Welcome to this week’s edition of Geopolitics & the Day After. Each week, we curate and synthesize key developments from global politics, economics, and financial markets, drawing from a wide range of trusted sources. Our goal is to provide you with a clear, concise, and insightful overview of the forces transforming the world today and shaping tomorrow. Below is an overview of what we cover this week:
Geopolitical Concerns shows how China, Russia, and Germany’s leaders are each turning legal codes, energy leverage, or election calendars into tools for managing the domestic costs of prolonged conflict, using courts, oil deals, and parliamentary votes to do what battlefield outcomes alone cannot.
Geoeconomics highlights how a war-driven surge in bond yields and a $2 trillion global interest bill are colliding with fraying trade imbalances and a retreat from dollar hedging, leaving governments with shrinking room to manage debt, deficits, and currency risk at once.
Global Junctions demonstrates that the same financial architecture propping up Nvidia’s valuation, and the AI bet it is meant to sustain, could just as easily accelerate its own unwind, while China’s export-driven manufacturing model spreads its own overcapacity problem worldwide.
Global Trajectories reveals that the institutions meant to track and contain large-scale change, whether a $32 trillion AI infrastructure buildout or the mountain systems reshaped by a warming climate, are struggling to keep pace with how fast that change is now arriving.
Geopolitical Concerns
The long arm of the Chinese state
The Economist
What Trump’s Venezuela Oil Grab Means for China
Bloomberg
The Long Shadow of the Iran Shock
Jason Bordoff and Meghan L. O’Sullivan, Foreign Affairs
Putin, war leader and election campaign chief, prepares Russia for military escalation
Benjamin Quénelle, Le Monde
After Far-Right Triumph, Merz Faces Narrow Path to Save His Government
Bertrand Benoit, The Wall Street Journal
Beijing is formalizing a body of law designed to project its authority beyond its own borders, mirroring the extraterritorial reach Washington has long claimed for itself through sanctions and export controls. New rules tighten Beijing’s grip over outbound technology transfers and strengthen exit bans on individuals, building on a 2021 law that already punishes foreign firms for complying with sanctions Beijing deems illegitimate. JPMorgan Chase and Citigroup now face $44 million in Chinese court claims for freezing a sanctioned Chinese oil trader’s assets at the direction of the US Treasury, and a Chinese court ruling this June ordered a Singaporean firm to pay damages for refusing to deliver cargo to a sanctioned Hong Kong buyer. This use of law as an instrument of state power draws its force from China’s dominance of global supply chains, the same leverage now being tested in Venezuela: Washington’s deal granting a private US-backed company rights to a fifth of Venezuela’s oil reserves has already halted China’s purchases of Venezuelan crude, once roughly 80 percent of the country’s exports. Still, Chinese state firms retain claims to an estimated 4.4 billion barrels under arrangements that predate the agreement. The larger loss, though, is financial — China has lent Venezuela more than $100 billion since 2000, and the new deal complicates Beijing’s ability to recover the $10 billion to $20 billion still outstanding.
That same willingness to convert economic leverage into geopolitical power played out even more starkly in global energy markets. The prolonged closure of the Strait of Hormuz, through which a fifth of the world’s oil and gas normally passes, cut roughly five million barrels a day from China’s imports through June. Beijing offset the reduction by drawing down commercial stocks and restricting fuel exports to its neighbors, giving China a source of influence once reserved for Saudi Arabia’s spare production, while the war’s costs kept falling disproportionately on poorer importing countries that lack comparable buffers. Meanwhile, Russia leaned into a similar asymmetry between external conflict and internal control, as President Vladimir Putin ruled out a new mobilization ahead of parliamentary elections scheduled for September 18-20. He nonetheless warned of retaliation for Ukrainian strikes on Russian oil refineries and prepared the public for what analysts describe as an intensifying campaign on two fronts, military and domestic. Germany shows a different kind of strain building, one rooted in years of stagnant growth and deepening dissatisfaction with Chancellor Friedrich Merz. The far-right Alternative for Germany party won 43.8 percent of the vote in Saxony-Anhalt’s state election, more than doubling the vote share of Merz’s Christian Democratic Union, leaving him under pressure to abandon the informal firewall barring cooperation with the AfD. Abandoning it could fracture his own coalition and would hand more influence to a party that has consistently opposed continued German support for Ukraine.
Geoeconomics
There Are Four Forces Pressuring Bonds: War Is No. 1
Jack Pitcher, Shradha Dinesh, and Ryan Dezember, The Wall Street Journal
Why Trump’s Venezuela Oil Grab Is No Quick Fix for Gas Prices
Benoît Morenne, The Wall Street Journal
The world’s $2tn interest bill
Ian Smith, Sam Fleming, and Emily Herbert, Financial Times
Michael Pettis, Foreign Affairs
World’s Unusually High Dollar Exposure Risks Fueling Selloff
Ruth Carson, Masaki Kondo, and Anya Andrianova, Bloomberg
Government borrowing costs are climbing worldwide as a mix of war-driven inflation, fiscal profligacy, and a flood of new debt issuance collide. The average benchmark 10-year yield across G7 economies has reached 4 percent for the first time since 2008. The rise is driven substantially by the surge in energy prices that followed the closure of the Strait of Hormuz and Ukraine’s campaign against Russian refineries, which pushed US diesel prices to $5.68 a gallon this month, within a fraction of the record set after Russia’s 2022 invasion of Ukraine. That pressure is unlikely to ease soon: the Trump administration’s plan to use Venezuelan oil to refill the depleted Strategic Petroleum Reserve could take months to yield new barrels and offers no clear mechanism for lowering prices at the pump in the near term. Governments now face the added burden of refinancing existing debt at borrowing costs not seen in nearly two decades, meaning interest bills climb even without any new deficit spending. The OECD’s 38 member states paid more than $2 trillion in debt interest in 2025, or 3 percent of collective GDP, and the United Kingdom, France, and the United States now each spend more on debt service than on defense. Total global government and corporate debt has climbed toward $300 trillion, and America’s Congressional Budget Office projects that US debt costs will double over the next decade and surpass Social Security spending after 2047.
Global trade and currency markets raise a similar question, as persistent trade surpluses in China and Germany, financed for years by the United States running the corresponding deficits, are approaching a reckoning that history suggests rarely resolves painlessly. China’s mounting debt burden, the fastest rise relative to GDP in recorded history, leaves Beijing under pressure to preserve its surplus for as long as possible rather than absorb the adjustment at home. Washington, though, has both the economic scale and the political centralization to push that cost back onto trade partners through tariffs and industrial policy, leaving Europe, fragmented and politically constrained, most exposed to absorbing whatever China and the United States decline to bear themselves. A related shift is unfolding in currency markets, where the dollar’s usual role as a safe haven during volatility is being questioned. Investors across Japan, Canada, and Taiwan hedged just 41 percent of their foreign-currency exposure as of June 30, the lowest share since at least 2015. The retreat reflects narrowing interest-rate gaps and doubts about the dollar’s reliability during market stress. A five-percentage-point increase in hedging ratios among major foreign holders of US assets would translate into roughly $230 billion in dollar-selling transactions, a scale large enough to accelerate the currency’s decline without requiring any investor to sell a single US bond or stock.
Global Junctions
Yes, We’re in an AI Bubble. Just Look to 1980s Japan.
Howard W. French, Foreign Policy
Nvidia is driving the AI boom. Good
The Economist
Artificial General Intelligence Forecasting and Scenario Analysis
Gopal P. Sarma, Sunny D. Bhatt, Michael Jacob, and Rachel Steratore, RAND Corporation
China’s hypercompetition goes global as Beijing frets over backlash
Wataru Suzuki and Stella Yifan Xie, Nikkei Asia
Nvidia’s latest earnings have sharpened a debate over whether the AI investment boom resembles Japan’s asset bubble of the late 1980s, whose collapse in December 1989 took 34 years to fully unwind. Nvidia’s net profit doubled to roughly $60 billion on revenue of $96.22 billion, and the company alone accounts for 8 percent of the S&P 500’s total value. Critics point to its growing practice of financing the very customers who buy its chips, a circular arrangement that recalls the vendor financing behind the dot-com crash of 2000-01, and warn that Nvidia’s biggest customers, including OpenAI, are still struggling to turn AI into profit at a pace that matches their infrastructure spending. Chinese firms building cheaper open-source models on less expensive chips are undercutting the economics on which the entire bet depends. Others argue the comparison overstates the risk. Nvidia’s balance sheet carries $99 billion in cash and gross margins of 75 percent. That lets it back its customer financing with cash rather than debt, and even if its cash flow flattened next year, the company would still be less leveraged than all but 39 non-financial firms in the S&P 500 by 2028.
Underlying the Nvidia debate is a forecasting problem that extends well beyond any single company: expert estimates for when artificial general intelligence arrives have shifted markedly toward the near term across independent methods. The field nonetheless lacks resolved forecasts for calibration, benchmarks resistant to gaming, or continuous real-time insight into model capabilities, leaving investors and policymakers to bet on a technology whose trajectory nobody can reliably measure. A similar mismatch between scale and oversight is playing out in Chinese manufacturing, as Chinese companies’ overseas revenue reached a record 22.7 percent of total revenue last year, more than double its 2019 share. Cutthroat domestic price competition, what Beijing calls involution, is pushing firms to pursue higher profit margins abroad instead of at home. Exports of solar cells, batteries, and electric vehicles jumped 52 percent in the first half of 2026 to $116 billion, though Chinese automakers’ total profit fell 20 percent while they exported 2.3 million vehicles in the same period, and Geely’s Hong Kong-listed arm now earns several times more on a car sold abroad than on one sold at home. Germany’s Economic Institute attributes roughly 400,000 of 520,000 manufacturing job losses since 2019 to this competition, and Beijing’s own anti-involution campaign, which includes summoning executives and banning new industrial subsidies, has done little to curb local governments’ support for national tech champions. Neither trajectory, AI’s capability curve or China’s industrial overcapacity, is one that existing institutions seem able to reliably forecast or contain.
Global Trajectories
AI Data Center Spending to Reach $32 Trillion by 2050, PwC Says
Saritha Rai, Bloomberg
China Makes Cars So Quickly That Regulators Are Getting Worried
Bloomberg
The world may have less time than it thinks on climate change
Tim Palmer, The Economist
Nepal’s Catastrophic Flood Heralds a New Age of Disasters
Laura Millan, Bloomberg
Global spending on AI data centers is set to reach $31.6 trillion through 2050, according to PwC. The firm says the buildout dwarfs the historical scale of railways, electrification, and the internet combined and could climb to $50 trillion if adoption accelerates further. Nearly half of that spending, $15.1 trillion, will land in the United States, and unlike prior infrastructure cycles, the bulk of the cost comes from replacing hardware every four to six years rather than a single upfront buildout, meaning the spending never really front-loads and shows no natural point at which it slows. A comparable acceleration is reshaping China’s auto industry, where artificial intelligence and virtual-reality testing could compress the time needed to bring a new vehicle to market to just 18 months, down from roughly two years, far below the three to five years typical of legacy foreign automakers. Chinese regulators have responded with a yearlong campaign of surprise inspections, tighter battery and steering standards, and a proposal to double mandatory road testing to 30,000 kilometers, after a wave of recalls that includes more than 4.27 million vehicles across nine automakers for door-handle defects. In both cases, the pace of the buildout is outrunning the institutions meant to validate it, whether that means power grids and semiconductor supply chains for data centers or crash-testing protocols for cars developed on compressed timelines.
That same gap between the pace of change and the capacity to track it is showing up in climate science as well. This summer’s persistent high-pressure system over Western Europe produced record heat and buckled rail lines across southern England. It may reflect ordinary chaotic variability, or it may reflect a deeper shift in the patterns that determine where such high-pressure systems form, a distinction current global climate models cannot resolve because they are not built to capture the small-scale weather interactions that keep a system like this locked in place. If the deeper shift is the real driver, the world could be approaching a tipping point that leaves little time to adapt gradually, rather than the decade or two of lead time earlier predictions assumed. Nepal’s Himalayan disaster in late August illustrates what that kind of acceleration already looks like on the ground. A section of glacier one to 1.3 kilometers wide detached near the Langtang Lirung peak, unleashing a deluge that raced downstream, raising river levels by up to 9 meters within half an hour, killing at least 1,250 people, and leaving 4,200 missing. Researchers have documented a marked increase in glacial lake outbursts and mountain floods since 2021, driven by permafrost thawing faster than anticipated at nearly every monitored site worldwide, yet warning systems remain sparse in remote Himalayan valleys where, as one researcher put it, there’s nobody up there watching the slope.
