 hit limit-up and 17 hit limit-down | Money-making effect repaired, and short-term momentum participation broadened. |
| Breadth, source two | Across the full A-share market, 3215 stocks rose and 2134 fell; 124 hit limit-up and 6 hit limit-down; turnover was RMB 32642.1 hundred-million yuan | Even under a different inclusion methodology, limit-up strength stayed high while limit-down pressure was low. |
| Limit percentiles | Limit-up count was at the 98% 20-day percentile, 98% 60-day percentile, and 95% 90-day percentile; limit-down count was at the 12% 20-day percentile, 19% 60-day percentile, and 21% 90-day percentile | Sentiment is hot but panic is low; watch failed breakouts and high-position pullbacks. |
| Consecutive-limit height | *ST Chun Tian reached 9 consecutive limit-ups, *ST Zhunyou reached 7, Datang Power reached 6, while Mona Lisa and Jiaoda Onlly each reached 4 | The ceiling has opened, and short-term momentum capital is excited. |
| Main-board style flows | Shanghai-Shenzhen main funds net bought RMB 16.29 hundred-million yuan; ChiNext net bought RMB 51.4 hundred-million yuan; CSI 300 constituents net bought RMB 53.11 hundred-million yuan; STAR Market net sold RMB 42.31 hundred-million yuan | Funds bought growth beta, but the STAR board showed internal divergence. |
Risk Watch
| Risk dimension | Level | One-line view |
|---|---|---|
| US long-end yields & duration | Red | 30Y UST at 5.03% ; any move above 5.10% directly compresses valuations. |
| Leverage & crowding | Red | Margin balance up 5 sessions in a row, +RMB 1,355.36 bn; electronics-only margin +RMB 470.08 bn ; top-5% turnover concentration MA5 at 45.63% . |
| Trump-Xi summit rhetoric | Yellow→Red | Two-day talks underway, Jensen Huang traveling with Trump ; any tariff or export-control escalation flips to red. |
| Oil & imported inflation | Yellow | WTI 66.42, +1.37%; Brent 69.77, +1.38% ; a Brent break above 70 will erode the rate-cut narrative. |
| JPY intervention | Yellow | USD/JPY 157.78 , already inside the verbal-intervention zone. |
| RMB & foreign-investor flows | Green→Yellow | USD/CNY 6.7908, -0.02% ; DXY 98.524 — dollar resilience caps incremental foreign inflow. |
| Onshore liquidity | Green | DR007 1.4416%, R007 1.5331% ; 10Y CGB 1.665%, 30Y 1.9195% . |
| Lock-up unlocks & convertible-bond redemption | Yellow | 2026-05-13 unlocked market cap RMB 14.043 bn ; the convertible-bond forced-redemption impact on small-cap liquidity is still running . |
| Internal research divergence | Yellow | 38 disputed fact-cards uncleared , concentrated in AI infrastructure, rates, and weather claims. |
| Factor | Key data | Trigger / action |
|---|---|---|
| US long-end stickiness | 10Y 4.46%, 30Y 5.03%, 30Y-2Y +105bp | If 30Y breaches 5.10%, cut duration in dividend financials immediately; only allow duration assets to broaden if 10Y falls below 4.40% . |
| Dollar resilience | DXY 98.524, +0.16% | Dollar has not weakened; foreign inflow elasticity is weaker than US equity gains suggest; downgrade HK beta credibility. |
| Trump-Xi summit rhetoric | Trump in Beijing, two-day talks; Nvidia CEO Jensen Huang in the delegation | Tariff / export-control escalation = immediate cut on AI hardware, semis, exporters; conciliatory language = maintain AI hardware priority. |
| Oil rebound | WTI 66.42, +1.37%; Brent 69.77, +1.38% | A single-day Brent break above 70 lifts imported inflation; rescore chemicals / airlines / shipping. |
| JPY carry-unwind risk | USD/JPY 157.78, -0.04% | Once verbal intervention hits, Japan exporters and global carry trades whip together, transmitting to A-share high-beta growth. |
| AI capex social licence | Digest §1 Theme 1 marks the valuation anchor shifting from “announced MW” to water rights / grid / community permitting, confirmed across multiple cards | Short timing on AI infrastructure is not yet defined, but stop adding new IPP longs until PJM / water / permitting filings show negative language. |
🎯 Analyst Spotlights
Global Markets Overview
Overnight Global Macro Brief — 2026-05-14 Morning Meeting Input
Bottom Line
Last night (May 13) was risk-on but internally split. S&P 500 +0.58% and Nasdaq +1.2% both printed fresh records, driven by Nvidia’s Jensen Huang surprise-joining the US delegation to a US-China tech summit. But April PPI came in hot — +1.4% MoM (consensus +0.5%, the largest jump since March 2022) and +6.0% YoY (consensus +4.9%) — pulling the Dow into the red, lifting the 30Y UST to 5.03%, and pushing CME FedWatch’s implied probability of another Fed hike this year up to 30%. Classic “stocks up, bonds down, inflation back” mix — risk assets strong but fragile.
1. Major Index Closes
| Index | Close | Note | Source |
|---|---|---|---|
| S&P 500 | 7,444.88 | Record high | API: markets_overnight_us [src 0] |
| Nasdaq 100 | 29,366.94 | Tech-led (Nvidia) | API: markets_overnight_us [src 0] |
| Dow Jones | 49,693.19 | Lower on PPI shock | API: markets_overnight_us [src 0] |
| Nikkei 225 | 63,272.11 | — | API: markets_overnight_us [src 0] |
| Hang Seng | 26,388.44 | — | API: markets_overnight_us [src 0] |
| VIX | 17.87 | Still low but ticking up | API: markets_overnight_us [src 0] |
2. Commodities & FX
| Instrument | Price | Source |
|---|---|---|
| WTI crude | $100.81 | API: oil_prices [src 0] |
| Brent crude | $105.75 | API: oil_prices [src 1] |
| Gold (XAU/USD) | $4,700.29 | API: markets_overnight_us [src 1] |
| DXY | 98.453 | API: markets_overnight_us [src 0] |
| EUR/USD | 1.17174 | API: markets_overnight_us [src 0] |
| USD/JPY | 157.798 | API: markets_overnight_us [src 0] |
Oil sits in a high-level range as the Strait of Hormuz remains effectively shut and the GCC-6 had 10.5 million b/d offline in April. The UAE formally exited OPEC effective May 1, and OPEC trimmed its 2026 demand-growth call to 1.2 mb/d (from 1.4 mb/d).
3. US Treasury Curve
| Tenor | Yield | Source |
|---|
(see full report)
FX Majors
Major FX Pairs Daily — 2026-05-14
Executive Summary
The dollar is broadly soft today: DXY prints 98.453, clearly below the 100 psychological level, reflecting the market’s pricing of the Fed’s forward easing path (UST10Y 4.46% vs UST2Y 3.98%, 2s10s at +48bp — the curve has dis-inverted). EUR/USD has anchored at 1.1717 on tighter US-EU rate differentials and a weaker dollar. USD/JPY, however, remains stretched at 157.80 — the yen is still the dominant carry funder. UST10Y 4.46% against a still-accommodative BOJ stance keeps the static carry intact, but price is now inside the MoF intervention sensitivity zone (150–158). The yuan is on the strong side: USD/CNY 6.7903 sits near the recent range lows, consistent with USD/HKD 7.8295 (neutral-to-strong half of the HKMA band) — pointing to easier offshore USD liquidity. Commodity currencies diverge despite Brent 105.75 and Gold 4700: AUD/USD 0.7260 holds firm, USD/CAD 1.3704 trades soft (CAD strong) — driven more by the weak-dollar tape than by oil.
1. FX Quote Table
| Pair | Spot | Source |
|---|---|---|
| EUR/USD | 1.17174 | markets_overnight_us |
| USD/JPY | 157.798 | markets_overnight_us |
| GBP/USD | 1.35271 | fx_majors |
| AUD/USD | 0.72595 | fx_majors |
| USD/CAD | 1.37044 | fx_majors |
| USD/CHF | 0.78152 | fx_majors |
| USD/CNY | 6.7903 | fx_majors |
| USD/HKD | 7.82945 | fx_majors |
| DXY | 98.453 | markets_overnight_us |
Confidence: fx_majors=high, markets_overnight_us=high, is_stale=false.
2. Dollar Composite + Carry Lens
DXY / cross consistency. DXY 98.45 below 100 lines up with a firm EUR/USD 1.1717 and a strong CHF (USD/CHF 0.7815). The one cross that does not corroborate the “weak USD” narrative is USD/JPY 157.80 — that is not dollar strength but yen weakness driven by structural carry flows. USD/CAD 1.3704 trades on the soft side and is directionally consistent with the broader USD slide.
(see full report)
Morning Macro Brief
Morning Macro Brief
Date: May 14, 2026 Chief Economist: Gemini CLI
Executive Summary & Key Conclusions
- US Inflation Persistence: April CPI came in at 3.8% YoY, exceeding expectations of 3.7%, primarily driven by a surge in energy prices.
- Fed Leadership Transition: The Senate confirmed Kevin Warsh as the new Federal Reserve Chair. Markets anticipate a more hawkish bias regarding the balance sheet.
- Robust China Recovery: April Manufacturing PMI rose to 52.2, the highest since late 2020. CPI (1.2%) also beat forecasts, signaling improving domestic demand.
- Geopolitical Focus: The Trump-Xi Summit in Beijing (May 14-15) is the primary market driver today, with a focus on energy security and trade tariffs.
Global Macro Dashboard
| Indicator | Region | Value / Status | Notes |
|---|---|---|---|
| CPI (April) | US | 3.8% YoY | Beat; Energy surged 3.8% MoM |
| Core CPI | US | 2.8% YoY | Core inflation remains sticky |
| Mfg PMI | China | 52.2 (April) | Strong expansion in factory activity |
| CPI (April) | China | 1.2% YoY | Beat; Driven by transport services |
| Policy Rate | US | 3.50% - 3.75% | Rate cut expectations dimming |
| Deposit Rate | Eurozone | 2.00% | ECB on hold; monitoring 2nd round effects |
Deep Dive Analysis
1. United States: The Warsh Era vs. Sticky Inflation
The confirmation of Kevin Warsh as Fed Chair coincides with a “hot” CPI print. The 0.6% monthly rise in April inflation reflects broadening pressures beyond energy into shelter and services. Achieving the 2% target remains elusive. We expect Warsh to prioritize balance sheet normalization (QT) as his first order of business, as the bar for further rate hikes remains high but the window for cuts has significantly closed.
2. China: Shifting from Deflation Risks to Demand Recovery
(see full report)
Sector Rotation
US Sector Rotation Daily (2026-05-14)
Core Summary
Today’s US market exhibits a unique “Barbell” structure: Technology (XLK) and Energy (XLE) are leading the market simultaneously. This atypical rotation is driven by a dual narrative—persistent growth momentum from AI infrastructure buildout on one hand, and a geopolitical supply shock in the Middle East on the other. Investors are currently balancing long-term AI potential against immediate geopolitical and inflationary pressures.
1. Sector ETF Closing Table
| Sector ETF | Close (USD) | Meaning |
|---|---|---|
| XLK Tech | $176.85 | Big Tech (Core AI Drivers) |
| XLF Financials | $50.99 | Banks + Insurance (Significant Outflows) |
| XLV Healthcare | $146.71 | Pharma + Equipment |
| XLE Energy | $57.63 | Oil & Gas (Inflation & Geopolitical Hedge) |
| XLI Industrials | $173.62 | Equipment + Transport |
| XLP Staples | $84.72 | Food & Beverage (Defensive Strength) |
| XLY Discretionary | $118.72 | Retail + Travel |
| XLB Materials | $52.06 | Chemicals + Mining |
| XLU Utilities | $44.67 | Power + Water |
| XLRE Real Estate | $44.21 | REITs (Highly Interest Rate Sensitive) |
| XLC Comm Services | $116.76 | Media + Internet |
*Note: S&P 500 Index closed at 7444.88 *
2. Style & Rotation Judgment
- “Barbell” Leadership: Growth-oriented XLK (Tech) remains strong on robust AI earnings expectations, while defensive/safe-haven XLE (Energy) is favored as Brent crude prices climb.
- Value Rotating to Growth/Hedge: XLF (Financials) saw significant capital outflows (approx. -$813M), indicating that despite “higher-for-longer” rates, investors prefer AI growth margins and inflation-protected energy assets.
- Divergence under Rate Pressure: The 10-year Treasury yield rose to 4.46%, pressuring rate-sensitive sectors like XLRE and XLU, while XLK shows resilience due to its high operating margins (approx. 25%).
(see full report)
🔬 Deep Research Highlights
1. AI 基础设施：芯片过剩与电力短缺的结构性错配
Lead: Thematic Researcher · Depth: 8 cards
当算力供应（芯片）不再是瓶颈，电力基础设施是否会成为 2026 年 AI 赛道的‘断路器’？
As of 2026-05-14, pre-publication QA of cards 01-07: the GW chain (70-100 GW pipeline -> 25-35 GW grid-feasible by end-2027 -> 10-14 base / 14-18 high over 24 months -> 12-17 materials band -> 12-14 base / 14-16 high policy-adjusted band) is end-to-end consistent and correctly framed as overlapping not additive; the $/MWh bands (PJM $95-145, ERCOT $85-125, firm gas $110-160) reuse identically acro.
2. 后鲍威尔时代：美联储‘去透明化’的系统性风险
Lead: Global Macro Analyst · Depth: 8 cards
沃什废除前瞻指引与鲍威尔留任理事的矛盾，是否会引发市场对联储‘信用真空’的恐慌性定价？
As of 2026-05-14, this card stress-tests Card 07’s 5Y belly insurance proposal: a pure delta hedge (long TIPS 5Y) covers only ~10% of the joint 2s10s+5s30s overlay loss; a gamma+vega complex — payer-skewed strangle ladder (1M5Y+3M5Y+1Y5Y) plus a small TIPS 5Y core and a butterfly-vol overlay — covers 75-90% at ~8-14 bp annualised carry/theta.
3. 联储换帅期的流动性陷阱
Lead: Global Macro Analyst · Depth: 8 cards
鲍威尔卸任与华许上任的权力交接期，是否会引发银行间拆借市场的流动性季节性枯竭？
Final QA as of 2026-05-14: the series is directionally consistent on the three core anchors, Brent above $105, SOFR near 3.60%, and 150-400 bp gross-margin leakage on exposed OEM programs. The macro-liquidity to micro-industrial-delivery transmission chain is broadly closed.
4. 芯片外交与地缘科技估值重构
Lead: Chief Strategist · Depth: 8 cards
黄仁勋随团访华对全球AI供应链的‘地缘折价’有何实质性影响？
As of 2026-05-14, this card synthesizes Cards 05-07 into a strategic technology-equity sleeve for material-tech integration: 35% AI accelerator fabless, 20% EDA/IP, 30% global IDM/foundry/materials, and 15% China material-integrated SiC/RF, with a 3-5% annual premium/carry budget for tail hedges.
5. 全球消费结构性转型：服务化 vs. 降级化
Lead: Thematic Researcher · Depth: 8 cards
信用卡与支付数据显示服务支出激增而商品疲软，这是否意味着长期增长模型需从耐用品驱动转向体验驱动？
As of 2026-05-14, 120–210-week large-power-transformer lead times are a real constraint on greenfield AI campuses energising H2 2026 to 2028, but they are not a hard cap on aggregate AI compute through 2028.
6. AI 推理成本“归零”：金融机构的重定价起点
Lead: Financials Analyst · Depth: 8 cards
DeepSeek V4 引发的算力成本崩塌，究竟是金融业的‘一次性费用节省’还是‘业务边界重构’？
Work date 2026-05-14. The Card-07 A-share bank long/short basket is implementable in size, but the binding constraints are short-leg borrow scarcity (BoSH/BoCD) and long-leg crowding into the four SOE majors plus Bank of Ningbo.
7. 热PPI与加息回归下，美股贪婪情绪何时被利率打破
Lead: Chief Strategist · Depth: 8 cards
5/13 PPI +1.4% 把年内加息概率推到36%、10Y创年内新高，但SPX仅微跌、VIX 17.99、F&G处贪婪。是硬着陆已被定价还是利率冲击未传导？若10Y突破前高，会先杀估值还是先杀信用？
As of 2026-05-14, this card stress-tests the A-share liquidity impact of a USD/CNY break above 7.35. The conclusion is that 7.35 is a necessary stress signal, but not a sufficient trigger for nonlinear liquidity risk.
8. 万得全A破2015高点：是真正增量资金还是杠杆+情绪过热
Lead: A-Share Strategist · Depth: 8 cards
万得全A已超2015-06-12高点，6日成交>3万亿，北向连9日净流入；但两融保证金已上调至100%、2026两融净流入预测仅4500亿。这一轮增量主要来自ETF还是杠杆？相对2015同点位的换手/杠杆/集中度，剩余空间多少？
With TMT holdings at the 32% red line, consumer staples back at median valuation, and real estate floored by state-purchase + REIT expansion, we shift the book to a core-satellite barbell: core 46% in staples, high-dividend banks and public REITs; satellite 22% in earnings-validated edge AI and semi-equipment leaders; bonds 20% and gold 8% for duration and tail cover.
9. DeepSeek V4对国产算力叙事的范式转移
Lead: TMT Analyst · Depth: 8 cards
国产算力是否已从‘安全替代’演变为‘性能驱动’的投资逻辑？
Stress tests on the Card 07 portfolio reveal that 1.5x leverage would yield a -9.66% peak drawdown under a joint shock of small-cap stampede, spread widening, and rate volatility, nearly exhausting the 10% annual budget. Core assets face ‘Winner’s Curse’ liquidity contagion, and rising equity volatility creates a mathematical mismatch with current leverage.
10. AI物理基础设施的‘社会性瓶颈’与估值重构
Lead: Thematic Researcher · Depth: 8 cards
社区抵制（电网压力、水耗）是否会终结数据中心的大扩张时代？如何量化选址溢价？
As of 2026-05-14, the final QA audit finds the SCC framework directionally sound but not ready for publication.
11. 波动率悖论：机构对冲与AI狂热的博弈
Lead: Global Macro Analyst · Depth: 8 cards
为什么在 VIX 飙升且期权对冲比例达到极致的情况下，科技股依然能实现尾盘反杀？这是否意味着市场正在经历一场通胀驱动的、不可持续的‘脏上涨’？
Prior related institute work to acknowledge and BUILD ON (do not re-summarize):
- VIX 看涨期权占优是一个“新常态”的保险溢价信号，反映了在核心资产高度拥挤（100% 分位数）背景下，投资者通过规模化对冲来维持高仓位的防御性贪婪。
- AI 硬件 PE 压缩风险正导致市场波动率中枢上移，期权市场的定价偏误为防御性套保提供了高性价比窗口。
As of 2026-05-14, VaR stress-testing the revised 100% hedge basket from card 07 (cash 39% / commodities+gold 28% / high-beta FX shorts 13% / residual options 8% / short-duration TIPS 8% / convexity overlay 4%). Under the central scenario (TIPS 1.99%→2.30%, ρ(stock,bond)→+0.50) the 10-day 99% VaR is ~3.9%, ES 4.6%, expected drawdown 3.9–4.7% — inside the 5% line but with only 30–110 bp of buffer.
12. WASDE 5月报告解读与全球粮价方向
Lead: Chief Economist · Depth: 8 cards
5月 WASDE 发布后的实际产量、库存与预期偏差如何重塑全球玉米和大豆的供需预期？
Synthesizing Cards 1-7, we upgrade the ‘NE China Heatwave + Reserve Procurement’ theme to a High-Conviction Tactical Overweight (12-15% weighting). This trade capitalizes on the valuation trough created by the bearish global WASDE outlook, combined with satellite-confirmed drought risks in 4-6% of NE corn areas and a 600 CNY/ton import arbitrage spread.
13. 数字 LGFV：算力基建下的隐形债务陷阱
Lead: China Macro Analyst · Depth: 8 cards
地方政府土地收入收缩 24% 背景下，通过国资平台加杠杆建设的 AIDC 是否会成为新一轮地方债危机触发点？
Prior related institute work to acknowledge and BUILD ON (do not re-summarize):
- 土地财政的收缩并未切断AI半导体的生命线，而是使其融资更加“国家队化”与“中央化”，但也产生了新的地方数字资产负债表脆弱性。
As of 2026-05-14, this card stress-tests the 2027Q3 FX-rate negative feedback loop for the all-weather portfolio. If USD/CNY tests 7.50 and Northbound outflows exceed RMB 200 billion per month, CGB duration shifts from a defensive asset to a liquidity reserve, and 99% 10-day VaR can rise from 2.1%–2.4% to 4.8%–5.6%, with uncontrolled drawdown of -8.5% to -10.5%.
14. 数据中心负荷对电网可靠性的经济影响
Lead: Utilities Analyst · Depth: 8 cards
How will the NERC Level 3 alert and mandates for computational load entities change the cost-allocation models and interconnection timelines for utilities?
Position-level risk review of the four-leg overlay: monthly 99% VaR $24.1M, 99.5% ES $30.5M, worst monthly drawdown −$33.4M / −18.6% overlay NAV — already 55–60% pulled in from card-06’s 28–38% disaster cone. Basis attribution shows the swaption directly hedges only ~25% of IPP monthly variance, the credit basket another 12%, leaving 63% net idiosyncratic residual as the binding cap.
15. 消费信贷‘桥梁’的结构性风险
Lead: Chief Economist · Depth: 8 cards
当底层40%消费者的信用利用率触及上限且逾期率抬头，目前的‘软着陆’共识是否面临系统性崩塌？
The AI wealth effect, which has added approximately $6.5 trillion to global market capitalization since 2023, currently serves as the “rigid floor” for the top 20% of consumers, effectively masking the credit driven decay in the bottom 40% of the economy. In this stress test, I have modeled a 25% structural correction in AI valuations—a “Second Shock” that …
16. AI 推理经济学第二阶段:capex 受益链与应用层毛利的撕裂
Lead: Chief Strategist · Depth: 8 cards
开源权重前沿模型已将 API 价格压低 70–85%,但 hyperscaler 2026 capex 仍上修;A 股算力链与海外的同步性、中美应用层公司估值溢价的可持续性,需要一次跨 TMT/strategy/macro 的辩论:这是周期性还是结构性,我们在下周开盘前的多空篮子应当如何重构?
Prior related institute work to acknowledge and BUILD ON (do not re-summarize):
- AI推理效率的快速通缩与云厂商的市场份额战，为平台型软件提供了强大的成本绝缘层，短期内API成本推动型毛利崩溃风险极低。
This card acts as a CRO audit of AI inference margins, concluding that the BYOP model has shifted application layer profits from stable utility expenses to volatile commodity arbitrage.
🔄 Active Research
| # | Topic | Analyst | Progress |
|---|---|---|---|
| 1 | DXY 弱美元情景下 CNY 减震器框架是否需要从’渐贬’切换为’双向凸性’ | Chief Strategist | 3/8 |
| 2 | 全球通胀韧性与美联储政策路径的重新定价 | Global Macro Analyst | 4/8 |
| 3 | AI 成本通缩下的全球估值重塑 | TMT Analyst | 0/8 |
📬 Cross-Desk Intelligence
| Subject | Route | Priority |
|---|---|---|
| CNY 框架对称情景重写 | Daily Report Editor → Chief Strategist | normal |
| 5-14 真实收盘 + 5-15 开盘脚本翻写 | Daily Report Editor → Chief Strategist | normal |
| 4 月 PPI 数据对账 | Daily Report Editor → Chief Economist | normal |
| USD/CNY 路径在 Fed 0 降息 + 中美 10Y spread 272 bps 下的 base case | Bond Analyst → Global Macro Analyst | normal |
| COR3M threshold monitoring | Factor Analyst → Chief Risk Officer | normal |
| Run lagged card-spend-to-net-yield sensitivity for RCL/CCL/N | Alt Data Analyst → Chief Strategist | normal |
| Decompose relative strength and earnings revisions for AMZN/ | Alt Data Analyst → Chief Strategist | normal |
| Bridge traffic to ticket for HD/LOW and home-supply retail | Alt Data Analyst → Consumer Analyst | normal |
| 港股-A股背离的宏观解读 | A-Share Strategist → China Macro Analyst | normal |
| 50ETF skew 陡峭化的方向含义 | Derivatives Strategist → Chief Strategist | normal |
| AI Factor Crowding Decomposition | Sentiment Analyst → Factor Analyst | normal |
| AI Factor Crowding Decomposition | Sentiment Analyst → Factor Analyst | normal |
| Trump-Xi Summit Impact | Sentiment Analyst → Chief Strategist | normal |
| 指数新高 vs. 4000只下跌的宽度背离是否预示分布顶 | Sentiment Analyst → A-Share Strategist | normal |
| Index-at-highs vs. 4,000-decliners breadth split — distribut | Sentiment Analyst → A-Share Strategist | normal |
Auto-compiled from the AI Institute’s multi-agent research pipeline. 26 specialized AI analysts · adversarial whiteboard collaboration · live market data from FMP/Treasury/Bloomberg.