On a Thursday morning in mid-July, a man in Chicago stepped outside and tasted metal. The sun over Toronto had turned the colour of a bruised yolk. In Manhattan, the skyline simply vanished behind a wall of particulate matter so dense that public-health officials recommended no one—no one—spend time outdoors. More than a hundred wildfires were burning across Ontario, and the smoke had drifted south in a slow, suffocating blanket that settled over the Great Lakes states and the Northeast corridor, trapping itself beneath a heat dome that had already pushed temperatures into the triple digits. One hundred and fifteen million people breathed air rated unhealthy or worse. In New York, Mayor Zohran Mamdani stood before cameras and said the word dangerous four times in a single sentence.
This is no longer an event. It is a season. The Canadian wildfire seasons of 2023 and 2025 were the worst on record; 2026 is tracking to join them. The smoke that once descended for a day or two now lingers for a week, and the political response has calcified into a familiar choreography: American senators threaten tariffs on Canada, Canadian premiers request military evacuations, and the sky stays orange. The Economist ran a cover this week on “global dimming”—the measurable decline in Earth’s albedo, the planet absorbing more sunlight than it reflects, a feedback loop in which warming begets further warming through mechanisms scientists are only now quantifying. The smoke over the Midwest is the visible symptom of an invisible arithmetic that has already tipped past comfort.
What makes the week’s atmospheric crisis emblematic is not its severity alone but its persistence without resolution. No one expects the fires to stop. No one expects the smoke to clear permanently. The infrastructure of daily life—schools, outdoor labour, childhood play—adjusts incrementally to conditions that would have been unthinkable a decade ago. And this quality of escalation without exit, of crisis normalised into weather, is the precise pattern that defined nearly every other domain of global affairs this week.
The videos appeared on social media before any government confirmed them: a highway bridge in southern Iran, reduced to rubble, its concrete spans collapsed into a dry riverbed. Then another. Then a water-treatment facility, its pumps silent. By Friday, the United States had conducted its seventh consecutive day of airstrikes on Iranian territory, and for the first time the target set had expanded beyond military installations to include what Pentagon officials euphemistically called “logistics infrastructure that can also serve civilian needs.” A control tower at an Iranian port. Bridges. Power stations. The architecture of ordinary life, rendered as collateral.
Iran responded by firing missiles and drones at American military facilities in Bahrain, Kuwait, Jordan, and Oman. Kuwait reported that its power and water plants had been hit, generators damaged, fires burning. The Islamic Republic’s parliament speaker—also its chief negotiator—called the conflict an “existential war.” The Houthis in Yemen threatened to close the Bab el-Mandeb strait, the Red Sea’s southern gateway, if American forces struck Iran’s electrical grid. And the Strait of Hormuz itself, that narrow waterway through which a fifth of the world’s oil normally transits, fell nearly silent. Observable commercial traffic dropped to a trickle: a handful of Iran-linked vessels using a northern route approved by Tehran, and almost nothing else.
The oil price told its own story. Brent crude surged more than eleven per cent over the week, its steepest weekly climb since April, settling near eighty-eight dollars a barrel. American petrol tanks filled at prices not seen since the war’s initial phase in March. And yet the conflict’s most revealing moment was not a missile strike but a presidential reversal: Donald Trump announced a twenty-per-cent toll on all cargo transiting the Strait of Hormuz, then abandoned the idea within twenty-four hours, replacing it with a vague promise of Gulf-state investment in American infrastructure. The Financial Times described the episode as a “thought bubble.” The Atlantic called it a symptom of a presidency whose efforts to subvert the coming midterms have “largely come up short,” leaving the president with fewer tools than grievance.
The structural comparison that analysts reached for this week was not Iraq or Afghanistan but Vietnam: the “short-war fallacy,” the belief that overwhelming firepower can compel political surrender. Lawrence Freedman, the emeritus war-studies professor at King’s College London, noted that leaders who possess powerful militaries but lack political strategies “set objectives that can be achieved, if at all, only through prolonged struggle.” The interim peace deal signed a month ago is in tatters. Neither side appears capable of backing down. The escalation spiral, as Bloomberg Economics put it, has become self-sustaining: “Neither side willing to back down.” And the global economy, which had briefly exhaled in June when the ceasefire held, inhaled again.
In a semiconductor fabrication plant in Hsinchu, Taiwan, the machines kept running. TSMC reported a seventy-seven-per-cent year-on-year jump in quarterly profit and announced another hundred billion dollars in American factory investment. Its earnings beat every estimate. Its guidance rose. And its stock fell.
The reaction was not irrational. It was the market’s way of saying that perfection is now the minimum requirement. Across the week, a selloff in chipmakers gathered pace, driving the Philadelphia Semiconductor Index twenty per cent from its record high—its worst week since the “Liberation Day” tariff rout of April 2025. The trigger was not a single event but a convergence: TSMC’s numbers, however strong, were not strong enough to justify the valuations already priced in; IBM cratered twenty-five per cent in a single session after acknowledging that its customers were redirecting spending from software toward AI hardware; and a Chinese startup called Moonshot released an open-weight model, Kimi K3, that its creators claimed rivalled the frontier offerings of OpenAI and Anthropic on key benchmarks.
The Moonshot release landed with the force of a second “DeepSeek moment.” Xi Jinping appeared in person at the World Artificial Intelligence Conference in Shanghai—the first time China’s paramount leader had attended—declaring that AI development “should not be a solo performance by a single country, but a symphony of international cooperation.” The line was aimed squarely at Washington’s export-control regime. Hours later, twenty-nine countries signed up to China’s new World Artificial Intelligence Cooperation Organization, a counterweight to the American-led Pax Silica alliance. The geopolitical architecture of AI was bifurcating in real time.
And yet the week also revealed the limits of the AI narrative as a market engine. SpaceX, the year’s most anticipated IPO, saw its shares dip below the one-hundred-and-thirty-five-dollar offering price for the first time, erasing more than a trillion dollars in paper value from Elon Musk’s rocket-and-AI conglomerate. Netflix disappointed with its slowest revenue-growth forecast in three years. Google was reported to be months behind on its flagship Gemini 3.5 Pro model. The fund-manager survey from Bank of America found that more than eighty per cent of respondents identified semiconductors as the world’s most crowded trade—the highest consensus reading ever recorded. The most crowded “most crowded” trade ever, as John Authers wrote in his Bloomberg column. Everyone in the room was looking at the same door.
The tension is structural: the AI buildout requires capital expenditure on a scale that dwarfs the revenues it has yet generated. The International Energy Agency noted this week that the capital spending of five technology companies now exceeds the entire world’s investment in oil and natural gas production. The question is no longer whether AI will transform the economy but whether the financial architecture supporting its construction can survive the gap between promise and delivery.
At nine o’clock on a Thursday evening, the President of the United States stood before a teleprompter and read, for thirty minutes, from a script about an election he lost six years ago. The speech was billed as a major address on election security. What it delivered was a catalogue of recycled allegations—China had “illicitly acquired” two hundred and twenty million voter files; the “deep state” had covered up foreign meddling; the 2020 result remained suspect. The declassified documents released alongside the speech contained no evidence that any foreign actor had altered vote tallies. ABC and NBC declined to carry the address live.
The speech’s most revealing detail, however, was not its content but its context. Earlier that day, it emerged that the president’s own teleprompter operator—the man physically loading the words Trump would read—had been placed on unpaid leave after the prediction-market platform Kalshi flagged approximately one hundred thousand dollars in trading profits tied to bets on what the president would say in his speeches. Separately, Trump Media and Technology Group announced a product called “Truth API,” a paid data feed giving algorithmic trading firms millisecond-early access to the president’s social-media posts. The New York Times reported that Trump had personally bought stocks in companies days before promoting them on Truth Social.
The tableau was complete: a president using the apparatus of the state to litigate a settled election, his speechwriter gambling on the content of the address, his media company selling privileged access to his words as a financial instrument, and the words themselves amounting to what the Atlantic called “a cry of frustration from a man whose efforts to subvert the 2026 midterms have largely come up short.” The SAVE America Act, his signature legislative vehicle for overhauling election administration, had stalled in the Senate. Executive orders had been blocked by courts. The Election Assistance Commission had been decapitated. What remained was spectacle: the prime-time slot, the grievance, the performance of authority without its substance.
And yet the performance mattered. Not because it would change any law or reverse any result, but because it deepened the epistemic fracture that the Atlantic identified this week as the “United States of disbelief.” The same issue carried a dispatch on conspiracy theories surrounding the death of Senator Lindsey Graham and the prolonged absence of Senator Mitch McConnell—rumours so pervasive that even Trump felt compelled to say the FBI was “wasting their time.” The line between prudent scepticism and corrosive cynicism, the magazine argued, had become nearly impossible to discern. The president’s speech did not cross that line. It was the line, drawn in prime time, for an audience of millions.
In the eighty-fifth minute of a World Cup semi-final in Atlanta, with England leading one-nil and the ghost of 1966 finally within touching distance, Lionel Messi received the ball near the right touchline. What followed took four minutes and two assists. Argentina scored twice. England’s players stood with hands on hips, staring at nothing. The final whistle confirmed a 2–1 victory for the reigning champions, and somewhere in the stands a banner was unfurled: Las Malvinas son Argentinas.
The Falklands. Always the Falklands. The two countries had not met at a World Cup since 2002, and the match carried the accumulated weight of 1982’s seventy-four-day war, 1986’s Hand of God, and a territorial dispute that the discovery of the Sea Lion oil field—expected to begin production in 2028—has rendered newly material. Argentina’s vice-president had called the English “invaders” and “usurping pirates” earlier in the week. The British government, after the match, issued a statement dismissing the banner as a violation of FIFA’s rules against political messaging. The Financial Times noted that the Sea Lion project would give the Falkland Islanders a per-capita income exceeding that of the United Kingdom. The balance of forces, on and off the pitch, was shifting in ways neither side could fully control.
And yet the tournament itself—the expanded forty-eight-team format, the record six million tickets sold, the forty-six million American viewers who watched the United States lose to Belgium—had succeeded by every commercial measure. FIFA’s embrace of dynamic pricing, its championship rings, its planned halftime show featuring Justin Bieber and Shakira, its cosying-up to authoritarians: all of it had been absorbed into the spectacle without dimming the public’s appetite. The Atlantic‘s Jacob Stern wrote of a “sports overload,” a calendar so saturated that the World Cup no longer interrupts the flow of professional sport but merely adds to it. And still the games delivered: Cape Verde’s forty-year-old goalkeeper Vozinha drawing with Spain and Argentina; Norway’s Erling Haaland bringing home a taxidermy raccoon; the tiny island nation’s Cinderella run. The beautiful game, indifferent to the machinery surrounding it, kept producing moments that no algorithm could predict and no amount of corporate packaging could manufacture.
A goldfish in Toronto, swimming toward one of two flags in a miniature pitch, had called fourteen of eighteen matches correctly by the group stage. Five AI models, tasked with the same exercise before the tournament, managed between fifty and sixty per cent accuracy. The fish outperformed the machines. The week’s most reliable oracle had no parameters, no training data, no compute budget. It had a tank, a flag, and an indifference to narrative that no language model could replicate.
In a conference room in Frankfurt, the chief executive of Morgan Stanley reported six point three billion dollars in quarterly equity-trading revenue, a sixty-nine-per-cent increase and an all-time record. Goldman Sachs posted seven point four billion, the highest quarterly trading haul in Wall Street history. JPMorgan, Bank of America, Citigroup: all shattered expectations. The KBW Bank Index hit a post-financial-crisis high. The markets were booming, Jamie Dimon said. Then he added, in the geological metaphor he has recently adopted, that “several risks are shifting below the surface like tectonic plates.”
The tectonic plates were visible everywhere. China’s GDP grew 4.3 per cent in the second quarter, its slowest pace in more than three years, as domestic consumption stagnated and the property crisis deepened. Yet exports surged twenty-seven per cent in June; the monthly trade surplus widened to one hundred and twenty-five billion dollars. The contradiction was not a contradiction: weak domestic demand did not idle Chinese factories; it redirected their output outward, flooding global markets with manufactured goods that local industries could not match on price. Germany’s car sector was the prime casualty. The Financial Times reported that Chinese brands accounted for ten per cent of European car sales in May, a milestone. In Britain, a Chinese-made SUV nicknamed the “Temu Range Rover” had become the country’s best-selling vehicle.
The fund-manager consensus, captured in the Bank of America survey, was “no landing”—neither soft nor hard, simply continuation. Cash allocations fell below four per cent, a level the bank’s strategists historically flag as a sell signal. The optimism was predicated on oil finishing the year between seventy and eighty dollars. Brent was already at eighty-eight. The assumptions were backward-looking, frozen at June 30, when the ceasefire held and Hormuz traffic was normalising. The data described a world that no longer existed.
And in the background, quieter but no less consequential: three reports showed Americans struggling to save for retirement as living costs rose. The Social Security trust fund is projected to be depleted by 2032. The national debt is heading toward forty trillion dollars. The cyclospora outbreak linked to Taco Bell lettuce had sickened more than sixteen hundred people across thirty-four states, and the food-safety infrastructure meant to catch such events had been hollowed out by funding cuts and layoffs. The FDA physically inspects roughly one per cent of imported food. The system, one expert said, was a “Jenga tower” with blocks being removed.
On a Monday in late July, King Charles III will ask Andy Burnham to form a government. The former mayor of Manchester, who won the Labour leadership uncontested after Keir Starmer’s resignation, will enter Downing Street with a mandate that is, in the words of one Bloomberg analysis, “never going to be more popular than it is now.” The challenge is not winning but governing: reviving an economy that has stagnated, managing a national debt that constrains spending, and holding together a party whose left wing is already aggrieved that Shabana Mahmood—seen as centrist, economically untested—appears set to become Chancellor over Ed Miliband.
Burnham inherits a country in which the number of higher- and additional-rate taxpayers has risen thirty-five per cent in three years through fiscal drag; in which Thames Water teeters on the edge of state control; in which the National Health Service waits lists have become a political metaphor for institutional decay. He inherits, too, a foreign-policy landscape in which the American security guarantee is unreliable, the Iranian war is reshaping energy markets, and the European Union is simultaneously deregulating its banks and tightening its carbon rules. The Financial Times noted that business leaders are already asking, “Who do we call?”—a question that reveals how thoroughly Starmer’s government had centralised decision-making, and how little infrastructure Burnham has built for the transition.
In Ukraine, a parallel drama of institutional succession played out in miniature. President Volodymyr Zelenskyy fired his defence minister, Mykhailo Fedorov, the thirty-five-year-old architect of the country’s drone-warfare revolution, and thousands took to the streets in protest. Fedorov had clashed with the old guard—generals who saw his technocratic vision as fanciful, defence contractors whose businesses his reforms threatened. His dismissal, six months into the tenure, suggested that the war’s institutional politics had become as contested as its battlefield strategy. The timing was cruel: Ukraine’s forces had recently scored significant gains against Russian shipping in the Sea of Azov, and the momentum was, for the first time in years, tilting in Kyiv’s favour.
Christopher Nolan’s The Odyssey opened in theatres this week, the first commercial feature shot entirely in IMAX 70-millimetre film. The cameras weigh hundreds of pounds. They roar. They require a sound blimp to mute their mechanical scream. Only about two dozen theatres in the world can project the film in its intended 1.43:1 aspect ratio; the rest will show cropped versions, excising as much as forty per cent of the image. Fans are flying cross-country for the true experience. Resale tickets have appeared on eBay for hundreds of dollars. The marketing insists there is only one way to see the film. Most people will see it another way, and most will not mind.
The New York Times critic Manohla Dargis called it “a classic in every sense, a transporting affirmation of the art and a work of pure cinema.” The Atlantic‘s A.O. Scott, writing from the Book Review rather than the film desk, noted that the original Odyssey is “one of the earliest and most powerful literary expressions of nostalgia”—a word compounded from the Greek for “home” and “pain.” Nolan’s film, arriving in a moment of streaming fragmentation, AI-generated content, and collapsing theatrical economics, is a two-hundred-and-fifty-million-dollar wager that people will still show up for a large-scale spectacle that evokes the grandeur of an earlier time. The wager appears to be paying: the domestic box office is having its strongest year since 2019.
And yet the week also contained a quieter image of cultural persistence. In Warsaw, the Museum of Modern Art hosted “In the Very Bowels of Change: Surrealism and Antifascism,” an exhibition tracing the movement from Buñuel’s Golden Age through the Spanish Civil War, the Nazi designation of “degenerate art,” and the postwar cabarets of divided Berlin. Claude Cahun and Marcel Moore’s “paper bullets”—notes taunting the Gestapo on occupied Jersey—were on display. Cahun, sentenced to death, had responded to the verdict by asking which sentence should be carried out first. The war ended before either could be.
The exhibition’s argument was that Surrealism was never merely a style—never the melting clocks of dorm-room posters—but a political technology, a method of psychic liberation aimed at the authoritarian mind. Its practitioners understood, as the week’s events repeatedly confirmed, that the struggle over reality is the first struggle of any conflict. The bridges burn. The smoke settles. The teleprompter loads its words. And somewhere, a goldfish swims toward a flag, indifferent to all of it, and calls the match correctly.
The pattern of the week was not crisis but the normalisation of crisis: the escalation that becomes weather, the speech that becomes noise, the war that becomes forever. What remains unresolved is not any single conflict but the question of whether the institutions meant to resolve conflicts—governments, markets, courts, alliances—retain the capacity to do so, or whether they have become, like the sidewalk sheds of New York City that Bloomberg CityLab profiled this week, structures erected temporarily to protect against falling debris and then left standing indefinitely, blocking the light, obscuring the view, serving no one, removed by no one, part of the streetscape forever.
Week of July 15 – 18, 2026
There is a goldfish in a Toronto storefront tank, fitted out as a miniature football pitch, who has been calling World Cup matches with the calm authority of a hedge-fund manager. Midway through the group stage, his handlers at ScoreGPT — a tongue-in-cheek leaderboard that also grades five large language models on the same predictions — had counted fourteen correct calls against four misses, an accuracy of roughly 80%. Draws are no-contests, because there is no third flag. The machines — ChatGPT, Claude, Gemini, Copilot, Perplexity — have been running between 50% and 60%. None has matched the goldfish. (Rest of World, The Global Dispatch, 17 July 2026.)
It is the most beguiling image of the week, and the most revealing. The machines, asked to call the most-watched sporting event on earth, were outperformed by a vertebrate with a four-second memory and a tail. The humans, asked to forecast the same thing, got four of the four semifinals right — Spain, France, England, Argentina — and then watched the tournament’s own logic collapse that call when the two favourites met each other in the second round instead of the final. So who is calling what? And, more pressingly, what does it mean that the defining technology of the 2020s has been, this week, on a slide?
The price of admission to that question, this week, is half a trillion dollars. Between 8 July and the morning of 17 July, SK Hynix — the Korean memory chipmaker whose HBM3E chips sit inside the Nvidia accelerators that train the large models that beat the goldfish, badly — lost roughly $504 billion of market capitalisation, then regained $56 billion of it on a single Tuesday, and ended the week still up some $603 billion on the year. (Bloomberg, Points of Return, 15 July 2026.) That is the kind of number that makes the trading floors in Lower Manhattan sound, again, like the viral “stonks” meme. Goldman’s equities desk booked $7.42 billion in a single quarter, the highest single-bank equities haul in the history of Wall Street. JPMorgan took $6.03 billion. Morgan Stanley took $6.3 billion. Citigroup, by comparison, “merely” jumped 45%. (Bloomberg, Businessweek Daily, 15 July 2026.) On the same week, IBM — once the biggest company in America — fell 25% in a single session, the worst day in its 110-year history, because its software business couldn’t keep up with the demand for hardware. (CNBC, The Tech Download, 15 July 2026; FT, International morning headlines, 15 July 2026.)
This is the first part of our story. It is about a financial system that has been rebuilt, in less than two years, on the assumption that the demand for artificial intelligence is infinite. It is, increasingly, also the story of the rest of the world — of straits and oil tankers, of orange skies and a President going on television, of a country that can no longer tell the difference between a coup and a campaign stop. The goldfish calls matches. The market calls the price. The week called the question.
Picture the scene: it is the morning of Thursday, 16 July 2026, and Lower Manhattan has gone ochre. A toxic, Canadian-fuelled haze has drifted south from more than 100 wildfires burning out of control in Ontario, dropping air quality in Brooklyn, Queens and Manhattan to “unhealthy,” and in Toronto to the same. Zohran Mamdani, the new mayor, holds a press conference. “We are reaching into a level of air quality that is dangerous for every single New Yorker,” he says, looking older than his salary suggests he should. (Bloomberg, Evening Briefing Americas, 16 July 2026.) A man in Maine has just been shot dead by an ICE agent whose ex-wife says he had a long history of psychiatric issues; a baby sleeps on a pavement in Durban; a six-year-old in Spain has just been told by her teacher that it might hit 50°C in August. (El País, English Edition, 16 July 2026; Bloomberg, CityLab Weekly, 18 July 2026.) The week has a colour, and the colour is the colour of a planet trying to cough.
Three thousand miles east, in the Persian Gulf, the U.S. Navy is in its sixth consecutive day of strikes on Iran. The pretext is the Strait of Hormuz, the 21-mile-wide chokepoint through which a fifth of the world’s oil and gas normally transits. Iran, having watched its negotiating partners come and go, has been quietly strangling the strait, ship by ship, since the spring. The U.S. has now widened the campaign: a sanctioned supertanker has been hit deep inside the Gulf, near Kharg Island; Iran’s Houthi allies in Yemen have threatened the Red Sea as the alternative route. (Bloomberg, Evening Briefing Europe, 16–18 July 2026; FT, In Today’s FT, 16 July 2026.) Brent crude is up 13% on the week, the biggest weekly advance since April. Asia, which depends on Hormuz for roughly 70% of its seaborne crude, is bracing for what one analyst called “a pricey winter.” (FT, Emerging Markets, 16 July 2026.)
What stops the price from spiking to historic levels is, of all things, Beijing. The world’s largest oil buyer has, in the past six months, been quietly stuffing its strategic petroleum reserves to such a degree that, when Iran began its squeeze, China was effectively able to cut its imports of Middle East crude, releasing global supply elsewhere in the world. “The Chinese cut in oil purchases helped other purchasers in the international market, alleviating the spike somewhat,” Juscelino Colares, a law professor at Case Western Reserve, told Newsweek. (Newsweek, The Bulletin, 17 July 2026.) It is one of those wonderful ironies of late-stage petro-politics: the country whose leader spent the week in Shanghai telling the World AI Conference that “AI development should not be a solo performance by a single country, but a symphony of international cooperation” (Xi’s actual words, 17 July 2026) is, in the same week, the country that has been quietly de-coupling the global oil market from the West’s worst crisis of the year.
And underneath both stories — the orange sky, the burning strait — is a third one. The Economist runs two covers this week. The first is on the Iran war; the second is on the Department of Justice. The third, for the print edition, is on global dimming: a long, almost elegiac piece of climate reporting arguing that the Earth has been absorbing a lot more sunshine than it used to, because of fewer clouds and lower sulphur emissions, and that this is the climate variable that is alarming scientists most. (The Economist, This Week, 16 July 2026.) The albedo — the fraction of sunlight that a surface reflects — is dropping with surprising speed. The planet is getting darker, in the precise technical sense, and warmer as a result. Wildfires like the ones choking New York are both a cause of that darkening and a consequence of it. They are also, as the FT points out, exempt from compliance calculations under the U.S. Clean Air Act, classified as “exceptional events” — a polite bureaucratic phrase for “we knew this would happen and we chose not to act.” (Bloomberg, CityLab Weekly, 18 July 2026.) The U.S. House of Representatives, in the same week, passes a bill to make Daylight Saving Time permanent. If it becomes law, the sun will not rise in New York until 8:20 a.m. in the depth of winter; in Bismarck, North Dakota, 9:30. (Newsweek, The 1600, 16 July 2026.) This is a country that cannot decide whether to acknowledge that the world is on fire, or to pretend it is just an hour later than it used to be.
The through-line here is interconnection, the kind that used to be a political slogan and is now a thermodynamic one. The orange sky over Manhattan is connected to the war in the Persian Gulf, which is connected to the strategic decisions of the Chinese state, which are connected to the gyrations of the global oil market, which are connected to the still-stuck pin that is the U.S. Federal Reserve’s next move on interest rates, which is connected to whether your mortgage is going to be affordable in two years’ time. The most arresting single sentence I read this week was buried in the Rest of World newsletter: “The capital spending of five technology companies now exceeds what the entire world invests in producing oil and natural gas.” (Rest of World, The Global Dispatch, 17 July 2026, citing the IEA.) That is a historical inversion. The Information Economy has, in five years, outgrown the Energy Economy in gross capital commitment. The climate is on fire; the silicon is hotter.
It is 9 p.m. in the East Room of the White House, Thursday, 17 July 2026, and Donald Trump is sitting behind the Resolute Desk, the optics of a serious president in a serious building, about to deliver a primetime address to the nation. The topic, the White House says, is “free and fair elections.” ABC, NBC and CNN decline to carry it live on their main platforms. CBS does. (Newsweek, The Bulletin, 17 July 2026.) He speaks for thirty minutes. He accuses China of stealing 220 million American voter files and “attempting to manufacture illegal ballots for Joe Biden.” He says the CIA had intelligence in 2020, “yet buried by rogue bureaucrats,” that would have stopped the alleged theft. (DW, Daily Bulletin, 17 July 2026; Newsweek, The 1600, 17 July 2026.) The intelligence community, as it happens, has said for years that the documents in question are scenarios and worst-case planning, not evidence of ballot manipulation. The voter files Trump cites are public records, available for free online or by political-party purchase. (Newsweek, The 1600, 17 July 2026.) The point is not the substance. The point is the setting, the timing, the audience.
Three days earlier, the same president’s longtime teleprompter operator was placed on unpaid leave after it emerged that he had made more than $90,000 in profits on the prediction market Kalshi by betting on the contents of the president’s own speeches. (CNBC, The Tech Download, 17 July 2026.) Three days after that, Trump Media and Technology Group announces “Truth API,” a paid-for data feed that will let trading firms receive the president’s Truth Social posts milliseconds before the rest of the world. (FT, International morning headlines, 16 July 2026.) On the same day, the FT reports that the Trump sons personally invested in a small American tungsten-mining company, Kaz Resources, and that their father personally lobbied the Kazakh government for the concession; the U.S. government has now earmarked $1.6 billion in taxpayer funds to help the family-adjacent company mine a metal in which China dominates global supply. (Bloomberg, Businessweek Daily, 15 July 2026; FT, The war over wealth tax, 16 July 2026.)
What we are watching, in other words, is not just a presidency but a vertically integrated one — a system in which the head of state, his family’s business interests, his social-media platform, his teleprompter operator, his prediction-market bets, his son’s mining concessions, and the country’s voting infrastructure have begun to fuse into a single, only partly legible apparatus. “He just needs to stir the pot enough to raise questions of illegitimacy for future elections that don’t go the way he wants,” writes Carlo Versano in The 1600. “Americans will hear ‘China’ and ‘rogue bureaucrats’ and ‘election integrity’ and tune the rest out.” (Newsweek, The 1600, 17 July 2026.) The architecture of grievance is now the architecture of the state. And the people who were supposed to be the institutional counterweights — the F.B.I., the C.I.A., the D.O.J., the Smithsonian, the Election Assistance Commission — are being hollowed out in the same week. (The Economist, This Week, 16 July 2026; ARTnews, 15 July 2026.) About a quarter of the D.O.J.’s lawyers have left. The cryptocurrency and public-corruption divisions have withered. The acting Attorney General, Todd Blanche, faces a Senate confirmation hearing in which survivors of Jeffrey Epstein call him “abrasive” and “condescending” in a private meeting, and his own party’s Senator Cornyn expresses “concerns” about his nomination. (Newsweek, The Bulletin, 17 July 2026.) The country that built the postwar international order is, in the summer of 2026, watching its instruments of self-correction fail, one by one, like a string of bulbs burning out on an old holiday string.
And into the breach: Hegseth. Pete Hegseth, the Defence Secretary, announces a new Pentagon programme to test every active-duty service member over the age of 30 for testosterone deficiency. (The Economist, Today, 17 July 2026.) It is, the magazine notes, “not a crazy idea” in itself — soldiers are, statistically, at higher risk of hypogonadism. But in the kulturkampf context of an administration that performs masculinity as a daily ritual — push-ups in public, helicopters in combat fatigues, “cockroach” market fears, geological “tectonic plates” — it reads as something else. It reads as a confession of anxiety. A state that is not confident in its own institutions, and not confident in its own legitimacy, is a state that will go looking for testosterone. (The Economist, Today, 17 July 2026.)
The second most arresting image of the week is invisible: it is a wire, twenty atoms wide, on the surface of a silicon wafer in a cleanroom in Hsinchu, Taiwan. That wire — printed, in 2026, by a deep-ultraviolet lithography machine that only one company on earth knows how to make — is the substrate on which the entire AI economy is built. That company is ASML, the Dutch monopoly whose shares jumped 75% in the year to mid-July, whose chief executive, Christophe Fouquet, says demand has been so strong that his customers are “not only increasing their capital spending, but also accelerating all their plans.” (FT, International morning headlines, 15 July 2026.) On 15 July, ASML raised its full-year sales guidance for the second time this year, to between €43 billion and €45 billion. (Bloomberg, Evening Briefing Europe, 15 July 2026.) The next day, TSMC — the company that actually fabricates the chips — reported a 23.4% jump in second-quarter profit and raised its own guidance. (CNBC, The Tech Download, 16 July 2026.) By Friday, both stocks were down on the week. The buyers had not materialised in the volume that the fundamental story implied. (Bloomberg, Stocks Slide as Chip Selloff Deepens, 17 July 2026.)
This is the second-order story of the week, and it is the one that the goldfish might, somehow, have seen coming. Eight of ten fund managers surveyed by Bank of America between 2 and 9 July now believe semiconductors are the world’s most crowded trade — the strongest consensus on this question in the survey’s history. (Bloomberg, Points of Return, 16 July 2026.) In South Korea, leveraged single-stock ETFs tracking SK Hynix and Samsung have driven the broader Kospi index up 6% in a single session and down 10% in the next, and the country’s own President, Lee Jae Myung, has had to publicly urge calm. (Bloomberg, Evening Briefing Asia, 15 July 2026; FT, Emerging Markets, 16 July 2026.) It is, the FT writes, the behaviour of “a casino for retail investors” — and, like all casinos, it is one bad hand away from a liquidity event. The leverage is the risk. The risk is being mispriced.
What is doing the mispricing, of course, is China. On 17 July, Moonshot, a Beijing-based AI startup with the heft of a Chinese OpenAI and the velocity of a DeepSeek sequel, released Kimi K3, an open-weight model that, on a clutch of standard benchmarks, matches or exceeds the most recent Claude and GPT-4 class systems. (Bloomberg, Morning Briefing Americas, 17 July 2026; FT, Emerging Markets, 17 July 2026; Newsweek, Geoscape, 17 July 2026.) The release was a near-perfect replica, in form, of the DeepSeek moment of January 2025: a Chinese lab, a shock announcement, a rout in the U.S. semiconductor complex, a global repricing of who has the technological lead. Xi Jinping, speaking at the same moment at the World AI Conference in Shanghai, called AI “a symphony of international cooperation” — a deliberate and, one suspects, slightly amused repudiation of the U.S. “Pax Silica” alliance that has tried to lock China out of the highest-end of the stack. (Newsweek, Geoscape, 17 July 2026.) The new China-led World Artificial Intelligence Cooperation Organization has, in seven months, signed up 29 countries. Pax Silica has 24. The race is not over. It is, in fact, the only race in town.
But there is a long view here that the headlines are not capturing, and it is worth pausing on. In the Economist‘s reading of the week, the structural fact is that no country — not America, not China, not the E.U. — can fully decouple from the others on AI. “Sovereign AI, independent of America and China, is a pipe dream,” the magazine notes in a Babbage editorial. (The Economist, Today, 16 July 2026.) The chips are made in Taiwan. The lithography is Dutch. The advanced memory is Korean. The model weights are split. The talent is everywhere and nowhere. The most-cited 25-year-old founder in the world right now is Liang Wenfeng of DeepSeek, whose net worth has doubled in the last funding round to $36 billion. (Bloomberg, California Edition, 17 July 2026.) We are not in a bipolar arms race. We are in a barycentre — a swirl, a knot, a system in which the position of every actor is determined by the position of every other. The “DeepSeek moment” is, in fact, the Moonshot moment, and the Moonshot moment will not be the last. The system has no final state. That is what makes the leveraged ETFs, the $64-billion capex plans, the $1.5-trillion semiconductor buildout, both intoxicating and terrifying. (FT, International morning headlines, 16 July 2026.)
The longer-historical reference is unavoidable. In his General Theory, Keynes warned that “worldly wisdom teaches that it is better for reputation to fail conventionally than to succeed unconventionally.” The leveraged-ETF crowd in Seoul, the small-cap miners in Idaho whose retirement plans have just become $1 million larger thanks to Micron’s 700% rally (Bloomberg, Businessweek Daily, 15 July 2026), the New York retiree reading the Times in the haze — all of them are taking, in their different ways, an unconventional bet, with conventional money, on the proposition that the demand curve for compute will not bend. The bet is rational. It is also, by definition, the most crowded bet in the world.
In a small town in Idaho, an investment manager named JT Belnap is having a routine follow-up call with a client. “Has anything changed since we met eight days ago?” “Well, it looks like I’m a million dollars richer.” The man works for Micron Technology, the Boise-based memory-chip company whose stock has almost ten-bagged in the last twelve months, and whose employees, having received decades of stock options, have just become, in paper at least, millionaires. (Bloomberg, Businessweek Daily, 15 July 2026.) One of them went out and bought the $100,000 truck he’d always wanted; another took her granddaughters to Disneyland; another made “a couple of sizable donations to charities.” The wealth effect — the macroeconomic term for what happens to a real economy when people who feel flush start spending — is now the single most important tailwind for U.S. consumer demand.
The image is, in its small way, the inverse of the orange sky over Manhattan. Up there, in finance, the world is making money faster than it can count it. Down here, on the street, the same people are breathing particulate matter, paying $12 for a pound of strawberries (Newsweek, The 1600, 16 July 2026), and being told by their President that the next election is going to be stolen by China. The two stories are not in contradiction. They are in complicity. The wealth effect is what gives a faltering administration its residual political viability; the orange sky is what the wealth effect is not being used to address. Goldman Sachs booked $7.42 billion in equities trading in three months. New York has no clean air.
The asymmetry is now a political fact. In California, a wealth-tax referendum is heading for the November ballot, and Silicon Valley’s billionaires have already spent tens of millions fighting it; the FT calls it “a bloody, and eye-wateringly expensive struggle.” (FT, FT Edit, 16 July 2026.) In a real way, the same week that saw Micron’s Idaho employees become millionaires also saw the structural argument over whether they should stay that way become the dominant domestic issue of the year. Larry Ellison’s Oracle, meanwhile, has been downgraded by S&P to BBB-, one notch above junk, because Ellison is in the middle of a $250 billion data-centre expansion that is burning cash faster than it can generate revenue. (Bloomberg, Evening Briefing Americas, 17 July 2026.) The same man whose family foundation is, in effect, lobbying to defeat a wealth tax is also the one most leveraged to the AI buildout. Ellison, like the leveraged ETF, is both the architect of the new system and its most exposed node. The figure of the cyborg billionaire — half flesh, half datacenter — is no longer a meme. It is the operating diagram of American capitalism in 2026.
The most disarming single piece of media I read this week was, of all things, a Monocle column by the editor-in-chief Andrew Tuck, titled “AI might Botox your writing but wrinkles are the lines that matter.” (Monocle, The Monocle Weekend Edition, 18 July 2026.) Tuck writes from a terrace in Palma, Mallorca, where every Friday morning he sits with a coffee and waits for the column to arrive in his “mental inbox.” He has noticed that the pitches he is receiving from writers he trusts have a slightly off quality — sentences that say something big but, on closer reading, are oddly hollow. He runs one through an AI checker; 70% generated. The fingerprints are everywhere. He then tells the story of a dinner with a famous magazine editor who now uses AI “to organise my initial thoughts, give me some prompts” — a Viagra for the blank page, he calls it — and ends with a small but heartfelt plea: “Don’t just reach for the AI Botox; show your writerly wrinkles.”
It is, in 2026, a radical position. The cultural pressure to use AI is now so total that the act of not using it has become a craft statement in itself. The same pressure is showing up everywhere, in every direction.
Look at the week’s other dispatches on the same theme. The Monocle Weekend Edition, on Saturday, leads with a piece on summer shopping; The Monocle Minute, on Friday, with an essay on white denim. (Monocle, 17–18 July 2026.) The Economist’s Today leads with a piece called “America should stop making it so hard to have fun” — a five-thousand-word polemic against the bureaucracy of U.S. nightlife, in which a 25-year-old needs more paperwork to open a bar than an 18-year-old does to buy a semi-automatic rifle. (The Economist, Today, 17 July 2026.) The Economist‘s own sub-feature this week: “Cities are rethinking what happens after dark.” Bloomberg’s CityLab Weekly runs a feature on New York’s “city of scaffolding” and its long-overdue glow-up. (Bloomberg, CityLab Weekly, 18 July 2026.) Monocle’s Thursday piece is on Vienna’s swimming culture, with its dozens of named lidos along the Alte Donau. (Monocle, The Monocle Minute, 16 July 2026.) The Friday Monocle piece is on Tokyo’s fan-jackets, the literal wearable air-conditioners that have become the surprise fashion hit of the Japanese summer. (Monocle, The Monocle Weekend Edition, 18 July 2026.)
The pattern is so obvious that it would be embarrassing to point out, except that I have to, because no one else is. The cultural signal of the week is the same in every publication: go outside. The Summer of Cinema, per Monocle, is urging audiences to break the spell of their screens, citing Toy Story 5‘s elegy for tablet-blinkered childhoods, Silo‘s dystopian vision of life underground, and Backrooms‘ doomscrolling-with-teeth as a triptych of cultural warning shots. (Monocle, The Monocle Minute, 16 July 2026.) The UK has just announced a ban on social media for under-16s. (Monocle, The Monocle Minute, 16 July 2026.) House Democrats in the U.S. have voted, this week, to strip $3.3 billion in annual aid to Israel; the Democratic coalition that has held for fifty years is shifting under their feet. (Newsweek, Perspective, 16 July 2026; FT, World News, 16 July 2026.) The Economist, separately, runs an essay from its Europe columnist on “the risky and dirty playgrounds” of Scandinavia, arguing that the world’s risk-averse culture is producing children who are too cautious to climb a tree. (The Economist, Today, 16 July 2026.)
The reason everyone is telling you to go outside is, of course, that it is, in many places, increasingly dangerous to do so. The orange sky. The 50°C summer in Spain. The air quality alert in Toronto. The three straight nights of 30°C heat in London that, in the FT’s reporting, have already killed more than 2,700 people in England and Wales this year. (FT, International morning headlines, 15 July 2026; Bloomberg, CityLab Weekly, 18 July 2026.) The fan-jackets of Tokyo, the swimming lidos of Vienna, the early-morning swims in the Danube that Monocle’s Francesca Gavin describes with such sensual specificity — these are all survival technologies, dressed up as lifestyle technologies. The FT HTSI, on Friday, runs a guide to “the world’s most breathtaking cinemas” — but it does so in the same week that the FT also runs, in its main edition, a piece on “the dull bit of climate change policy.” (FT, International morning headlines, 16 July 2026.) The “dull bit” is the bit about adaptation. The dull bit is the bit about the people who are going to die in this country, this decade, because we did not act.
Hong Kong’s nightlife has, meanwhile, quietly decamped from Lan Kwai Fong to a century-old dim sum hall, a heritage building in Stanley, a record store in Central, an industrial building in Kwun Tong. (Bloomberg, Hong Kong Edition, 16 July 2026.) The new scene, in the words of one of its young organisers, is “more experience-driven, more interested in music, crowd, space and community than in bottle service or status.” It is also, structurally, a response to the surveillance and QR-code-and-RAT-test infrastructure of the post-pandemic city. The young Hong Kongers going to dim sum raves are not just being hedonistic; they are rebuilding a civil society, one warehouse party at a time, in a city in which the public square is closing.
In other words: there is a worldwide, cross-demographic, cross-ideological movement, expressed this week in fashion columns, in cinema reviews, in policy decisions, in the migration of nightclub maps — to reclaim the outside. The cultural and political valence of the move is, depending on where you stand, hopeful or desperate. In Vienna, it is the rediscovery of a 19th-century civic tradition. In Hong Kong, it is an act of civil disobedience. In New York, in the orange haze, it is a brief, panicked opening of the window. The outside has been a luxury, a politics, a medicine, and a refusal. This week, it was all four.
Sunday, 19 July 2026, will see the men’s World Cup final in New Jersey: Argentina against Spain. The first is the defending champion, led by a 38-year-old Lionel Messi who, on Tuesday night, broke English hearts in the 85th minute with a pass that was, as one of the England defenders conceded afterwards, “the kind of thing you simply cannot legislate for.” (Newsweek, The Bulletin, 16 July 2026; FT, In Today’s FT, 16 July 2026.) The second is the team that the world’s AI models, before the tournament, picked to win it all — and which, to be fair, is also the team most observers have now picked to win it all, having comprehensively outclassed France in the other semi. (FT, In Today’s FT, 16 July 2026; Rest of World, The Global Dispatch, 17 July 2026.) The goldfish has, of course, already called it, and the goldfish picked Argentina.
It is the small details of the tournament, more than the matches themselves, that are doing the political work. The Argentine players, after their 2-1 semi-final win over England, held up a banner claiming the Falkland Islands. The UK has called for a FIFA investigation. (DW, Daily Bulletin, 16 July 2026; FT, In Today’s FT, 16 July 2026.) The Falklands, in 2026, are not a frozen conflict; they are a hot memory that flares up every time a Buenos Aires shirt appears on a global screen. The geopolitics of sport have always been the most legible geopolitics, because the jerseys are the simplest possible sign. The question, of course, is whether a 21st-century World Cup can still do the kind of political work that the 1986 final did, when Maradona’s “Hand of God” goal against England became a global meme long before the word meme existed. (The Economist, Today, 15 July 2026, on the Odyssey; FT, International morning headlines, 16 July 2026.) The answer, this week, is: maybe not the same work, but different work. The 2026 World Cup has, in fact, been a global commercial triumph, with 6 million tickets sold and 46 million Americans watching the U.S.–Belgium round-of-16 match — the most-watched football broadcast in U.S. history. (Bloomberg, Businessweek Daily, 17 July 2026.) It has also been, by FIFA’s own design, a deeply Americanised product: a halftime show with Justin Bieber and Shakira, championship rings for the winners, “dynamic pricing” that pushed some tickets past $29,000 for the trophy-lounge package. (Newsweek, The 1600, 17 July 2026; Bloomberg, Businessweek Daily, 17 July 2026.) The U.S. has, in 28 days, learned to love football on its own terms — which is to say, in a way that involves celebrities, sponsorships, and a sentimental attachment to teams that can be calibrated by TikTok following.
And yet the U.S. team itself lost in the round of 16, and the country has, in the words of one of the Businessweek reporters, “not even Messi” been able to deliver the breakthrough. The U.S. is the host that did not get to play on Sunday. The world came to the U.S. for the party, and the U.S. mostly watched, and that, too, is a kind of cultural statement about where the country is in 2026. The 1994 World Cup was the moment that the U.S. decided it wanted to host a global sport; the 2026 World Cup is the moment it discovered that wanting to host and being a participant are two different ambitions. The 2026 World Cup will be a financial and ratings success. It will not be, in the way that the 1986 and 1990 tournaments were, a defining cultural event for the host. The U.S. has, instead, been spending its summer arguing about the SAVE Act.
The other stories of the week do not so much interrupt the above as provide a long, slow bass note underneath it. The new Hungarian prime minister, Peter Magyar, is three months into his dismantling of Viktor Orbán’s 16-year machine; the Hungarian parliament this week voted to amend the constitution to remove Orbán-loyalist President Tamás Sulyok. (Bloomberg, Eastern Europe Edition, 17 July 2026; FT, FT Edit, 16 July 2026.) In Ukraine, President Zelensky has sacked his popular defence minister, Mykhailo Fedorov, the architect of the country’s drone warfare, and the country is in protest. (FT, In Today’s FT, 16 July 2026; DW, Daily Bulletin, 15 July 2026.) The new British prime minister, Andy Burnham, takes over on Monday from Keir Starmer, who is leaving with a “foundation laid” speech that is, in the FT’s view, “mixed”; his expected chancellor, Shabana Mahmood, is described in one profile as “the migration hardliner who refused to serve under Corbyn,” a phrase that means almost everything it needs to about the state of the British centre-left. (FT, In Today’s FT, 15 July 2026; FT, World News, 16 July 2026.) In the U.S., Mark Carney’s bridge deal with Washington is, in the Canadian press, “less clear what exactly was agreed to,” with both governments giving different accounts of the profit-sharing terms. (Bloomberg, Canada Daily, 18 July 2026.) The Pope is silent on a lot of things but has, in the past month, weighed in on AI. China’s Q2 GDP grew at 4.3%, the slowest in three decades outside the pandemic. (FT, International morning headlines, 15 July 2026; CNBC, The Tech Download, 15 July 2026.) The Congo’s Ebola outbreak is the fastest the WHO has ever managed. Africa is negotiating with the World Bank over a $400 million climate-finance loan guaranteed by the U.K. Bloomberg’s Next Africa opens with the lede: “Africa’s largest fund manager is in crisis. Again.” (Bloomberg, Next Africa, 15–17 July 2026.) The Davos-man vision of an integrated, prospering, climate-resilient global economy is, in 2026, a memory of a memory.
What is replacing it is not yet clear. The most interesting new international institution of the week is not an institution at all; it is a preference — the preference of 29 countries, including Russia, Pakistan, and Indonesia, to sign up for the China-led World Artificial Intelligence Cooperation Organization, and the preference of 24 (overlapping) countries, including the U.S., India, Japan, the U.A.E. and the Philippines, to sign up for Pax Silica. (Newsweek, Geoscape, 17 July 2026.) The new map of the world is not being drawn in the Security Council; it is being drawn in the data-centre permitting offices of provincial capitals. Whoever gets the chips, the power, and the cooling water, gets the rest.
The week, then, was a hinge. Not because of any single news story, but because of the interlocking character of the stories. The orange sky over Manhattan is connected to the burning strait of Hormuz, which is connected to the strategic decisions of Beijing, which are connected to the leveraged ETFs in Seoul, which are connected to the bankruptcies of small-cap miners in Idaho, which are connected to the wealth-tax ballot in California, which is connected to the immigration raids in Maine, which are connected to the broadcast of a presidential address that almost no major network would carry. The political crisis of 2026 is, at some level, a thermodynamic crisis. We are running out of the cheap energy and the cheap trust that the old system needed. The new system — the one being assembled, in fits and starts, by the trillion-dollar data centres and the open-weight models and the prediction markets and the Truth APIs — is not, yet, a system at all. It is an archipelago. The goldfish lives in a small tank. The algorithms live in the cloud. The climate is in the sky. The world is, in the meantime, somewhere in the middle, watching, and being told — by every columnist, every cinema, every wearable air-conditioner, every summer lido, every public-health official in New York — to go outside.
The goldfish, for what it is worth, has been outside this whole time. It is, after all, in a tank in Toronto, which is, this week, the city with the worst air quality on the continent. Its predictions, so far, are 80% accurate. The five large language models are between 50 and 60. The Federal Reserve is unable to make up its mind about whether to cut rates. The President is unable to make up his mind about whether the previous President won. The U.N. climate chief is unable to make up her mind about how to talk about the dimming of the Earth. The leveraged ETFs are unable to make up their minds about whether they are gambling or investing. The goldfish, in its little tank, just keeps swimming toward the flag. It does not write columns. It does not give speeches. It does not bet on its own future on Kalshi. It does, when pressed, tell you which team will win. Eighty per cent of the time, that is more than the rest of us have managed this week.
Perhaps that is what the algorithm is, in the end, for. Not to replace the goldfish. To remind us that we are, all of us, swimming in a small tank with a few flags at either end, and that the only honest question — the one we keep refusing to ask — is which side we are swimming toward, and why. The week of July 15–18, 2026, will not be remembered for any one of its events. It will be remembered, if it is remembered at all, for the moment we noticed, collectively, that the tank is getting smaller, the flags are getting more numerous, and the air outside, this summer, is on fire.
[Written, Researched, and Edited by Pablo Markin. Some parts of the text have been produced with the aid of Agent, Minimax, and, Qwen, Alibaba, tools (July 21, 2026). The newsletters were sourced from ARTNews, Artforum, The Atlantic, Bloomberg, CNBC, Deutsche Welle, The Economist, The Financial Times, Le Monde, Monocle, The New York Times, Newsweek, Nikkei Asia, Noema Magazine, El País, Rest of World, Radio Free Europe/Radio Liberty, Semafor, The South China Morning Post, The Sydney Morning Herald, and The Wall Street Journal.]
Laurent Warlouzet. Liberty, Solidarity and Community: Capitalism and European Integration, 1945 to the Present. Cambridge: Cambridge University Press, 2026. ISBN 978-1-009-68263-3 (hardback), 978-1-009-68264-0 (e-book), 390 pp. Open Access on Cambridge Core, Hardcover $115.00, Paperback/eBook $35.00.
This is the English translation, revised and expanded through 2025, of the French original Europe contre Europe. Entre liberté, solidarité et puissance depuis 1945 (Paris: CNRS Éditions, 2022, 496 pp., ISBN 978-2-271-13846-0). The new edition incorporates material on the war in Ukraine, the environmental backlash, and “Trump 2.0” (Warlouzet, “My book,” Academia.edu). The English text is also published in Open Access.
Laurent Warlouzet is Professor of Contemporary History at Sorbonne Université, affiliated with the Centre d’histoire de Sciences Po and a specialist in the history of European integration, European political economy, and the varieties of capitalism (Wikipedia, “Laurent Warlouzet”; Sorbonne CV, July 2022). His earlier English-language monograph, Governing Europe in a Globalizing World: Neoliberalism and Its Alternatives (Palgrave, 2018), covered 1973–1986 and already prefigured the present book’s interpretive grid. The argument that “capitalism is not just a contest between free-marketeers and their opponents” but involves a third pole — a community or neomercantilist logic — was first developed in the French edition and is now the explicit organising principle of the 2026 English volume (Warlouzet, “My book,” Academia.edu).
The book is dedicated, in Warlouzet’s own words, to making sense of “the intensity and diversity of European economic cooperation from 1945 to the present” by reading it through three ideal-typical logics of capitalist governance (Warlouzet, “Europe against Europe,” iaces.ie).
Warlouzet’s thesis is a sustained polemic against the conventional two-sided reading of post-war European capitalism. Most of the literature on European integration pits liberalism (the Single Market, competition policy, free trade) against social Europe (welfare, redistribution, environmental and labour regulation). Warlouzet insists that this dichotomy is incomplete: there has always been a third, structurally distinct camp that he calls community capitalism — protectionism, industrial policy, defence of the European “group” against external economic and military powers (Warlouzet, Liberty, Solidarity and Community, Introduction, Cambridge Core).
The trinity of liberty, solidarity, and community serves as an ideal-typical framework:
Liberty capitalism aims to “free the market to unleash growth” — pure and perfect competition, trade liberalisation, competition policy, the single currency as a disciplining device.
Solidarity capitalism “reins in the free market to protect the weak and the environment” — welfare provisions, gender equality, regional transfers, social and environmental regulation.
Community capitalism “safeguards the group through protectionism and military might” — industrial policy, defence of European champions against US/Chinese competition, the “Brussels Effect,” and (since 2016) a renewed rhetoric of European strategic autonomy (Introduction; Chapter 1, “The Trinity of Capitalist Governance,” Cambridge Core).
Three claims, advanced in the introduction, animate the whole volume:
1.All societies seek to balance the three types of capitalist governance; none is ever purely liberal, social, or neomercantilist.
2.The European Union is the ideal case study for understanding how polities negotiate compromise between the three.
3.The management of community capitalism is the most pressing contemporary challenge, made acute by the return of protectionism, the war in Ukraine, and the Trump administration’s transatlantic tariffs (Introduction; Chapter 10).
The volume is organised in four parts plus an introduction and a synthetic conclusion. The chronology deliberately crosses the conventional periodisation of European integration, which is one of the book’s most thought-provoking methodological choices (Introduction, “Book Outline,” Cambridge Core).
Introduction
Part One — Concepts and Frameworks
1.The Trinity of Capitalist Governance
1.The European Union as a Political Hybrid
Part Two — The Threefold Model (1957–1992)
3. A Regulated Market at the Core
1.Solidarity: A European Welfare State Flanking the Single Market
1.A Community without Communitarianism: Europe’s Failure as a Military and Industrial Powerhouse
Part Three — Between Crisis and Transformation (1970–1992)
6. European Attempts to Promote Alternatives to Neoliberal Globalisation (1970–92)
1.Common Currency and Neoliberal Turn? (1970–92)
Part Four — The High Neoliberal Era and Beyond (1992–Present)
8. The European Union as a Driver of “High Neoliberalism” (1992–2015)
1.Solidarity: Expanded and Contested Social and Environmental Action
1.The Resurgence of the Community Approach in the Twenty-First Century
Conclusion: Chronology, Alternatives, and Current Challenges
Back Matter — Notes, Primary Sources (Archives and Interviews), Acknowledgements, Index.
The empirical base is striking: 22 archival fonds across 8 countries, supplemented by an extensive bibliography, primary interviews, and “grey literature” (L’Histoire, June 2022, quoted on the CNRS Éditions page). The 2026 English edition also extends coverage through 2025, including the COVID-19 recovery instrument (NextGenerationEU), the energy and climate packages, the war in Ukraine, and the second Trump administration.
Warlouzet announces the trinity, justifies the long period (1945–2025), and warns against the “stereotypical dichotomy” between pro-market liberals and anti-market egalitarians. He situates the book in three literatures that “rarely enter into dialogue” — European integration studies, comparative political economy (varieties of capitalism), and international political economy — and proposes to combine them (Introduction, Cambridge Core).
The conceptual foundation. Warlouzet defines each ideal type through a table (Table 1.1) mapping policies, instruments, and actors to one of the three principles. The category of community draws explicitly on Max Weber, Albert Hirschman, and Elinor Ostrom — the latter two being the more surprising reference points and a useful reminder that “community” here is not the communitarianism of the 1990s but a structural category of group-defence (Chapter 1, Cambridge Core).
A short, dense institutional history arguing that the EU’s distinctive features — its multilevel, hybrid, supranational-intergovernmental character — are not an accident but a functional adaptation to the need to combine the three logics. This chapter is essential for the rest of the book: it explains why the EU, uniquely, can be protectionist, social-democratic, and liberal all at once.
Traces the making of the Common Market and the 1992 Single Market. Warlouzet underlines a paradox familiar to specialists: the Treaty of Rome (1957) was “based on the constitution of a large integrated ‘common market’ compatible with national policies, and based on the principles of solidarity and community” (Chapter 3, Cambridge Core). The “single market” project of 1985–1992 is read as a liberty project tempered by flanking social and regional measures, foreshadowing the Delors era.
The European welfare state is “embryonic” rather than federal: it emerged slowly between 1945 and 1985, peaked under Delors, and consisted of three layers — protection of the weak (labour law, gender equality, later environment), redistribution (regional funds), and macro-economic coordination. Two alternative paths were pursued and ultimately abandoned: indicative planning à la Monnet, and comprehensive social and fiscal harmonisation (Chapter 4, Cambridge Core).
The most counter-intuitive chapter. Warlouzet shows that, despite a long rhetorical tradition of “European power” (De Gaulle, early Commission documents, the Fouchet Plan), the European project consistently failed to become a military or industrial powerhouse. The Community became a civilian normative power, with the “Brussels Effect” rather than industrial policy or hard power as its main instrument. This chapter is essential to the argument: the EU’s long community deficit is precisely what makes the post-2016 resurgence in Chapter 10 historically significant.
The book is at its most original here, documenting projects that were seriously considered and then dropped: the 1970s plans for a European company law harmonised across the continent, the 1980s project of a European industrial policy (Bureau de Bruxelles, ESPRIT), environmental regulation, gender equality directives, the “social clause” debate, and the Multiple-Action Programme for worker consultation. The point is to recover “the alternatives that were – and still are – present” (Introduction; Conclusion).
A careful deconstruction of the standard narrative that frames the euro as a straightforward neoliberal project. Warlouzet shows that monetary union was conceived well before the neoliberal turn and had multiple justifications, some of them closer to solidarity (intra-European transfers, symmetrisation of shocks) or community (European monetary power on the world stage). The Delors Committee combined all three logics — a point that complicates the Anglo-American “euro = ordoliberal straitjacket” reading (Chapter 7, Cambridge Core).
The 1992 Maastricht settlement, the 2000 Lisbon Strategy, the competition-policy activism of the Monti and Kroes era, the Bolkestein Directive, the 2010–12 eurozone crisis management and the “new economic governance” (Six-Pack, Two-Pack, Fiscal Compact) are read as a coherent liberty phase. The German ordoliberal imprint, the role of France under Sarkozy, and the ECB’s turn under Draghi are reassessed against the trinity.
Despite neoliberalism, solidarity did not disappear. Warlouzet documents three vectors: (a) the legal regulation of globalisation through social and trade legislation, (b) financial redistribution (structural funds, the COVID-19 recovery instrument), and (c) environmental regulation, including climate policy. The Delors era is the high point; the Barroso and early Juncker eras mark a relative ebb; the von der Leyen Commission and the European Green Deal mark a partial return, contested by the right (Chapter 9, Cambridge Core).
“While the late twentieth century was characterised by the growing prominence of liberty capitalism, the second half of the 2010s has witnessed a resurgence of community capitalism” (Chapter 10, Cambridge Core). The chapter traces this turn through Brexit, the Trump challenge, the COVID-19 crisis, the war in Ukraine, the Inflation Reduction Act, and the rise of far-right parties — including the strong showing in the 2024 European elections. The tension between a nationalistic form of community capitalism (often far-right) and a European-level community capitalism (industrial policy, strategic autonomy) is identified as the central political question of the coming decade.
Warlouzet proposes a four-phase chronology of post-war European capitalism:
1.Embedded liberalism (1945–1973): Keynes at home, Smith abroad.
2.Global attempts at mixed capitalism (1973–1992).
3.High neoliberalism (1992–2015).
4.The return of community capitalism (2016–present).
Two further conclusions follow. First, although imperfect and cumbersome, the European system of compromise “has nonetheless provided a degree of choice for Europeans”: “Far from the image of a neoliberal technocratic dictatorship, the European Union can be protectionist and/or socio-environmentalist if Europeans want it.” Second, Germany was influential but not dominant: the historiography that attributes European integration to German ordoliberal hegemony is, on Warlouzet’s reading, an over-simplification (Conclusion, Cambridge Core).
A major contribution is the genealogy of the three principles. Warlouzet roots the trinity in:
Liberty: the Ordoliberal/Friedmanite tradition, the 1986 Single European Act, the Washington Consensus.
Solidarity: Beveridge, the Delors Commission’s “social flank” rhetoric, the European Pillar of Social Rights, Catholic social teaching in CDU/CSU, the Nordic tradition.
Community: a Weberian and Hirschmanian notion of group defence, the French Gaullist tradition of “puissance,” the Italian IRI/state-holding tradition, the Bundesbank’s defence of the Deutsche Mark, the recent literature on strategic autonomy and the “Brussels Effect” (Anu Bradford).
By making community an analytic category of equal weight to liberty and solidarity, Warlouzet recovers a dimension of European integration that Anglo-American scholarship has tended to relegate to “the intergovernmental” or to nation-state “preferences” (Warlouzet, Governing Europe in a Globalizing World, 2018).
Andrew Moravcsik (Princeton, author of The Choice for Europe), in a back-cover blurb, calls it “a magnum opus” that “sweeps across 75 years of integration history right up to current crises” and credits Warlouzet with showing that “the EU’s unique mix of policies” reflects “a specific balance that European capitalism strikes between liberty, equality and community” (Cambridge Core, book page). In Foreign Affairs, Moravcsik is more critical, writing: “Although Warlouzet is sometimes tempted to exaggerate the range of potential choices governments faced, in the end, his book proposes some clear answers” (Wikipedia, “Laurent Warlouzet,” citing Foreign Affairs).
The Spanish-language journal Historia y Relaciones Internacionales (madrimasd.org) emphasises the methodological contribution: Warlouzet recovers the projects that were seriously considered and abandoned, allowing us to see “as many possible futures as Europeans could seize” (”tres proyectos en pugna,” 8 March 2026).
The French academic journal Histoire Politique (Carine Germond) endorses Warlouzet’s “typologie ternaire” as “un prisme à une étude magistrale” (Germond, Histoire Politique, 4 November 2022).
The French edition of 2022 produced a remarkably broad and positive critical response:
Marc Semo, Le Monde, 7 January 2022: “L’Europe est aujourd’hui à un tournant. Doit-elle aller vers plus d’intégration, voire de fédéralisme ? Elle a discrètement fait des pas significatifs dans cette voie depuis 2008, et plus encore pour faire face à la pandémie de Covid-19. Mais les opinions publiques restent majoritairement réticentes. Le flou et l’ambiguïté créatrice restent donc de mise.”
Johann Chapoutot, L’Histoire, June 2022: “Ouvrage de recherche, à la pointe de l’art, ce livre est également la référence, bientôt le classique, appelé à rendre l’Europe moins difficile.”
Sébastien Maillard, Études, May 2022: “Les visions libérale, sociale et néomercantile forment ainsi un triptyque à l’aide duquel l’auteur démontre comment la construction européenne repose, à des degrés variables selon les moments, sur ‘la concurrence qui stimule, la coopération qui renforce et la solidarité qui unit.’”
Maxime Lefebvre, Politique étrangère, Spring 2023: “C’est une somme encyclopédique sur l’histoire de la construction européenne.” Lefebvre nevertheless recommends the book as a “contribution remarquable, précise, utile et sérieuse.”
Samuel B. H. Faure, La vie des idées, 2024: notes that the volume constitutes “an insight into and an interpretation of the history of European integration from a political and social perspective” while also being a “policy manifesto for the political and economic Europe to deal with future challenges and crises.”
Geoffrey Maréchal, La Cliothèque, 27 June 2022: “Dans un ouvrage d’une très grande densité est complexe, Warlouzet nous offre une relecture de l’histoire de l’Union Européenne à travers le prisme des politiques sociales et économiques qui ont pu animer ces six dernières décennies. L’ouvrage constitue à n’en point douter un titre marquant sur ces questions.”
Ines Soldwisch, Francia-Recensio (Heidelberg): “Das anzuzeigende Buch von Laurent Warlouzet ist sowohl ein Einblick in und eine Interpretation der Geschichte der europäischen Integration aus politischer und sozialer Perspektive, gleichzeitig aber auch eine Programmschrift.”
Giada Lagana, LSE Review of Books, 10 March 2022: praises the depth of historical resources and a style that is “brilliantly written,” describing it as a book for “any readers wanting to understand the evolution of economic and social policies in and beyond Europe.”
Le Grand Continent, 1 April 2026 (in connection with the English edition): publishes an essay by Warlouzet himself on the question “Is Trump transforming European capitalism?” citing the new book as a key analytical reference (legrandcontinent.eu).
Three criticisms recur, useful for a balanced appraisal.
1.Ideal types and their limits. The trinity is heuristically powerful but can flatten the heterogeneity of national trajectories. As the French reviewer from La Cliothèque notes, the book is “d’une très grande densité est complexe”; the compression into three logics sometimes blurs the role of domestic political institutions and party systems.
2.Choice vs. constraint. Moravcsik’s Foreign Affairs caveat is the most theoretically pointed: the “range of potential choices” available to European governments may be smaller than Warlouzet’s archival recovery of abandoned projects suggests. The “alternatives” were often politically unthinkable rather than merely dropped.
3.The “community” cluster. Community capitalism is the most heterogeneous of the three. It encompasses protectionism, industrial policy, the “Brussels Effect,” military power, the resurgent far-right’s nativist protectionism, and French Gaullist puissance. Some readers may wish the chapter had spent more time disaggregating these, especially given the very different normative valences of European-level community (Brussels effect) versus national-level community (Le Pen, Meloni, AfD).
Warlouzet positions the book at the intersection of three bodies of work that he says rarely communicate:
European integration theory (Haas, Moravcsik, Sandholtz, Stone Sweet).
Varieties of capitalism (Hall and Soskice, Amable, Ebbinghaus and Manow).
International political economy (Helleiner, Hirst and Thompson, Rodrik).
The explicit dialogue with the varieties of capitalism literature is the most productive. Where Hall and Soskice distinguish liberal market economies (LME) from coordinated market economies (CME) at the national level, Warlouzet shifts the unit of analysis to the governance principle (market, solidarity, community) and tracks how the European level has institutionalised all three simultaneously. This moves the debate beyond the “Europe as a coordinated market economy” thesis (the “capitalist diversity” of Amable) and beyond the liberal-intergovernmentalism of Moravcsik, while remaining in dialogue with both.
The book is also in productive tension with the new institutionalist literature (Stone Sweet, Fligstein) and with the historical sociology of the Eurozone crisis (Matthijs, Blyth). Warlouzet’s trinity offers a more fine-grained periodisation of the 2008–2015 crisis than the conventional “austerity vs. stimulus” framing, and it captures the post-2016 turn (Brexit, Trump, COVID-19, Ukraine) that much of the older literature has not yet absorbed.
Compared to his earlier English-language monograph Governing Europe in a Globalizing World (2018), this volume is both broader in time (1945–2025 rather than 1973–1986) and more theoretically ambitious: the trinity is now fully theorised, not just used as a heuristic.
Warlouzet explicitly proposes the book as a “programme” for thinking about Europe’s present (Soldwisch, Francia-Recensio). Five current debates are illuminated by the trinity:
1.The Inflation Reduction Act and the Net-Zero Industry Act: a return of community capitalism at European level, contested by liberty purists and conditioned by solidarity concerns (cohesion, just transition).
2.The NextGenerationEU recovery instrument: an unprecedented solidarity innovation that nonetheless has a strong community logic (strategic autonomy).
3.Trade policy after Trump 2.0: tariffs and the weaponisation of economic interdependence force the EU to choose between a liberty response (retaliation within the WTO) and a community response (industrial policy, protection of European champions).
4.Defence and rearmament: a clear community turn after decades of failure (Chapter 5), now driving “Rearm Europe” initiatives.
5.The 2024 European elections and the rise of the far right: Warlouzet notes that “most far-right parties promote a nationalistic form of community-based capitalism, with restrictive immigration in particular, combined with opposition to a solidarity-based approach at the European level” (Chapter 10, Cambridge Core). The book provides a vocabulary to distinguish this nationalistic community from a European-level community response.
Warlouzet’s own contribution to Le Grand Continent (1 April 2026) argues, in the wake of Trump 2.0, that European capitalism is being forced to choose between a defensive liberty and an assertive community, with solidarity caught in between. The book gives the reader the conceptual tools to follow the argument as it unfolds.
Liberty, Solidarity and Community is a major contribution to the historiography of European integration. Its strengths are three: (i) an original conceptual framework that escapes the liberal-vs-social binary; (ii) an empirical base of unusual breadth, drawing on 22 archival fonds across 8 countries; and (iii) a synthetic conclusion that proposes a four-phase chronology of post-war European capitalism and identifies the post-2016 turn as the central political question of our time.
The book is dense, conceptually demanding, and at points the community category is asked to do a lot of work. It is also occasionally vulnerable to the objection that the “alternatives” it recovers were politically more constrained than their archival salience might suggest. But these are the prices of a genuinely synthetic history, and on balance the book delivers the most ambitious long-run reinterpretation of European capitalism in the post-1945 period since Alan Milward’s The Reconstruction of Western Europe (1984) and Andrew Moravcsik’s The Choice for Europe (1998).
For students of European integration, varieties of capitalism, and international political economy, the book is essential reading. The fact that it is now available as open access on Cambridge Core is itself a small solidarity gesture in line with the book’s argument.
Cambridge University Press, Liberty, Solidarity and Community (book page, chapters, introduction, conclusion, index): https://www.cambridge.org/core/books/liberty-solidarity-and-community/355531CFEE184AFF8B706B24D25EC368
Excerpt (Introduction): https://assets.cambridge.org/97810096/82640/excerpt/9781009682640_excerpt.pdf
CNRS Éditions, Europe contre Europe (press reviews and blurb): https://www.cnrseditions.fr/catalogue/histoire/europe-contre-europe/
Carine Germond, review in Histoire Politique, 4 November 2022: https://journals.openedition.org/histoirepolitique/7442
Giada Lagana, LSE Review of Books, 10 March 2022: https://eprints.lse.ac.uk/114773/
Maxime Lefebvre, Politique étrangère, Spring 2023: https://politique-etrangere.com/2023/05/19/europe-contre-europe/
Sébastien Maillard, Études, May 2022: https://www.revue-etudes.com/critiques-de-livres/europe-contre-europe-de-laurent-warlouzet/24352
Ines Soldwisch, Francia-Recensio: https://journals.ub.uni-heidelberg.de/index.php/frrec/article/download/94499/89529
Samuel B. H. Faure, La vie des idées, 2024 (cited via samuelbhfaure.com/book-reviews)
“tres proyectos en pugna en la historia de la integración europea,” Historia y Relaciones Internacionales (madrimasd.org), 8 March 2026: https://www.madrimasd.org/blogs/Historia_RRII/2026/03/08/131584
“Trump est-il en train de transformer le capitalisme européen,” Le Grand Continent, 1 April 2026: https://legrandcontinent.eu/fr/2026/04/01/trump-est-il-en-train-de-transformer-le-capitalisme-europeen/
Laurent Warlouzet, “Europe against Europe: making sense of the intensity and diversity of European economic cooperation,” IACES: https://www.iaces.ie/post/europe-against-europe-making-sense-of-the-intensity-and-diversity-of-european-economic-cooperation
Wikipedia, “Laurent Warlouzet” (English and French): https://en.wikipedia.org/wiki/Laurent_Warlouzet, https://fr.wikipedia.org/wiki/Laurent_Warlouzet
Warlouzet, “My book…”, Academia.edu: https://sorbonne-universite.academia.edu/LaurentWarlouzet
YouTube lectures: “From a market-oriented Europe to a global Euro-power?” (Centre Européen); “Europe vs. Europe: Social, Neoliberal and Industrial Europe” (Warlouzet lecture).
Reviewer’s note: All quotations in §2–§4 are taken from the open-access Cambridge Core edition; the French reviews of §6 are from the 2022 CNRS original. Page references to the French original (496 pp.) and the English edition (c. 390 pp.) are given where available.
[Written, Researched, and Edited by Pablo Markin. Some parts of the text have been produced with the aid of Agent, Minimax, tools (July 21, 2026). The featured image has been generated in Canva (July 21, 2026).]
OpenEdition suggests that you cite this post as follows:
Pablo Markin (July 20, 2026). Book Review: Liberty, Solidarity and Community: Capitalism and European Integration, 1945 to the Present. Open Economics Blog.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.