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.