How the US Blue Wave Washed Through Global Markets

This Bloomberg Big Take episode description sketches the market's first read on the 2026 US midterm results, in which Democrats gained ground in both the Senate and the House. The framing is a familiar post-election sorting exercise: big tech shares and French bonds are cast as beneficiaries of the outcome, the Federal Reserve as the day's winner, and Elon Musk's companies as losers.

Host Sarah Holder is joined by Bloomberg Opinion executive editor Robert Burgess and markets-and-politics reporter Matthew Griffin to work through who gained and who lost. The item is only an episode description, not a transcript, so it offers a headline-level map of the reaction rather than evidence or argument: no magnitudes, no mechanism linking the election result to French bonds or to the Fed, and no discussion of whether the moves held beyond the day.

Its value is as a same-day sentiment snapshot of how the midterm result was initially priced, not as an analysis of what the result means for policy or markets.

Ramp's Economist Found Cracks In The AI Thesis | The Brainstorm 156

This ARK Brainstorm episode description presents Ramp lead economist Ara Kharazian's research, drawn from 80,000 businesses and over $275 billion in annual spend, arguing that AI spending is extraordinarily concentrated: 1% of businesses drive 80% of it, and those heaviest adopters are now slowing their spend even as token volumes keep climbing. The description also preserves the counterargument made on the show, with host Brett pushing back that three months is not a trend, that prices have halved over the same window, and that the gap between the median firm at $15 per employee per month and the top 1% at $7,000 to $10,000 per month is itself the story.

The debate is framed around revenue quality for the frontier labs, concentration risk among venture-capital-backed buyers, and where AI spend could land five years out. Because only the episode description is available rather than a transcript, the underlying data and the strength of the rebuttal cannot be checked here, but the item still delivers a genuinely contested question: whether slowing spend among the biggest adopters signals saturation, or merely cheaper tokens masking rising usage.

From Conversation to Application: When a Portfolio Workflow Should Become a Tool

This FactSet essay argues that conversational AI in portfolio construction has a natural life cycle: natural language is the right interface while a workflow is still being discovered, but once the mandate, constraints and outputs have settled, repeating the same conversation on every run is wasteful and, more importantly, unreproducible. A sentence is not a durable specification, the author warns, because two phrasings of the same intent can produce slightly different optimization problems.

The mature pattern is therefore conversation that creates the interface: the settled workflow is packaged into a purpose-built panel exposing only the inputs that vary, while a notebook underneath preserves the full specification — universe, risk model, objective, constraints and defaults — in reviewable, version-controlled code. The piece works through concrete designs: different panels for portfolio managers and quantitative researchers, a curated set of pre-approved objectives, and constraint sweeps that turn a single optimization into a visible trade-off frontier, for example evaluating tracking-error ceilings from 2% to 4%.

It is a vendor-adjacent practitioner argument, grounded in FactSet's own workflow product, and it is candid about the governance catch: a panel that quietly behaves differently from one day to the next is worse than no panel at all, so version control, change review and frozen references for published results are non-negotiable.

Analysts Increasing Quarterly EPS Estimates for S&P 500 for 3rd Straight Quarter

FactSet senior earnings analyst John Butters reports an unusual pattern in S&P 500 earnings revisions: despite market concern about higher oil prices, analysts raised fourth-quarter EPS estimates during October, with the bottom-up Q4 estimate increasing 1.4% to $94.79 from $93.49 between September 30 and October 29. That defies the normal seasonal pattern, in which first-month-of-quarter estimates typically fall — by 1.1% on average over the past five years and 1.9% over the past twenty — and marks the third consecutive quarter, and the fifth time in six quarters, that first-month estimates have risen.

Sector breadth, however, was split rather than broad: five of eleven sectors saw increases, led by Energy at +20.3%, while five declined, led by Consumer Staples at -4.0%. Full-year 2026 estimates also rose 0.6% to $363.94 over the same window.

The piece is a clean, data-dense FactSet update rather than a deep interpretive essay; it documents the revision trend carefully but offers no explanation for why estimates are rising against the historical pattern, and no view on whether the trend survives the rest of the quarter.

Responsible AI’s Next Challenge: Preserving Human Judgment

This short MIT Sloan Management Review note closes five years of joint responsible-AI research with Boston Consulting Group by posing the question it says now defines the field: as AI systems act more independently, how do organizations preserve human judgment when and where it matters most? Rather than presenting new empirical findings, it distills the year's work into a progression — first looking beyond model performance to workforce impact, then insisting that human oversight means more than verifying outputs and requires genuine expertise and independent judgment, and finally drawing a deliberate line around agent autonomy, deciding where independent action is appropriate and where human intervention remains essential.

The through-line is institutional rather than technical: responsible AI is achieved by designing organizations, processes and systems that let people and AI work together, not simply by building better models. The note is a framing and sign-off piece, not the full article it points to, so its claims are programmatic — a research agenda for governance, workforce design and autonomy limits — rather than evidence about what any particular system does, and readers looking for case-level detail will need the underlying articles themselves.

Macro Signposts | Campaign Spending: The Hidden Boost Behind U.S. Consumer Resilience

PIMCO's Wilding and Westwood solve the puzzle of 2026 US consumer resilience: it is less a story of household strength than of accounting and one-off fiscal support offsetting a large energy shock. Gasoline and other energy goods prices rose roughly 40% this year, and with energy at about 3% of spending, the mechanical drag on spending elsewhere is about 1.2 percentage points — or about 0.6 points even if households absorb half by saving less — yet real consumption still grew about 2.75% annualized through the first three quarters, close to 2025's pace.

The missing offset, in BEA national accounts data, is political spending: because political organizations sit in the nonprofit sector serving households, their advocacy is recorded as personal consumption, and the election-driven rise in professional advocacy spending added roughly 0.5 points to real consumption growth on average over the first three quarters, nearly cancelling the energy drag. The 2026 midterms were the most expensive on record, with nearly $12 billion spent on federal races alone versus $8.9 billion in 2022, disproportionately funded by wealthy households whose contributions would otherwise largely have been saved — which also helps explain the falling measured saving rate.

A second, separate boost came from the One Big Beautiful Bill Act, whose retroactive household provisions added roughly $160 billion to household cash flow in 2026 through larger refunds and smaller tax bills. Both supports fade in 2027, leaving energy prices as the swing factor: if they retrace toward pre-conflict levels, real incomes recover some lost ground; if not, the energy drag should finally become visible in consumption.

State of the Themes

Citrini Research's tenth State of the Themes declares the broad thematic trade that worked since May 2023 over in its old form: stocks are no longer cheap, sentiment is no longer bleak, and narratives have moved out on the risk curve toward speculative, unprofitable companies priced on 2030 addressable markets. The core argument is that AI infrastructure has hit peak expectations rather than peak adoption — hyperscaler capex is approaching $1 trillion annualized while ecosystem revenue, though at least $200 billion in ARR, cannot grow fast enough to close the gap, quarterly remaining-performance-obligation growth has slowed to its lowest in years, and funding is tightening as banks were left holding $15 billion in bridge loans for a halted Anthropic-backed Texas data-center project.

Token pricing wars, notably OpenAI's July 30 price cut, signal commoditization, and data-center bottlenecks in power, labor and permitting — sharpened by gubernatorial politics in Texas, Ohio and Georgia — are turning from a momentum theme into execution risk. Yet Citrini does not call an AI bust or an economic breakdown: broad job openings are rising again, surveys sit at multi-year highs, and a collapsing breakeven payroll level as net immigration turns negative means even modest payroll gains absorb labor supply, while tax breaks, tariff relief and onshoring persist independently of AI.

The response is a rebuilt Citrindex tilted to reversals, pairs and fresh ideas, with low-beta beginning to outperform after a historically long momentum thrust, plus new buckets in boomer software, alcohol, financials and entertainment.

Asia Pacific Infrastructure: Beyond the Myths

KKR's Insights piece challenges five common misconceptions about investing in Asia Pacific infrastructure, summed up as too fragmented, too risky, too difficult to exit, and already picked over. Against those perceptions, KKR argues from its experience on the ground that the region's complexity can raise barriers to entry but also create opportunities for investors with deep local expertise, and that investment risk depends on more than geography alone.

It also points to an increasingly diverse buyer ecosystem that is expanding potential exit pathways, countering the view that exits are too difficult, and to structural demand for digital infrastructure, energy and transportation that continues to generate opportunities despite the belief that the best chances have already been captured. The piece is framed as an introduction to a longer report and webcast rather than a detailed breakdown of individual markets or assets, so its contribution is mainly a reframing of the region: fragmentation and complexity are presented not only as obstacles but as a source of advantage for locally embedded investors, while the breadth of demand across digital, energy and transport infrastructure is presented as the reason the opportunity set is still reshaping rather than exhausted.

Europe’s strained social model

Bruegel argues that the rise of populist and nationalist forces across Europe, where two-thirds of the EU population will go to the polls through the rest of 2026 and in 2027, is rooted in a systemic crisis of the European social model that for over half a century balanced market dynamics with universal healthcare, state pensions, education, labour-market regulation and collective bargaining. The pillars of that post-war stability have eroded: political fragmentation has pushed the combined vote of the once-dominant democratic coalitions well below 50% in many EU countries, while societies have fractured economically, culturally and generationally.

Three pressures stand out. The demographic squeeze has pushed the total fertility rate below 1.4, against a replacement level of 2.1, while the ratio of old people aged 65 and over to young people aged 15 to 24 has risen from less than one in 1970 to two today, and spending mainly related to the old, pensions and health, has more than doubled to over 20% of GDP, against a little below 5% for education.

Household atomisation has left one-person households at over 35% of all households, with more than 40% of new marriages ending in divorce, shifting care burdens onto underfunded state services. Labour-market insecurity has also grown: over a quarter of women work part-time, temporary contracts affect over 11% of the workforce, union membership has weakened, the labour share of GDP has fallen, and youth unemployment in the EU stands above 15%, against 6% overall, with artificial intelligence likely to intensify skill and reorganisation pressures.

The authors reject both mainstream belt-tightening and nationalist promises to preserve benefits only for native citizens, arguing instead that reform remains primarily a national responsibility, that welfare should be recast as an engine of human-capital formation and economic dynamism rather than merely redistribution, and that mainstream parties must offer a credible vision combining dynamism with tangible social protection, pointing to Nordic experience where high protection coexists with entrepreneurship and technology adoption when embedded in a coherent social compact.

Podcast | China’s critical minerals off switch isn’t as ironclad as Washington thinks

This Trivium China podcast episode, hosted by Andrew Polk with co-host Cory Combs and guest Jessica DiCarlo, an assistant professor of geography at the University of Utah, argues that the idea of Beijing simply flipping an on-off switch on critical minerals processing and shipments underappreciates how complex and fragmented the sector's regulation is in China. The discussion starts from the US-China trade truce extended during Xi Jinping's late-September visit to Washington, noting that more than a month later there was still no Federal Register notice or amended MofCom announcement formalising it, a gap DiCarlo describes as managed interdependence rather than de-escalation.

Within China, she argues, local officials, provinces and firms, not just Beijing, shape what actually gets mined and shipped, and processing, not mining, is China's genuine chokepoint. Cobalt illustrates the interconnection: about 96% of Congolese cobalt goes to China for processing, while China's real edge may be guaranteed buyers up and down the supply chain, and price swings keep sinking Western projects such as Idaho's cobalt mine.

The episode also flags minerals at the margins, including hafnium, zirconium, boron and beryllium, whose strategic importance may grow as definitions of critical minerals keep changing, and argues that a just transition for mineral mining belongs in the security debate rather than on its sidelines.

Global asset managers realign business in Asia

Some foreign asset managers are downsizing or realigning their businesses in Asia, but market participants quoted here argue that this probably does not signal a broader retreat from a region where wealth is rising. The article cites PGIM, which exited Taiwan's onshore market in March last year and India in April 2026, with a spokeswoman saying the moves were intended to sharpen its focus on long-term growth across its core businesses and markets, while insisting PGIM remains committed to serving clients and growing across key markets in the region.

Josephine Kwan, asset and wealth management industry leader at PwC Hong Kong, reads such decisions as isolated changes to operating models rather than evidence of a region-wide pullback, describing Asia Pacific as arguably the most innovative and dynamic of the world's major asset and wealth management markets, supported by rising wealth, deepening capital markets and significant headroom for greater investor participation given demographic trends. Evonne Gan of Broadridge Financial Solutions adds that as Asian markets become larger and more competitive, it is becoming harder for managers to be everything to everyone, so many are concentrating resources where they have genuine differentiation and scale, a sharpening of strategic priorities rather than a withdrawal.

Kwan expects a range of approaches, since the region is not a homogenous market, with mergers, acquisitions and collaborations expanding footprints in some places and individual markets serving as testbeds for innovative products in others.

Five points

Paul Donovan's UBS daily update reads the Democrats' decisive midterm win mainly through its consequences for investors, starting at the state level, where data-centre decisions are often made and where some Democratic governors are publicly opposed to such projects, so builders are likely to have to offer more incentives to local communities. At the federal level, he sets out five points: fiscal gridlock seems more likely, which has supported US government bonds; President Trump's nominees are less likely to be confirmed, including any future vacant Federal Reserve governor positions; administration officials will have to divert time to preparing for congressional hearings; and the Senate shift places Democrats in a strong position in two years' time, when the available seats favour them, which matters for the next administration's legislative agenda.

He also notes that Republican unity bears watching, since Trump's very personal endorsements did not translate into electoral success. Finally, the California ballot initiative on a billionaires' tax was not yet decided, but its heavy campaign spending is itself an economic contribution, and the initiative fits a global trend alongside Italy's recently proposed wealth exit tax, as private wealth reaches record highs while indebted governments look for ways to mobilise it.

Inside the cost-of-living squeeze

Bank of America Institute's short briefing argues the affordability squeeze is narrower than inflation headlines suggest, with housing as the clear exception. Households are spending a slightly smaller share of income on gas and groceries than they did several years ago, despite the cumulative rise in prices, which the authors read as evidence that the squeeze is less broad-based than it appears.

Housing breaks the pattern: both mortgages and rents have risen since 2019, some lower-income renters look particularly stretched, and around one in four higher-income mortgage holders also spend more than half their income on housing. Wage gains and changing behaviour have helped households absorb higher prices, but real wage growth has mostly softened, and the authors suggest weak sentiment may reflect the sacrifices consumers make to keep spending rather than an imminent collapse in spending itself.

That combination, spending holding up while sentiment stays weak, is the briefing's central tension: it suggests households are coping for now, but leaves open whether they can keep absorbing higher prices if real wage growth stays soft or housing costs rise further.

De-Dollarization and the Status of the US Dollar

Robert Hodrick argues that dollar dominance is eroding at the margin in pricing, not collapsing in usage, and that the main threat comes from US fiscal complacency rather than from any rival currency. The dollar's share of reserves fell to 56.8 percent in 2025 from 71.0 percent in 1999, yet dollar reserve assets still rose from $1.35 trillion to $7.44 trillion as total reserves grew to $13.1 trillion, and almost 90 percent of foreign-exchange transactions still involve the dollar.

The telling change is in Treasury pricing: five-year swap spreads averaged 16.02 basis points in 2010-15 but minus 20.24 basis points in 2021-25, and covered-interest-parity deviations for government bonds turned negative, implying Treasuries no longer enjoy the safe-asset discount they once did. No alternative is ready: euro safe assets lack joint liability, renminbi debt is only about 1 percent of international debt, and BRICS shelved a common currency in September 2026, though CIPS, stablecoins and BRICS Pay show demand for workarounds is real.

The Midterm Predictions Were Good

Matt Levine's post-election column is a defence of prediction markets as a consumer product: on election night, Kalshi and Polymarket synthesised scattered results into a single answer hours before television called the races, with Kalshi showing a 98 percent chance of Democratic Senate control by 8:30 p.m. and favoured candidates winning 98 percent of resolved races.

The deeper point is economic rather than celebratory. Election traders in aggregate cannot make money, since the market is zero-sum before fees, so the valuable information they produce is exhaust, subsidised by fun traders, hedgers and sharps, while platforms capture value elsewhere: daily sports gambling and a push into general derivatives such as compute futures.

Levine is candid about the awkwardness of that business model, since sports gambling is addictive in a way biennial election betting is not, and about the unresolved risks of insider trading and manipulation, which he judges too small to swamp elections of this size. The column then pivots to two market-structure curiosities: a criminal quant code-theft trial involving Headlands Technologies, and Republic's planned securities tied to college athletes' future income.

锐评 | 美国中期选举：民主党大胜意味着什么？

中国金融四十人论坛的评论文章认为，民主党基本确定夺回国会两院，其意义更多在于改变华盛顿权力格局，而非直接改变美国经济政策或中美关系。众议院易手符合预期，且已是民主党1990年以来最好的中期选举成绩；更值得关注的是参议院的突破，民主党从共和党手中拿下7席，其中堪萨斯、得州、南卡上一次选出民主党参议员分别在1932年、1988年和1998年，而2024年特朗普在这三州还分别领先哈里斯16.1、13.7和17.9个百分点。1946年至2022年的20次中期选举中，总统所属政党有18次在众议院丢席，但两院同时易手的只有4次，这一次生活成本、伊朗战事和特朗普执政风格的不满叠加，引发了更广泛的选民反弹，约八成登记选民希望下届国会优先降低生活成本，约七成希望优先结束伊朗战事。文章判断，新国会就任后民主党可以加强监督、阻挡人事提名，但特朗普仍握有否决权，民主党短期内很难交出明显成果，大胜也未必转化为2028年大选优势；经济政策博弈将集中在军费和医保，AI立法最可能先从发电、输电审批和数据中心供电成本分摊入手；对华方面，新国会短期重心是国内议程，强硬表态的噪音难改中美关系战略稳定的大局。

Lessons for the renminbi from the yen internationalisation story

Bruegel uses the yen's rise and retreat to argue that China is building a materially larger renminbi role without completing the hardest step of internationalisation: opening the capital account. The yen's reserve share rose from under 1 percent in 1976 to 8.7 percent in 1991, fell to 3.4 percent in 2009 as Japan's economic weight shrank from 18 percent of the world economy in 1994 to 4 percent in 2025 and its banks retrenched, then recovered to 5.8 percent by end-2025 on openness, deep markets and hedgeability.

China has moved faster on use than on holding: renminbi goods-trade settlement rose from 1.7 percent in 2010 to 28.7 percent in 2025, renminbi claims are 31 percent of Chinese banks' cross-border claims, and renminbi SWIFT payments reached 3.3 percent in September 2026, level with the yen's 3.4 percent. But the reserve share slipped back to 2.0 percent in the first quarter of 2026 from 2.9 percent in 2021, foreigners hold only 3 to 4 percent of China's domestic assets against 9 to 10 percent of Japan's, and China remains heavily dollarised on the asset side once state banks are counted.

The lesson is that trade and funding can expand under managed convertibility, while genuine reserve status requires confidence that holders can exit, which Beijing has so far refused to grant.

Daily Comment (November 4, 2026)

Confluence's daily comment ties the midterm sweep to markets, the AI trade and a cluster of geopolitical items, while warning that gridlock and deficit politics will keep volatility alive. Democrats are on track for roughly 225 to 232 House seats and a 54-seat Senate majority, which the authors expect to slow legislative momentum, invite oversight investigations unlikely to end in removals, and force confrontations over data-centre energy demands, tariffs, presidential war powers and a debt ceiling expected to be reached in 2027.

On AI, the tone is two-sided: AMD guided fourth-quarter revenue to about $17 billion against $16.3 billion expected and Super Micro raised its forecast to $19.5 billion, reassuring investors, yet Alphabet, Amazon and Meta sank after earnings on debt and cash-flow fears, Microsoft excepted, and Nvidia's financing of sales to smaller clouds shows how dependent the buildout is on continued borrowing. The comment also notes Trump's plan to end military aid to Israel by end-2026, Nvidia fulfilling just 1 percent of Chinese customers' RTX chip orders and reported White House pressure on Oracle's $7 billion Tencent deal, and a cyberattack on Crédit Mutuel Arkéa days after a breach at Binance costing over $6 billion.

An Active 120th Congress a Neutral-to-Constructive Global Macro Influence

Kim Wallace reads the midterms as a decisive repudiation of President Trump's brand of governance that nevertheless leaves markets in a neutral-to-constructive position rather than a policy vacuum. Senate Democrats will hold a minimum of 54 seats, a majority total last seen a decade ago, and Democrats will hold at least 231 House seats, with margins not seen since 2009, powered by unexpected Democratic strength in the South and by close races breaking more widely than expected.

Wallace draws a historical parallel to 1974, when voters punished a party for tolerating a distrusted Republican president, and argues the breadth of the victory empowers congressional leaders to negotiate from strength with their own conferences, the Republican minority and the president. The governing test, in his view, is the cost of living, which voters have cited since the financial crisis.

On data centres, he argues the election confirms rather than creates a slowdown: Democrats outperformed in Texas statewide and in Virginia's first and second districts, and a capital shakeout plus growing political constraint is guiding both parties toward a 2027-28 rightsizing of national data-centre policy, disappointing investors priced for the old pace but potentially producing a more sustainable rollout.

Election Stunner – A net positive for the Regionals?

Bill Hebel argues the midterm results lean positive for regional banks through rates, regulation, capital and credit rather than through any single policy change. With the 10-year yield off its roughly 5.3 percent high by about 40 basis points, and assuming it stays around 5 percent, banks should reverse a good portion of the negative mark-to-market hit to capital seen in third-quarter results.

On regulation, Hebel pushes back against the reflex that a blue wave means punishment for banks, noting Democrats have recently been more aligned with banks in pushing back on the Clarity Act and seeking reasonable crypto guardrails, and that financial-services hearings carry less political weight nearly 20 years after the crisis than a pivot to AI, data centres and the environment. Capital tailoring under Michelle Bowman, including lifting asset thresholds, is in full flight and likely to be finalised later this year, delivering expected relief on risk-weighted assets, while M&A remains driven by administration appointees rather than Congress.

Any congressional restraint on tariffs or war could help credit by easing corporate margin squeeze and consumer budgets, though Hebel warns Trump could double down. Regionals have already retraced roughly half their August to early-October selloff.

Labor market conditions highlight that less AI impetus would be stabilizing

Gerard MacDonell makes a careful distinction between what would be good for the business cycle and what would be good for equities: a cooling AI buildout would stabilise the economy but disappoint markets priced for relentless AI growth. The labour market looks indistinguishable from full employment, with unemployment near the Fed's natural-rate estimate and prime-age employment near its cycle peak, while underlying inflation in goods and services is running at about 2.5 percent, above the 2 percent target, after stripping up to 30 basis points for tariffs and energy from 3.1 percent core PCE inflation.

With trend productivity near 2 percent and labour-force growth near zero to 0.3 percent, MacDonell puts the speed limit on aggregate demand growth at 2 to 2.25 percent. GDP rose 2.1 percent over the four quarters to the third quarter of 2026, fitting that limit, but real final sales to private domestic purchasers rose 3.3 percent, and 4.8 percent annualised in the middle two quarters, with about three-quarters of a percentage point of impetus coming from the AI buildout.

If AI spending merely stabilised, its growth impetus would fall toward zero and demand would drop to or below the speed limit, easing inflation pressure and plausibly leading to Fed cuts. The worry is financial, not macro: equity prices have implicitly factored in continued aggressive AI growth, and MacDonell's central case is that the impetus diminishes slightly rather than vanishing.

AI and the Midterms – the Clearing Event?

Dauvin Peterson treats the midterms as a clearing event for AI stocks: the shock is over, the chapter is closed, and the market can now work through individual situations rather than pricing a single binary outcome. The central overhang is Texas, where a leadership transition effective in mid-January creates more than two months of uncertainty over next steps, on top of a broader move by several states toward data-centre pauses, including Ohio, Michigan, New Mexico and Georgia.

Peterson's key market call is that compute prices will remain strong, because a more expensive and slower US buildout, set against continued AI demand growth, pushes any supply-and-demand balancing further out; compute futures expected within one to two weeks could create a positive dynamic for CoreWeave, Nebius, Oracle and, debatably, SPCX. The financing side is darker: chipmakers worry about getting deliveries plugged into data centres on time, and powered shells and builders without fully permitted Texas projects will struggle to finance buildouts.

His underperform list is Texas- and grid-delay-focused, while hyperscalers should act defensively and shells with capital or non-Texas pipelines should hold up better.

Little NEW Implications from the Election Given Recent Re-rating

Dennis DeBusschere cautions against extrapolating election-night moves, arguing the midterms are incremental for equities because AI buildout baskets had already rerated on shakier fundamentals and less supportive politics. The election does not change his view that results would not determine the medium- to long-term trend in equities, and a longer buildout cycle need not be all bad if it alleviates fears about whether AI revenue will justify capital spending.

The concrete policy signal is at state level: the six governor races watched most closely all went to candidates leaning toward a pause or moratorium, together representing roughly 58 percent of planned US AI data-centre buildout. Ohio's Acton implies a moratorium, New Mexico and New York elected governors backing a pause, Pennsylvania's Shapiro win means existing regulations will make projects harder and costlier but is net more positive than a freeze, and Texas matters most, where Hinojosa is expected to extend the existing permit pause to extract concessions, putting the timing of 2027 projects at risk.

Beyond AI, the oversold conditions in the average stock and non-AI cyclicals such as retail that helped the recent rally are now gone, leaving 22V more cautious on those names alongside the risk of too-high inflation into year-end.

科技股失宠、非 AI 股翻身? 三大防御配置策略一次看！

美股第三季财报季呈现罕见分化：标普500指数超过七成成分股已公布财报，11个产业板块均实现获利年增速，整体预估获利增速已上修至接近30%，但M7科技巨头的EPS预估增速却由9月底的33.6%下修至26%，为2022年第四季以来首度落后于其他493家公司约33%的预估增速。市场表现同步分化，费半指数自6月底至10月30日下跌25%，10月单月跌15%，纳斯达克100同期跌6%，而道琼10月上涨1.7%、等权重标普上涨4.9%，资金由科技与通讯板块流向其余板块。文章认为，AI变现本身仍在好转，微软云端营收年增17%、Meta广告业务增速26%，本季EPS难看更多源于一次性因素：Meta认列98.9亿美元诉讼和解费用、Alphabet的SpaceX投资收益由第二季约940亿美元转为第三季未实现损失、苹果面临服务获利波动。市场真正的担忧在资本开支与现金流：五大云2026年合计资本开支指引超过8000亿美元，营运现金流预估约7100亿美元不足以支应，第三季自由现金流合计负350亿美元，市场共识预计未来两年资本开支迈向1至1.2万亿美元；2027至2029年折旧摊销分别达3000亿、4000亿、6300亿美元，2029年甚至超过2026年全体营业利益总和，前三季融资已近3000亿美元，11月2日甲骨文发行100亿美元债券和100亿美元新股当日股价跌逾8%，而截至10月底五大云表外隐性承诺已突破4万亿美元。估值随之压缩，亚马逊、Alphabet、甲骨文预估市盈率分别滑落至21.9倍、17.6倍、13.5倍，接近或跌破近十年低点，Meta为19.3倍，只有微软24.9倍相对有韧性。

A Blue Congress, Mr. Veto And A Treasury Twist - Gavekal Research

Gavekal Research argues that despite the Democratic Party's victory in the 2026 US midterm elections, the result may prove less damaging than feared for the US AI capex boom and for US assets more broadly. The reasoning turns on the institutional calendar: Republicans will still have the opportunity to push through legislation in the lame duck session between now and January, and after that Donald Trump will retain his presidential veto over new legislation, which together sharply limit what a Democratic Congress can enact.

The main residual risk sits at the state and local level, where adverse regulatory action remains probable, particularly for the AI buildout. Will Denyer's argument is therefore one of contained damage rather than comfort: federal gridlock and the veto blunt the national policy swing, but investors should watch state and local regulation as the channel through which the election result can still bite, especially for data centers and AI infrastructure.

Only a teaser excerpt was available in Readwise, so this analysis is grounded in that excerpt, but the argument's structure — lame duck, veto, and state-level risk — is a coherent framework for judging how much a midterm wave can actually change the investment landscape.

More Binary Risks In A World Of Binary Choices - Gavekal Research

Gavekal Research reads the 2026 US midterms as a rout for the Republican Party that was visible in advance in the landscape itself: the lack of GOP signs, Trump bumper stickers, and red hats, even in conservative parts of America, augured the beating that followed. The piece draws a sharp irony from the result, arguing that Donald Trump's war to achieve regime change in Iran ended up triggering regime change across the United States instead.

Markets reacted immediately, and the Nasdaq did not like the outcome one bit, a reaction Gavekal links to the fact that Democratic Party candidates almost all run on opposing data centers. The argument connects electoral politics directly to the AI infrastructure trade: if the winning coalition is hostile to data-center expansion, the permitting, power, and local-politics environment for the AI buildout becomes a live investment variable rather than background noise.

Only a teaser excerpt was available in Readwise, so this analysis is grounded in that excerpt, but its central observation — that the political mood visible on the ground anticipated the result, and that anti-data-center politics is now a market factor — is a distinctive lens on an election night otherwise dominated by seat counts.

The midterms are done. Time for the fiscal clashes to begin.

The Atlantic Council's Econographics team argues that the Democrats' victory in the 2026 US midterm elections sets the stage for major showdowns on the federal budget and the debt ceiling, and that fiscal clashes, rather than the election result itself, are the next chapter of the story. The logic is straightforward: a changed Congress confronts a White House of the opposite party just as the calendar forces decisions on funding the government and raising the borrowing limit, and those forced deadlines convert a political realignment into a market-relevant negotiation.

Budget and debt-ceiling fights are familiar territory — they recur because the US system builds in hard deadlines and divided control — but each round carries its own dynamics, depending on the size of the majorities, the state of the economy, and how much brinkmanship each side is willing to tolerate. Only a brief excerpt was available in Readwise, so this analysis stays within the excerpt's central claim rather than attributing further detail, and the low score reflects the thinness of the available material.

Even so, the claim itself is a useful signpost: after the votes are counted, the fiscal calendar becomes the arena where the midterm outcome is priced.

The AI Power Boom: Investing Through Carbon Markets

Artificial intelligence is driving one of the largest increases in electricity demand in decades, and power is becoming one of the defining economic constraints of the AI buildout. The International Energy Agency projects global data center electricity consumption will nearly double, from 485 terawatt-hours in 2025 to 950 terawatt-hours in 2030, with AI-focused demand roughly tripling over the same period.

As data centers multiply, the question for investors shifts from whether AI demand exists to whether power can be supplied, permitted, and priced fast enough to sustain it — and that is where carbon markets enter the frame. The piece, published by KraneShares, positions carbon markets as an investment route through the AI power boom: as electricity demand surges, the cost and allocation of emissions become a direct input into data-center economics, energy sourcing, and the relative attractiveness of clean versus carbon-intensive power.

Only an excerpt of the article was available in Readwise, so this analysis is grounded in that excerpt rather than the full text, but the core claim stands on the IEA numbers alone: AI's power appetite is now large enough to move energy and carbon markets, not just technology stocks.

The AI reset - Fundamentals take centre stage

Eastspring Investments argues that the recent sell-off in global technology stocks reflects greater scrutiny, not a breakdown of the AI growth story, and that the market is moving from a broad-based AI thematic re-rating toward a phase that places greater emphasis on fundamentals, earnings delivery, and returns on AI investment. The latest earnings season supports that reading: AI demand has not weakened and continues to grow rapidly, with Microsoft Azure revenue up 48 percent year-on-year, Amazon AWS up 39 percent, and Alphabet Google Cloud up 103 percent.

But investors are becoming more selective about whether AI investment translates into earnings and cash flow. Microsoft paired Azure's growth with 19 percent overall revenue growth and positive free cash flow, while Alphabet's advertising revenue came in below expectations and free cash flow turned negative; Intel showed the same divergence, with Data Center and AI revenue up 78 percent but Client Computing down 5 percent and Foundry still loss-making.

As AI investment expands from GPUs into memory, advanced packaging, networking, and other infrastructure, the opportunity set is broadening: high-bandwidth memory, advanced packaging, high-end substrates, thermal and power management, PCBs, and copper-clad laminates all benefit from AI server specification upgrades, and Eastspring sees selective opportunities in companies with strong earnings growth, improving fundamentals, and reasonable valuations.

Four key conclusions from the US midterm election

Eastspring Investments frames the 2026 US midterm elections as a broad Democratic victory whose implications for investors extend well beyond Washington. That framing is the substance of what was available in Readwise: the piece positions the election not as a self-contained political event but as a starting point for tracing consequences through markets, policy, and the economy, and it promises four key conclusions for investors drawn from the result.

The emphasis on implications beyond Washington is itself the analytical point — headline election outcomes are a poor guide to portfolio consequences unless they are followed through the channels that actually move asset prices, such as fiscal negotiations, regulatory direction, and shifts in the policy environment facing companies and sectors. Read as a framing device, the piece counsels against reacting to the result in isolation and in favor of asking what a changed Congress can and cannot do, where gridlock constrains action, and where state and local policy may matter more than federal headlines.

Only a brief excerpt was available, so this analysis stays within that framing rather than attributing specific conclusions the excerpt does not contain, and the low score reflects the thinness of the available material rather than a judgment on the full piece.

Ping An bets on private equity, real estate and gold as market risks mount

Chinese insurance giant Ping An Group is increasing allocations to private equity, income-producing real estate, and gold to strengthen portfolio resilience amid market volatility, elevated inflation, and geopolitical uncertainty. Private equity gives the insurer early access to hard-tech and pre-IPO investments, while real estate and gold serve as steady income and geopolitical hedges, and the combination is designed to match long-term liabilities as well as hedge inflation.

Deputy CIO Bill Lu frames alternative assets as playing a growing role in diversification and risk hedging. The firm relies on specialised external managers for niche and cross-border strategies under a strict monitoring framework, while gradually building its own in-house capabilities.

The shift illustrates how large Asian insurers are moving beyond traditional fixed income toward alternatives, not for yield chasing alone but as a structural response to an environment where volatility, inflation, and geopolitical risk have become persistent rather than episodic features of the market.

Apac to lead global VC fundraising growth

The overall outlook for venture capital remains positive, with artificial intelligence expected to support stronger fundraising and improved liquidity, according to Preqin, and Asia-Pacific is positioned to lead global venture fundraising growth. The piece is brief, but its central claim matters for allocators: after a long funding winter, AI has restored a credible growth narrative that lets venture funds raise again and gives existing portfolios a plausible path to exits, and that dynamic is strongest in Asia-Pacific, where AI adoption, deep technology pools, and returning institutional appetite coincide.

For investors, the implication is less about any single fund than about the direction of travel — venture fundraising is recovering, liquidity is improving from a low base, and the region leading that recovery is Asia-Pacific rather than the traditional US and European centres. The thinness of the item itself is a caveat: it reports an outlook, not fund-level data, so it should be read as a sentiment signal about the venture cycle rather than as evidence about any particular market or vintage.

Investors embrace outcome-based ETFs beyond benchmark investing

Asian asset owners are increasingly using active, outcome-based ETFs to generate income, manage downside risk, and improve portfolio flexibility, moving beyond simple benchmark tracking. Insurers, pensions, and family offices are deploying strategies such as covered calls and buffer funds as strategic portfolio building blocks rather than tactical trades, using the products to generate reliable income, navigate market volatility, and improve portfolio resilience amid higher interest rates, geopolitical uncertainty, and persistent market swings.

Adoption is still in its early stages, but regulatory convergence and demand for predictable income are driving a long-term shift toward outcome-oriented ETF solutions. The significance is structural: where ETFs were once primarily cheap beta, Asian institutions now want engineered outcomes — income with defined risk parameters — and product providers that can deliver those outcomes inside a liquid, transparent wrapper stand to capture a growing share of institutional flows, even as the shift raises questions about complexity, fees, and how these products behave in a genuine market dislocation.

Nobody Knows Anything, 2026 Midterms Edition

Barry Ritholtz uses the 2026 US midterm elections as a fresh exhibit for his long-running argument that nobody can reliably forecast politics or markets. Democrats took roughly 250 House seats and about 55 Senate seats, a historic swing that stock futures, prediction markets, and polls all misread in different ways: futures sold off sharply on election night before rebounding, prediction markets underpriced Democratic candidates, and polls captured the direction but badly understated the magnitude, missing the so-called Texas shock in which Democrats won a Senate seat by around ten points and the governor's race by seven percent, while Kansas and South Carolina elected Democratic senators for the first time in nearly a century.

Exit polls showed 43 percent of voters casting ballots to oppose Trump versus 16 percent supporting him, and turnout models missed Democratic motivation. Ritholtz traces the fraying of the 2024 Republican coalition to three signature policies — aggressive ICE immigration enforcement, a muscular foreign policy backing multiple wars, and a haphazard tariff policy — and closes with the broader lesson that tribalism, a poor information diet, and overconfidence in one's own judgment are a dangerous combination when capital is at stake.

Charts of the Week : AI is boosting business creations both in the US and Europe

BNP Paribas Economic Research's Charts of the Week argues that artificial intelligence is now visibly lifting business creation on both sides of the Atlantic. The increase in new businesses since mid-2025 is continuing and significant in the United States, France and Germany, with a smaller rise in the United Kingdom, and it is driven across the board by the information and communication sector, mostly by AI.

The trend is also spreading into other business services, while consumer services, industry and construction show more mixed developments. The most notable claim is about breadth: in both the United States and France, the expansion is broadening beyond its original drivers, suggesting that AI is making it easier to start businesses across all sectors rather than only inside the technology industry.

Read as a diffusion story, the piece implies that AI's economic footprint is widening from a sectoral boom into a general reduction in the cost and friction of founding a firm. The note is, however, only a short chart summary: it gives directions and relative comparisons but no absolute figures, no methodology and no data on whether the new businesses survive, hire or generate revenue, so it supports a directional reading of AI diffusion rather than a quantified growth call.

Is Private Equity Cooked?

In this a16z essay, Jen Kha argues that technology buyouts are in structural, not merely cyclical, trouble, while top-tier venture is better positioned at the start of a new technology wave. Buyout, she contends, is a late-cycle strategy that works once business models are predictable enough to borrow against, whereas venture thrives amid uncertainty; top-quartile venture beat top-quartile buyout through both the internet and mobile/cloud build-outs.

The evidence marshalled is stark: US buyouts returned 14.5% a year over the past decade, trailing the Russell 3000 despite heavy leverage, while US venture returned 17.6% unlevered; PE-backed middle-market companies carry 4.7 to 6.0 times adjusted EBITDA in net debt, exits now take about seven years, and roughly $140 billion of risky PE-backed loans mature in 2028 with another $160 billion in 2029. The 2021-2022 vintages loom largest, with 70% of debt foreclosed in 2025-2026 coming from those deals, Thoma Bravo's Medallia loss held up as proof that even elite operators could not escape the vintage, and self-marked valuations, continuation vehicles and NAV loans accused of flattering reported results.

Kha concedes that capital is sticky, smaller non-software buyouts can still work, and median venture remains riskier, but advises limited partners to interrogate their managers' 2021-2022 payment-in-kind interest and exit routes, and warns that shifting into private credit or secondaries mostly means holding the same buyout companies with less upside.

Investing in Phaze

In this brief a16z investment announcement, Martin Casado explains why the firm is backing Phaze, a high-performance remote desktop company, on the thesis that AI agents become productive only when given full computers, browsers and applications like a human worker. The industry first tried agent-specific APIs and abstraction layers, he writes, but models in practice want to act more like humans, not less, and the most sophisticated users already connect agents to applications and operating systems remotely, even when those apps are operating-system-specific or carry special hardware requirements.

Phaze is presented as the answer: a remote-access product built for interactive, latency-sensitive workloads such as Blender or Unreal Studio, and designed from the ground up for agent compatibility, management and administration. Casado leans heavily on the founding team, noting that Benjy Boxer and Chris Dickson previously built Parsec, the leading remote desktop product for graphics-intensive work, which a16z backed through its acquisition by Unity.

The post discloses no investment terms, customers, pricing or performance benchmarks, so it reads as a statement of conviction about agent infrastructure rather than evidence that Phaze has yet proven its edge.

Scenarios for future banking: implications for profitability, liquidity and solvency

A Bank for International Settlements study asks who will perform banking's core functions a decade from now, and what that shift means for profitability, liquidity and solvency. It classifies some 9,000 banks in more than 150 jurisdictions into three traditional business models using 20 years of bank-level data, then builds four scenarios — slow transformation, technology acceleration, platform expansion and political uncertainty — combining different paths for technology, regulation, consumer preferences and macro-financial conditions.

The framework deliberately structures expert judgment rather than producing forecasts. Traditional banks retain around three quarters of banking-system assets in every scenario, but asset shares understate the change: new providers may capture payments, lending origination, distribution, advice and fee income while banks keep the balance sheet, and technology-driven models could contest roughly 40% of the revenue pool.

Earnings become more volatile as fees replace interest income, digital channels let deposits leave faster, and operational and third-party dependencies grow, adding pressure on mid-sized banks. Four supervisory priorities apply across all scenarios: reassessing deposit outflow assumptions, strengthening operational and third-party resilience, treating weak earnings as a solvency risk, and improving data on emerging models.

WEEKLY WEBCAST: Some AI Lessons In A Booming Economy

In this Yardeni QuickTakes weekly webcast excerpt, Ed Yardeni and Elias examine the lessons from last week's market turbulence, when investors soured on AI-related stocks and pulled the S&P 500 down despite good performance from most sectors. The excerpt argues the week still yielded several insights encouraging for the AI trade, while showing how little tolerance investors now have for under-delivering AI companies — a harshness it blames on analysts' unattainably high estimates, but which at least inhibits bubble formation by punishing disappointment quickly.

Separately, it describes the US economy as firing on all cylinders, citing the release of Q3 real GDP growth and arguing that underlying demand was even stronger than the headline number suggests. The item is a short, truncated teaser for a paywalled webcast rather than the full discussion, so its conclusions are stated without the supporting data or argument that would let readers test them.

Is The AI Selloff A Buying Opportunity? | Warren Pies

In this Forward Guidance episode summary, Warren Pies of 3Fourteen Research explains why he took equity risk off while still believing the AI build-out has room to run, framing the central question as whether the AI trade is busting or merely catching its breath. The discussion flags financials versus staples as flashing caution, treats Federal Reserve rate hikes as the biggest risk to the AI investment boom, and asks who ultimately finances the build-out, including debt issuance and CoreWeave's bonds.

It also turns to energy, arguing that the Iran blockade and blown-out diesel cracks are reshaping the oil landscape, and closes by weighing how fragile the rest of the economy is outside the AI complex. Note that only the episode summary was available, not a full transcript, so the analysis is necessarily thinner than a full-episode treatment.

Blue Wave Is Complicated for AI

Stock futures initially fell overnight as Democrats won control of the House and Senate, then staged a sharp reversal as investors welcomed greater political certainty, producing a midterm relief rally led by the Dow and S&P 500. Solar and other clean-energy stocks were among the big early winners, but the tech sector was a mixed bag: Nasdaq futures lagged because the results ramp up political pressure on the AI data-center buildout, pressuring Tesla, SpaceX, CoreWeave and Oracle, while Lockheed Martin and RTX fell as the power shift and voters' apparent rejection of the Iran war cast doubt over future defense spending. Offsetting that, strong earnings from Advanced Micro Devices and Super Micro Computer lifted part of the tech sector, producing what the piece calls a rare fracture in the AI trade that, if it widens, poses a challenge for markets.

Chartbook 483 After the midterms: How does America come down from Trump's “speedball economy”?

Adam Tooze argues that the Democrats' 2026 midterm landslide was a voter rejection of what he calls Trump's speedball economy: haphazard tariffs, violent ICE deportations, corporate favoritism and corruption, all against a backdrop of cost-of-living anger. AI data centers were directly on the ballot and suffered a devastating reversal, with restriction leading 29 to 1 across 30 ballot measures, a Democrat winning the Texas governorship, and 60 GW of data-center capacity and $200 billion of investment at stake in a state hosting 25% of data centers coming online by 2030.

On fiscal politics, Treasuries rallied on the result, with 10-year yields down four basis points to 4.92%, as markets priced less fiscal impulse from a divided government. Tooze warns, however, that coming down from the speedball may not raise purchasing power: a weaker dollar and an AI slowdown could close the K-shaped economy and help housing and trade-exposed sectors, but devaluation also makes imports more expensive.

He adds that economic nationalism is not just a MAGA hobbyhorse, with deep roots in the labour movement and a hawkish Democratic mainstream on China, and that the new majority inherits a weaponized Strait of Hormuz and a regional quagmire it must find an off-ramp from.

Venture capital backs biotech's clinical push

Citing JP Morgan's Q3 2026 Biopharma Licensing and Venture Report, Preqin reports that global biopharma venture deals totaled $54.5 billion across 959 rounds from 2025 through Q3 2026, including $37.2 billion across 537 rounds in the US and $9.2 billion across 214 rounds in Europe including the UK. So far in 2026 there have been 68 rounds of $100 million or more and 80 rounds in the $50-99 million bracket, helped by an open IPO window, receptiveness to reverse mergers, and more M&A.

Recent large rounds include Oxford-based OMass Therapeutics' £230 million series C, Matchpoint Therapeutics' $150 million series B, Boston-based Soufflé Therapeutics' $135 million series B, and Shanghai-based NeuShen Therapeutics' series B of more than $80 million. The piece also notes that China-based programs represented 51% of big-pharma licensing deals with upfront payments of at least $50 million this year as of September, up from 39% across 2025, and that APAC family offices active in private credit rose 47% between 2024 and 2026.

SpaceX lifts endowments—especially one

The University of North Carolina's exceptional fiscal 2026 returns trace back to a $2 million investment in SpaceX made in 2009 through Founders Fund, when Elon Musk's company was worth about $500 million, according to PitchBook data. SpaceX listed on Nasdaq in June, raising $86.25 billion at a $1.78 trillion valuation; UNC has sold about $1 billion of shares and still holds a stake worth more than $1 billion.

The UNC Investment Fund, which also includes at least $4 billion in non-endowment capital, returned 37.8% in the fiscal year ended June 30, the highest reported so far among the 50 largest US endowments, lifting assets under management from $12.5 billion to around $17 billion. UNC Management Company president and CIO Stefan Strein told trustees that SpaceX accounted for about half of the performance, making the story a striking illustration of how a single early venture bet can dominate an institutional portfolio's year.

Iran: Severe Stress Now, More Risk If Conflict Returns

Torsten Sløk's three charts tell one connected story: since the conflict began, Iranian consumer inflation has surged to roughly 90% year on year, well above the previous peaks of around 50% in 2019-20 and 2022; the open-market rial has fallen to nearly 3 million per US dollar after a near-vertical slide; and crude exports, Iran's main source of foreign currency, have collapsed from about 1.8 million barrels a day to essentially zero. The mechanism is a feedback loop: lost oil revenue means fewer dollars entering the country, which pushes the rial lower, which raises import prices and feeds inflation, and the loop continues for as long as exports remain shut in.

AI and the great cashflow transfer. The mining boom that drove the Australian economy. Psychopatriarchy & the caes…

This Chartbook instalment is a collage rather than a single argument, and its most useful frame is the first: the AI boom can be read as a huge cash flow transfer from the hyperscalers to the manufacturers of chips. Tooze pairs that with Australia's mining boom of the 2000s and into the 2010s, presented as a huge driver of national growth driven by an export surge and commodity prices, inviting a comparison with today's chip-driven wave. The issue then quotes at length a Guardian essay on psychopatriarchy, described as jet fuel to racism, colonialism, imperialism and capitalism, which is polemic rather than evidence.

Apple cracks the door open to Siri AI

Apple has released a developer preview of its rebuilt Siri AI, letting apps of any kind build their own actions into the assistant and teach it multi-step jobs through new skills, a rare crack in its walled garden. The catch is timing: developers can only test with internal users for now, App Store submissions start in spring, and Apple's rules for developers will come later, so iPhone users will feel nothing yet.

The same betas bring China its first Apple Intelligence tools, minus Siri AI. The newsletter's judgement is that the opening matters, but leaves the work to developers while Siri still feels generations behind frontier labs.

AI on the Ballot (Stratechery Article 11-4-2026)

Ben Thompson argues that the 2026 US midterms, in which Democrats swept Congress and Gina Hinojosa won the Texas governorship pledging a moratorium on data-centre construction in the state holding over 25% of planned US capacity by 2030, mark a shift in what constrains AI: from compute, to memory, to power, and now to politics. Unlike globalisation, where affected towns had no say, building data centres requires permission, so voters who felt voiceless before are now blocking projects, with nearly every relevant local ballot measure opposing data centres winning except in North Dakota. Winners are hyperscalers with permitted sites; losers are neoclouds, Nvidia short term, and startups.

Chinese FDI hotspots in Europe face political shifts

Political change in Europe's top destinations for Chinese greenfield investment threatens the relationships that secured multibillion-dollar factory and infrastructure pledges: Hungary's Viktor Orbán and Serbia's Aleksandar Vučić lost elections in April and October respectively, and Spain's Pedro Sánchez, who also courted Chinese manufacturers, has called snap elections for 29 November. Together the three countries attracted $41.4bn of Chinese greenfield FDI pledges between 2020 and September 2026, 43% of Europe's total. Hungary is reviewing CATL and BYD projects, and Serbia's new government may scrutinise Chinese-backed projects, though existing ties give Chinese firms reasons to stay.

What the 2026 US midterms mean for APAC markets

Invesco's message to APAC investors is that the US midterm result changes less than it feels like it does: early results pointed to a strong Democratic showing and Democratic control of Congress, ending two years of unified Republican government, but midterm losses for the president's party, averaging 22 House seats over six decades, are historically normal. Gridlock could reduce the prospect of major additional fiscal stimulus, easing pressure on US growth, inflation, Treasury yields and the dollar, easing pressure on Asian currencies. But the administration retains executive authority, so tariffs, export controls and technology restrictions remain the key risks for China, Taiwan and Korea.

Concrete’s enabling role in building a sustainable and resilient future

Oxford Economics' research for the Global Cement and Concrete Association argues that five megatrends, population growth and urbanisation, infrastructure and housing deficits, climate transition and resilience, geopolitical uncertainty and digitalisation, are reshaping construction demand, with construction activity forecast to increase by more than 70% to US$16.6 trillion annually by 2050. Cement and concrete are presented as critical to delivering essential infrastructure and strengthening resilience. The hard part is meeting rising demand while reducing carbon emissions, improving productivity, managing costs and ensuring sustainability, through innovation and resilient supply chains.

Is AI making diversification harder?

Aberdeen argues that portfolios which look diversified may not be: global equities, emerging markets and small caps are increasingly driven by the same AI theme, with correlations to an AI ETF proxy of 0.79 for global equities, 0.69 for emerging markets and 0.75 for small caps. Technology's weight in the MSCI ACWI Index has risen from around 10% to 31% , contributing roughly half of overall index risk, so owning more equity markets no longer guarantees different sources of return. Cash, bonds and property show lower AI correlations but sacrifice return, hence the pitch for listed infrastructure serving data-centre demand as AI-adjacent growth through different drivers.

Global equities were little changed on Tuesday as Americans voted

Emirates NBD's morning note captures a market caught between two stories: on Tuesday, global equities were little changed, with the S&P 500 up 0.4% on a chip rebound and the Nasdaq 100 up 0.7%, while emerging markets fell 1.7% as the selloff hit Korea and Taiwan, each down 3.6%. Overnight the election result arrived, with Democrats taking the House at about 248 seats to 187 on the NBC estimate and closing in on the Senate, sending US futures, the dollar and yields lower. The note reads the outcome as gridlock plus investigations, with AI oversight and the defence budget most exposed, against the post-midterm history cited by the note.

只给 AI 看两段英文 NetLogo 视频教程，它就复刻出中文版，那还要老师做什么？

王老师的秋季研究生复杂系统仿真课要讲 NetLogo，但旧英文视频语速快、操作有错、版本也不一致，学生跟练困难。他让 Claude Code 先看两段圣塔菲研究所 Melanie Mitchell 入门课视频，再在 NetLogo 7 里真机复现、录屏、剪辑并用中文讲解。首版成片较短，改成跟练版后 Ant1 为 41 步、13 分 39 秒，Ant2 为 45 步、17 分钟出头，每讲完一步画面停住并打出【跟练时间】。随后他又修正了窗口切换后首次点击只激活窗口、没有真正点中 setup、ticks 停在 1229 的问题，换成真正生效的录屏并保留 ticks 归零画面；再按实际操作量而非原视频节奏缩短停留，并删去原视频对比、只按眼前结果讲。结论是 AI 已跨过看懂与复现的一关，老师不可替代的，是替学生看片、把跟不上的地方变成具体要求。

Why have long-term bond yields risen sharply? (Chinese Only)

近期环球长债息急升，美国 30 年期国债孳息一度升至约 5.33 厘、创 2007 年以来新高，10 年期亦接近 4.75 厘，欧洲、日本及英国长债息同样升至多年甚至数十年高位。作者归因于三股力量交织：美元投资级债券供应大增，以科技巨头为代表的超大规模云端服务商为兴建 AI 数据中心、电力设施及相关基建而发债，年初至今发行量已接近 2200 亿美元、远高于 2025 年全年的 1080 亿美元；孳息曲线陡峭化，2 年期受压回落、10 年期与 30 年期显著上升，反映市场重新为沃什领导下联储局的政策取态定价，联储局严格坚守 2% 通胀目标、减少前瞻指引，使长端要求更高的风险补偿；中东局势未明、布兰特原油重上每桶 90 美元以上，推高通胀预期。作者认为这并非政府债务危机信号，更多反映发债量上升、曲线重塑及通胀风险，市场对美债仍有承接、只是要求更高补偿；加上通胀逐步回落至目标、沃什更重视的 12 个月通胀率已降至 2.2%、美国财政部加大回购有望舒缓压力，债息急升或只是暂时性，但短期仍会波动。

Why are we optimistic about mainland China’s AI sector? (Chinese Only)

在全球 AI 竞赛中，美国在 OpenAI、Anthropic 等科技巨头带领下保持领先，但内地 AI 正沿着独特的发展路径加快发展，优势主要来自成本结构、政策环境与开放技术策略的协同。首先，内地电力、数据中心建设及相关配套成本相对较低，使大模型训练与推理的词元成本更具竞争力，有望建立长期的成本护城河，而美国较高的电力及数据中心营运成本往往推高模型服务定价门槛。其次，中央将 AI 列为国家战略重点，从顶层设计到地方落实持续释放政策红利，加快 AI 由实验室走向实际应用场景，与美国较依赖市场竞争的路径形成对比。第三，DeepSeek、智谱、月之暗面旗下的 Kimi 等领先模型往往在推出新一代版本后同步公开相关研究成果，加快先进技术在实验室之间的传播与重复应用，而美国领先模型较多呈现各自为营的闭环特征。加上算力芯片、框架软件及数据治理等关键环节国产替代步伐加快，内地 AI 有望由跟跑迈向并跑、甚至在部分范畴担任领跑角色，作者提醒可留意投资于内地 AI 企业的 ETF，但亦需要注意相关投资风险。

The three key events shaping the near-term outlook for US equities (Chinese Only)

踏入 10 月下旬，美股正迎来第三季业绩期、联储局议息及 11 月美国中期选举三大事件，短期市场情绪、资金流向、估值及风险偏好均可能受到影响。业绩方面，根据 FactSet 截至 10 月 23 日的数据，市场预期标普 500 指数第三季盈利按年增长约 29.5%，连续三季录得逾 25% 增幅、亦是连续第八季录得双位数增长，11 个行业盈利及收入均预计增长，其中以能源及科技业、尤其半导体表现最为突出；市场焦点是大型科技企业及云端服务商能否以盈利和现金流支持庞大的 AI 数据中心投资，若现金流受压或资本开支放缓，相关企业估值或需调整。议息方面，市场普遍预期联储局维持联邦基金利率于 3.75 厘至 4 厘，由于会议不会公布经济预测及点阵图，焦点落在主席记者会以及政策声明对通胀、就业和利率路径的表述，市场认为 12 月仍可能加息一次，但未必代表新一轮紧缩周期开始，利率路径的清晰度较单次议息决定更重要。选举方面，预测市场及民调显示民主党很可能重夺众议院控制权、部分预测机率逾 90%，参议院选情胶着，分裂政府令重大立法更困难、短期有利政策延续及市场稳定，但民主党若扩大优势，市场或关注 AI 监管、税务与能源政策；历史上中期选举后 12 个月标普 500 大多录得正回报。作者提醒，选举未必决定后市方向，但若结果大幅偏离预期或引发债务上限及政府运作争议，短期波动或加剧，投资者宜结合业绩及联储局信号作出判断，而非只看单一事件。

云启参投「主动科技」，推动脑机接口走向临床应用｜云启伙伴

主动科技由临港实验室孵化，把柔性电极、专用芯片、编解码算法到系统集成做成全链路自研，围绕运动、视觉与情绪三条产品线推进，核心产品「天枢」面向中风、脊髓损伤、渐冻症导致的运动与语言功能障碍。公司成立不到一年，已完成天使轮与天使+轮两轮融资、累计约 8 亿元，云启-交大 AI 天使基金参与投资。近日，上海主动科技有限公司宣布完成 4.64 亿元人民币天使+轮融资，由高榕创投、国投先导、上海科创投联合领投，云启资本、康君资本、中网投、国泰海通、浦东创投、上海科创基金、广药资本、闵金投跟投，老股东继续追加投资；据公开披露的融资信息，本轮规模创下国内脑机接口领域天使+轮融资纪录，天使轮与天使+轮合计金额亦创下该领域天使轮系列累计融资纪录。今年 8 月 3 日，公司公布 3.3 亿元天使轮融资，不到三个月，天使+轮再次获得新老股东支持。三条产品线共用同一套电极、芯片与算法底座，运动线上跑通的工程经验可以迁移到视觉和情绪两条线，这是公司选择做平台而不是做单品的理由。运动脑机接口走得最快，2025 年 11 月，核心团队研发的 1024 通道侵入式脑机接口系统取得型式检验报告，为后续临床研究与产品转化打下基础。本轮资金将用于持续推进侵入式脑机接口研发、临床转化与产业化建设，并探索脑机接口与人工智能、具身智能的结合。

一个小人物在数据江湖

这是一篇以化名「小武」为主角的非虚构叙事，讲述一个 AI 应用创业者在三年创业无果、积蓄赔光后转做数据倒爷的经历。过去一年，他在海外地下小网站爬数据、去过流水线般的千人标注工厂、在硅谷追踪过 Surge 的专家网络，见识到游戏玩家操作数据、大学实验室的实验过程数据、以及破产公司整套运营数据都可能被买卖：他为一家 SaaS 公司找到国内模厂买单，以 300 万美金成交，清单几乎完整记录了该公司十多年从获客、签约、交付、研发、售后到内部管理的各个环节。更高效的路径是 API 中转站：国内大厂通过中转站调用海外模型 API，中转站把沉淀下来的任务轨迹抽走转卖，形成用卖数据收入补贴 token、吸引更多调用、再产出更多数据的闭环，直到一家大厂买数据时买到了自己使用中转站的轨迹，才引发降温。作者的核心观察是，任何人、任何行业、发生的任何事都可能成为数据，简单数据在跌价、高质量专家数据反而在涨，普通人参与更难；灰色交易席卷广泛，源于模厂对模型能力的不惜一切追求与每个个体的无力感——大家心知肚明 AI 是通过淘汰人来进化的，却没有人愿意停下，都想抢泰坦尼克号的救生船；上个月成立的国家数据集团，则可能加快这个行业的转折。

Normal Americans punished MAGA

Noah Smith argues that the 2026 US midterms were a referendum on Donald Trump, and that MAGA's crushing defeat exposed a central delusion: the belief in a hidden reservoir of shy Trump supporters who would rescue the party, as in 2016. Instead, polls, forecasts and prediction markets had called a Democratic victory for months, and the election-night surprise favoured Democrats, with even Trump's traditional base of white voters without a college degree abandoning the GOP in droves.

Kansas, South Carolina and Iowa turned blue, and Texas went blue too, with Democratic challenger Gina Hinojosa unseating Republican governor Greg Abbott and progressives winning statewide office. Smith traces the backlash to the cost of living: Trump was elected to tame inflation but raised taxes on almost everything Americans buy from overseas and raised gas prices by attacking Iran, plunged the US into an expensive Middle East quagmire, promoted Vladimir Putin's interests, and presided over corruption that voters now rank above grocery and gas prices as their top concern.

His broader thesis is that normal, swing voters do exist in great numbers, and that pushing them hard enough with corruption, mismanagement and disdain for their daily struggles will eventually make them rise up and punish the incumbent. The warning runs both ways: Democrats, he says, risk a red wave in 2030 or 2034 if they misread the mandate and chase activist priorities such as Palestine over the cost of living, embark on a borrowing binge, try price controls, push billionaire taxes voters just rejected in California, or reopen large-scale illegal asylum admissions.

The AI-enabled risk function of the future

McKinsey argues that after two decades of building risk governance, data and controls, banks face still-mounting demands, and AI now offers a redesign of the risk function, not just automation of single tasks. It describes five shifts — risk work becoming predictive, dynamic, faster and more consistent, with controls embedded in processes and reusable platforms — plus a three-pillar model of a strategic horizontal function, lean AI-enabled domains and a risk-AI centre of competence. Claims include 60 to 70 percent of manual tasks eliminated and 30 to 50 percent productivity gains in early wholesale lending; five roadblocks are flagged, and delay is judged costlier than leadership.

Asset allocation decisions - November Barometer 2026

Pictet's Strategy Unit has turned more cautious for November, cutting equities from overweight to neutral as tighter financial conditions and geopolitical uncertainty worsen the risk-return trade-off. Within equities it tilts defensive, upgrading healthcare to overweight while reducing financials, and upgrades the Swiss franc to overweight as a hedge against further political upheaval. The implication is a defensive, hedge-aware allocation rather than an outright risk-off call.

AI's rising bill of expectations; India blog; the Netherlands in focus; Income restored, and more

BlackRock's weekly highlights argue that AI spending is booming but the bar for returns is rising, so it stays selective while maintaining its core AI conviction. It also points to a deeper shift in India, where headline stock lag may obscure deepening domestic capital markets, to capacity constraints in the Netherlands despite strategic strengths in security, energy and technology, and to richer income opportunities at higher yields that still require flexibility on bond risk. The piece is broad rather than deep on any single topic.

Fitch on China: AI Emerging as Broader China Macro-Credit Theme

Fitch argues AI is increasingly becoming part of China's macro-credit story, expanding from semiconductors and computing infrastructure into automation, power generation, enterprise software and industrial applications, as China seeks new growth drivers beyond property, investment and credit expansion. The same newsletter also upgrades China's bank operating environment to bbb/stable, with six state banks upgraded on 27 October 2026, while warning that local and regional governments' land revenue is likely to stay soft under tighter land and development rules, with the transition extending beyond 2026.