2026-08-25 ATS briefing

The AI Curse

Ref: A1 Item-ID: article-a5278bd168 Source: Project Syndicate

Prince William County's doubled electricity bill is the price of the county's own obsolescence, transferred back to its residents as a line item. The windowless building humming beside the town consumes more power than the town itself, and it needs nothing from the people who live there. Holmes reaches for the resource-state analogy, and it holds. A citizen's leverage over the state has always been the capacity to withhold something the state wants: labor, taxes, sons, votes. Each sanction presupposes a counterparty with an appetite. The data center has no appetite for the county's residents. It needs their grid capacity, their land, their water, but not their work, not their sons, not their votes.

The dispute will be conducted in the language of rate structures and transmission corridors, and that language will obscure the actual question of whether the county's people retain any bargaining position at all. The AI infrastructure competes with them for the resources they do control, and the family's doubled bill is the material form of their diminished standing.

What distinguishes this from earlier rounds of automation is the scale of fixed capital involved. A factory that replaces workers still needs workers to build the next factory. The data center's demand is for electricity and cooling, not for the human beings who live beside it. The county's residents have become what the resource-state's population always was: an externality to be managed rather than a constituency to be served. Where there is no need, there is no democracy, and the only remaining political fact is the citizens' need for protection they have no leverage to claim.

Where Will the Global Bond Rout Hit Growth and Equities?

Ref: A2 Item-ID: article-efa824d285 Source: Project Syndicate

Roubini's framing rests on a distinction between two kinds of bond sell-off. In the US, he argues, higher yields reflect something real: the end of the post-2008 stagnation, productivity gains from tech investment, and stronger private credit demand. That is a supply-side story where the economy can grow into higher rates. Japan, Germany, the UK and France have no such luck. Their yields are rising for fiscal reasons, with stagnating productivity and ageing populations, so the same interest-rate move does different damage.

The analytical hinge is whether higher yields are a symptom of growth or a cause of distress. Roubini's answer is that the US can absorb the shock because the underlying economy is generating enough surplus to service the debt. The others cannot, which means their governments face a choice between austerity and monetisation, with the political fallout that implies. For equities, the implication is that US stocks may be resilient while European and Japanese markets carry the adjustment.

What is missing is any account of where the extra US demand for credit comes from. If it is private-sector borrowing to finance AI infrastructure, then the yields are pricing in an investment boom that may or may not pay off. Roubini treats the tech investment as a given productivity gain, but it is also a speculative bet on future returns, and the bond market is the place where that bet gets tested. The rout in the weaker economies is the more straightforward story: fiscal crisis dressed up as a rates shock. The US case is the one where the optimism could prove circular, with high yields justified by growth that depends on the very borrowing those yields are meant to price.

Maximum Pressure on Iran Will Fail

Ref: A3 Item-ID: article-a38fb391d5 Source: Foreign Affairs

Six months into Trump's war on Iran, the administration's economic squeeze is collapsing under its own weight. Swanson, who served on the negotiating team before the strikes began, argues that the pressure campaign has exhausted the very credibility it depends on: Washington can no longer threaten military escalation credibly after six months of strikes failed to break the regime, and its sanctions now read as desperation rather than strategy.

The piece's central observation is that Iran's regime has always been both ideological and transactional, and that American policy has never known how to handle that duality. Washington demanded Iran choose between being a nation or a cause, in Kissinger's framing, while Tehran simply refused the choice, alternating between revolutionary posturing and pragmatic accommodation as circumstances required. The 2015 nuclear deal and the 2023 Saudi normalisation both came from the transactional side; the proxy network and anti-American posture from the ideological one. Trump's war assumed that crushing the ideological side would force Iran into rational nationhood. Instead, the bombing strengthened the regime's internal standing while weakening its military, and the economic war has taught Tehran that it can impose reciprocal costs through the Strait of Hormuz.

What makes this more than another establishment lament is the admission buried in it: the war has so degraded American leverage that no future administration can return to the old playbook. That is the real strategic rupture. The United States has spent 47 years cycling between military threats and economic coercion, and this war has demonstrated the limits of both simultaneously. Iran has learned it can survive American punishment; Washington has learned it cannot escalate further without unacceptable costs. Neither side can return to the status quo ante, but neither has achieved its objectives either.

For all its Atlantic Council credentials, the piece inadvertently sketches the contours of a genuine strategic impasse. The regime is internally combustible, as January's protests showed, but externally resilient. The United States retains overwhelming military superiority but no usable political strategy. That combination tends to produce either reckless escalation or grudging accommodation, and Swanson's own account suggests the latter is the only path left open.

Iran says it is ready for what US calls 'greatest financial offensive ever'

Ref: A4 Item-ID: article-e42ebe3e20 Source: BBC News

Bessent's "Operation Economic Outcast" arrives months after the US naval blockade already cut off most Iranian oil exports, which makes the Treasury Secretary's "ending" the threat rhetoric sound less like strategy and more like a script for domestic consumption. The sectoral determinations against digital assets, gold, aviation and shipping, plus nearly sixty sanctioned entities, target the financial plumbing that moves whatever revenue still flows. But the explicit threat to isolate any nation trading with Iran is where the plan collides with reality.

Ninety per cent of Iran's oil goes to China, which has never recognised US sanctions and has already said it will protect its interests. Beijing's response is the quiet centre of this story. The US is trying to sever the financial arteries of a major supplier to the world's second-largest economy while simultaneously propping up demand for its own government debt through bond market buybacks, an intervention that produced a one-day dip in long-term rates before they bounced back. A financial offensive of this scale depends on the dollar's dominance holding, yet the US is having to shore up its own borrowing costs at the same moment.

Iran's economy minister speaks of a two-year plan and "tools" to play the game, which is either bravado or a realistic reading of how sanctions enforcement actually works: it relies on third-party compliance. Capital Economics calls the direct impact a "damp squib", and the logic holds. You cannot sanction a country into submission when its principal customer is a rival power with its own reasons to keep buying. The Strait of Hormuz stays blocked, Brent sits at $92, and American voters face $4 gasoline ahead of the mid-terms. The "greatest financial offensive ever" is aimed as much at November as at Tehran.

The Age of Gated Recessions

Ref: A5 Item-ID: article-f5f75c9eb3 Source: Project Syndicate

Basu catches the Bureau of Labor Statistics in a moment of apparent self-contradiction: 23,000 jobs lost in July, unemployment down to 4.1%. The resolution is less mysterious than the headline suggests. A shrinking labour force pulls the rate down even as payrolls contract, and that shrinkage quietly redistributes the meaning of the numbers. The aggregate is not wrong so much as it is useless for anyone not already at the top of the distribution.

The piece's argument is that the US economy now operates as a set of gated communities. Some regions and workers enjoy boom conditions, rapid income growth, tight labour markets, while others face recession-like deterioration in their own streets. AI is the accelerant, concentrating gains in places and professions that can deploy it while hollowing out the work that cannot. Basu's framing captures something the standard business-cycle vocabulary misses: the economy no longer rises and falls as a single unit, so the policy levers that assume it does are firing blind.

This is a political problem dressed as an economic one. If the pain is localised and the boom is localised, the people experiencing recession are a minority, and the people experiencing boom have no reason to support transfers or stimulus aimed at places they do not live. The electoral arithmetic of redistribution collapses at the moment it is most needed. Basu does not say this, but the implication is that the working class is being fragmented by geography and sector into constituencies that no longer recognise each other as sharing a condition. That is the real gate, and it is not built of data.

Colombia’s new president Abelardo de la Espriella announces immigration crackdown

Ref: A6 Item-ID: article-2ea9e767c6 Source: The Guardian

De la Espriella announced this crackdown from Barranquilla, the city that has become his alternative capital while Bogotá deals with earthquake recovery. The timing matters. With 300 dead and 30,000 homes destroyed, the state's administrative capacity is stretched thin, and the new president is choosing to direct it at the 848,000 Venezuelans living irregularly rather than at reconstruction. That is a political choice dressed as law enforcement.

The legal basis is shaky. Colombian law has no category of "illegal" immigrant, only "irregular" status, which is not a crime. The migration authority reportedly lacks the staff to carry out the promised operations. So the announcement functions less as an operational plan and more as a signal to Trump, whose "Americans first" framing de la Espriella has borrowed wholesale, and to the regional bloc of Milei and Kast. The substance is secondary to the alignment.

What the article does not dwell on is the economic logic underneath the xenophobia. Colombia hosted roughly 2.8 million Venezuelans through a decade of bipartisan openness, absorbing them into informal labour markets that needed the supply. The crackdown targets the irregular majority, not the documented workers who contribute taxes and consumption. This is the classic move of a state that wants to keep migrant labour while appearing to expel it: the rhetoric of deportation, the reality of selective enforcement. Whether the operations actually materialise at scale is almost beside the point. The threat itself disciplines the migrant population, pushing them further into the shadows where wages are lower and protections weaker.

The opposition's framing of "aporophobia" captures the moral dimension, but the material one is that Colombia's ruling class has found a convenient scapegoat for the earthquake's failures. The gangs de la Espriella has declared "terrorists" include illegal mining operations that have deep ties to local economies and, in some cases, to state actors. Declaring war on them by "reason or by force" is a promise he cannot keep without tearing apart the clientelist networks that sustain Colombian politics. The immigration crackdown is the achievable half of the platform, and it targets the population least able to fight back.

Beyond China's humanoid robots, a quieter machine revolution is unfolding

Ref: A7 Item-ID: article-5bf5bccc5b Source: BBC News

Chen Tianyu tells his students that if they are not needed by the world of AI and robotics, they will be eliminated. He is not being cruel; he is describing the logic of a state project that has decided its demographic future. With more than two million robots already in Chinese factories and over half of the world's industrial robots now made in China, Beijing is automating not to cut costs but to survive a looming labour shortage. By 2035, over a third of the population will be over 60, and the country is set to lose nearly 60 million people in the next decade. The machines are coming because the workers are not.

The quiet revolution is visible in places like CRP Technology in Chengdu, where a prototype robot can only scan QR codes and perform simple repetitive movements. Pang Kai, one of the engineers, hopes it will think by itself within a decade because "maybe in 10 years, we won't have enough workers in China." The target is not to replace labour but to substitute for its absence, a distinction that matters for the 120 million people still in manufacturing. If the transition moves too quickly, job losses could be catastrophic; if it moves too slowly, the economy stalls. Cao Li at Leapmotor admits that 100% automation in stamping, welding and painting is already achieved, and that full replacement of the final-check workers is "entirely possible" and "may come relatively quickly."

The state's $20bn plan is a wager that vocational schools can churn out engineers faster than factories shed workers. The US leads in the "brain" of robotics, the AI programming, while China dominates the "body" through its manufacturing ecosystem. Shenzhen can source robot components in under an hour; Europe takes a week. The open-source approach has let Chinese AI firms catch up despite US export controls on advanced chips. The next leg of this race will be determined not by who builds the most impressive humanoid, but by which system can absorb the social cost of its own success.

Syria’s tanker trade: what the data shows

Ref: A8 Item-ID: article-34bf15d8f3 Source: Hellenic Shipping News

The tanker data from Banias, Tartous and Latakia shows a state rebuilding its energy trade from a standing start, but the shape of that rebuild matters more than the headline 354% jump in imports. Syria went from importing seventeen barrels for every one exported to a near 1.1:1 ratio, with crude and fuel oil exports that did not exist in 2025 now accounting for four fifths of outbound volume. That is a country liquidating its most readily available asset rather than resuming normal commerce.

The clean products deficit of 3.95m barrels against a stated domestic demand of 120,000 to 150,000 b/d is roughly what you would expect from a refining system running at partial capacity. The export side is the anomaly. Murmansk alone accounts for 4.68m bbl of inbound crude, and Russian load ports collectively supply 47% of arrivals. The crude goes in, gets processed or simply trans-shipped, and fuel oil comes back out. Banias is the pivot: 96% of wet imports and 74% of exports pass through it.

The timing clusters around the April reopening of the Al-Tanf crossing, though the first export cargo left in March, so the causal story is messier than the border reopening alone. What the data cannot show is who is paying whom, and in what currency. Russian crude arriving at a Syrian port and leaving as fuel oil looks less like market exchange and more like a state-to-state arrangement where the price is political. The 1.1:1 ratio flatters Syria with the appearance of balance; the composition of that balance suggests dependency dressed up as recovery.