2026-08-14 Observatory briefing¶

The Next Global Economic Crisis Could Be Made in China¶

The argument rests on a familiar sleight of hand: China’s overcapacity is presented as a policy error, a failure of will in Beijing, when the arithmetic the author himself lays out points elsewhere. A system that channels state-directed credit into sectors regardless of profit, that props up a third of its industrial firms operating at a loss, and that compels local governments to defend employment and tax revenue at any cost is not a system that can simply choose to rebalance. It is a machine whose political legitimacy is fused to its output. The anti-involution campaign and the 15th Five-Year Plan’s consumption rhetoric are gestures toward an exit that the structure forecloses.

Froman’s own numbers betray the voluntarist framing. Thirty percent of global industrial production, with a projected 45 percent by 2030, is not a trade surplus; it is a concentration of productive capacity that has outgrown the planet’s ability to absorb it. The comparison to the post-war United States is telling but incomplete. Washington’s dominance was underwritten by a global dollar system that let it export capital and import goods indefinitely. Beijing has no equivalent external circuit. Its surplus is a one-way valve, and the political backlash Froman documents — the shrinking appetite for deindustrialisation in the US, Europe, and the Global South — is the valve closing.

The real contradiction is not between China and its trading partners but within the Chinese model itself. The state must keep the industrial machine running to preserve social stability, but running it requires export markets that are politically exhausted. The author’s proposed solution — a preemptive, gradual rebalancing toward domestic consumption — assumes the state can decouple its survival from its productive apparatus. Yet the entire structure of local government finance, bank lending, and employment depends on that apparatus. A genuine rebalancing would mean allowing the unprofitable firms to fail, which would detonate the fiscal and social arrangements that keep the CCP in power. The crisis Froman fears is not a malfunction of Chinese capitalism; it is the form its continued reproduction would take.

The Long Shadow of the Iran Shock¶

Source: Foreign Affairs

The opening months of the Hormuz closure have produced a paradox that the policy establishment is struggling to digest: the largest supply disruption in the history of the oil market, and yet Brent crude hovering around $100. The authors are right to resist the complacent reading. But their own account suggests something more structurally significant than a lucky combination of cushions. The shock has been absorbed not because the system is resilient but because its centre of gravity has shifted.

The key variable is China. Its ability to reduce imports by drawing down inventories, curbing refinery runs and tightening domestic consumption effectively functioned as a global shock absorber. This is not the familiar story of Chinese demand growth driving prices; it is the opposite — a state-directed contraction of demand that stabilised the world market while Beijing absorbed the domestic cost. The United States, meanwhile, enjoys its own insulation as a net exporter. The uneven distribution of pain is therefore not an accident of geography but a reflection of who holds strategic autonomy in the energy system. Europe and much of Asia are left exposed, paying the price in rationing, fiscal strain and depleted gas storage ahead of winter.

The authors frame this as a problem of rebuilding resilience. But the deeper issue is that the instruments of resilience are themselves the products of state capacity and market power that most countries simply do not possess. The IEA’s emergency stock system, the Saudi and Emirati bypass pipelines, the US strategic reserve — these are the accumulated infrastructure of a particular hierarchy of states. When the authors warn that inventories are depleted and alternative routes vulnerable, they are describing the exhaustion of that hierarchy’s capacity to manage a crisis it created through the US-Israeli strikes that provoked the closure.

The modest price response also reflects the peculiar condition of the refining sector. Crude prices may have stayed low, but diesel and jet fuel have spiked — a sign that the bottleneck has moved downstream, where capacity is concentrated and inflexible. This is where the real economic damage will register: not in the headline barrel price but in the cost of moving goods and people. The global economy has not escaped the shock; it has merely deferred its transmission through a different layer of the supply chain.

Salvage work begins on tanker leaking oil off Oman, risk firm says¶

Source: Al Jazeera

The Caroline Bezengi has been bleeding crude into the Arabian Sea since June, and the only reason a salvage operation is now being framed as a fresh development is that the oil has finally reached Omani beaches. The timeline is the story. A vessel with no functioning flag state, no verifiable insurer, and a crew that reported an "unidentified explosion" sat leaking for two months while the response window closed. Ambrey's announcement is not the beginning of a rescue; it is the formalisation of a loss.

The war exclusion invoked by the International Oil Pollution Compensation Funds is the tell. The tanker is a suspected member of Russia's shadow fleet, and the compensation regime has determined that an act of war likely caused the damage. That single determination voids the entire liability architecture designed for precisely this scenario. The polluter pays principle collapses when the polluter is a shell entity flagged across three states, insured by nothing, and operating in the service of evading sanctions. The legal framework assumes a responsible party exists; the shadow fleet is engineered to ensure none does.

Greenpeace's satellite data shows the spill expanding from 45 to 1,300 square kilometres in roughly two weeks, while Omani authorities' own estimate lagged at 400. The discrepancy is not bureaucratic incompetence; it reflects the absence of any entity with the capacity and incentive to measure accurately. The marine reserve off Hallaniyat was established last year, and the monsoon season has made containment a secondary concern to stabilisation. The salvage crews are not racing to prevent damage; they are racing to prevent the wreck from becoming a permanent fixture of the seabed.

The international response remains unnamed beyond Ambrey and one oil spill firm. That opacity is the point. Every state involved has an interest in the shadow fleet's continued operation, just not in its visible failures. The spill is the cost of doing business in a sanctions regime that outsources its enforcement to the market and then pretends surprise when the market externalises the consequences.

Maersk delivers strong Q2 and raises full year guidance¶

Source: Hellenic Shipping News

Maersk’s upgraded guidance is a textbook illustration of how capital profits from the very frictions it claims to manage. The CEO’s framing of a "new era of heightened volatility" is accurate, but the volatility is not exogenous weather; it is the material expression of trade flows that have become structurally unbalanced since 2024. Far East exports, particularly Chinese, are surging into infrastructure that cannot absorb them, producing congestion that functions as a tollbooth for the carrier. The 22% increase in average freight rates against a 4.1% volume rise shows the margin expansion is not coming from moving more boxes, but from the scarcity premium attached to moving them at all.

The rerouting around the Strait of Hormuz is instructive. Capacity displaced from the Gulf was redeployed to other lanes, but the disruption did not reduce Maersk’s profitability; it enhanced it. The crisis in one region becomes an opportunity in another, and the company’s integrated logistics arm — with its Gulf landbridge solutions — monetises the same disruption twice. This is not profiteering in any conspiratorial sense; it is the normal operation of a vertically integrated monopoly responding to the price signals of a system in which bottlenecks are increasingly permanent features.

The raised guidance to USD 10.5-12.5bn EBITDA rests on an assumption of 4% global volume growth, but the more telling number is the EBIT swing in Ocean: from a USD 192m loss in Q1 to USD 935m profit in Q2. That oscillation is the new normal — not a cyclical upturn but a structural condition where earnings depend on the persistence of congestion. The USD 1.7bn Vietnamese terminal investment and the Brazilian electrification project are bets that this imbalance is durable enough to justify fixed-capital expansion. For the wider economy, the implication is that supply chain "resilience" has become a rent-generating asset class, and the carriers holding the infrastructure are the primary beneficiaries.

Halting the march of the African armyworm: a natural cure offers hope to farmers¶

Source: The Guardian

The discovery of a soil fungus that kills mature African armyworms is a genuine scientific advance, and the Guardian frames it as a story of hope for smallholders. But the hope is carefully circumscribed by the political economy the article only gestures at. Mathulwe’s breakthrough matters precisely because the existing arsenal of chemical controls fails against mature populations — and it fails, in part, because those controls are commodities sold to farmers who must buy them anew each season. A biological agent that persists in the soil and acts naturally disrupts that logic, which is why the research is notable and why its path to market will be contested.

The article’s own data does the analytical work. The 2017 outbreak destroyed up to 40% of Zambia’s maize. Zimbabwe’s inflation at 736% and Malawi’s at 35% in 2024 are not incidental colour; they are the conditions under which a farmer decides whether to purchase any input, chemical or biological. A fungus that can be cultured and applied locally, perhaps even propagated on-farm, threatens the input-supply chains that extract value from African agriculture. The question is whether the product, when it emerges from the lab, will be patented, packaged and priced like the chemicals it replaces, or whether it will be treated as a common resource.

There is also a temporal contradiction the piece does not name. The fungus works on mature caterpillars — precisely the stage at which chemical controls fail and the point at which outbreaks are already devastating fields. That is excellent for emergency response. But the armyworm’s damage is compounded by its mobility as a moth, and the article notes the fungus was found in one specific region of KwaZulu-Natal. A biological control that works only where it is already present, and must be distributed ahead of an outbreak that migrates across borders, requires coordination that fragmented state agricultural departments and cash-strapped governments are poorly positioned to provide. The science is sound; the delivery system is the weak link.

Zambia elections haunted by ghost of incumbent president's arch-rival¶

Source: The Guardian

The feud between Hakainde Hichilema and the late Edgar Lungu has outlived its object, which is precisely its political utility. The Tonse Alliance is not merely campaigning on policy; it is campaigning on the corpse of the former president as a symbol of the incumbent’s disrespect. By promising to repatriate Lungu’s body, Mundubile and Zulu have found a way to indict Hichilema’s character without contesting his economic record directly. That the government fought the burial in South African courts suggests the ruling party understood the symbolic stakes: a state funeral presided over by Hichilema would have laundered the feud into a display of national unity, closing the wound on the incumbent’s terms.

The economic numbers tell a story of stabilisation that has not reached the table. Inflation down from 24.6% to 6.5%, average income up from $1,127 to $1,318 — yet prices nearly doubled over the same period. The macroeconomy has been repaired for creditors and investors, particularly those eyeing copper deposits for the green transition, while the lived reality of wage labour has deteriorated. Hichilema’s 50% increase in student accommodation allowances, announced as a giveaway to young voters, is a targeted concession that reveals the pressure: the state can afford to buy off a constituency, but it cannot deliver the general rise in living standards that would make such gestures unnecessary.

The authoritarianism allegations are the sharpest contradiction. Hichilema’s team defends itself by invoking Lungu’s repression — the 15 arrests, the closed media houses, the switched-off internet — as if the baseline for democratic conduct were merely the absence of the previous regime’s worst excesses. But Transparency International’s monitoring of police bias against opposition rallies suggests the ruling party has internalised the playbook it once suffered under, deploying it in milder form. Vandome’s observation that Hichilema’s own experience of electoral insecurity is now counting against him captures the tragedy: the man who was arrested 15 times has learned the lesson that elections are won by controlling the terrain, not by trusting it. The ghost haunting this election is not Lungu’s corpse but the institutional memory of how power is actually exercised in Zambia.

AirAsia delays London, Bahrain flights amid capacity 'reset'¶

Source: FlightGlobal

The headline frames AirAsia’s decision as a "reset," but the substance is a retreat. The group is postponing its London and Bahrain launches while absorbing a quarterly loss driven by fuel costs. For a carrier whose entire model rests on stimulating demand through rock-bottom fares, the arithmetic has broken. Fuel is the one input cost that cannot be engineered away, and when it rises, the low-cost proposition — thin margins, high utilisation, ancillary revenue chasing — loses its elasticity. Passengers still fly, but not at the volume that makes a 5,000-mile A330neo route viable.

The "reset" language is instructive. It suggests a temporary recalibration rather than a structural problem, yet the underlying condition is one of overcapacity meeting input-cost inflation. AirAsia’s expansion into long-haul was always a bid to escape the brutal price competition of Southeast Asian short-haul markets. Delaying it means those aircraft and crews stay grounded or are redeployed into the very markets where yields are already compressed. The group is not shrinking because demand vanished; it is shrinking because the cost of carrying a passenger now exceeds what the market will bear.

What is notable is the timing. This is not a crisis of demand but of cost — fuel prices are squeezing carriers across the region, and the weakest balance sheets bend first. AirAsia’s aggressive order book and network ambitions were built on assumptions of cheap energy and endless growth. Those assumptions have collided with a fuel market that no longer cooperates. The postponement is a quiet admission that the growth model, as constructed, cannot absorb the current cost environment. For the broader industry, it signals that long-haul low-cost operations remain a marginal experiment, viable only when every external variable cooperates.

Malta Air 737 window-loss event: Fractured blade punctured engine inlet forward of fan¶

Source: FlightGlobal

The NTSB’s finding on the Malta Air 737 window loss is a textbook case of cost discipline migrating from the balance sheet to the fuselage. A fractured fan blade, showing evidence of fatigue, punctured the engine inlet and sent debris through the cabin wall. Fatigue in a rotating component is not an act of God; it is a maintenance schedule deferred, an inspection interval stretched, or a part kept in service past its economic replacement point. Ryanair Group’s entire business model is built on extracting maximum cycles from airframes and engines while holding unit costs below every competitor in Europe. The metallurgical crack that opened at 11F is the physical trace of that accounting logic.

The bird remains complicate the picture usefully. A bird strike is the kind of event that operators cite to shift blame to the environment, but the NTSB’s dual finding — fatigue plus bird — suggests the blade was already compromised before impact. The bird may have been the trigger, not the cause. This matters because the regulatory response will likely focus on the immediate technical fix rather than the economic pressures that determine how often blades are inspected and when they are retired.

For the low-cost sector, the implications are uncomfortable. The margins that made Ryanair the most profitable airline in Europe were never going to be achieved by flying newer, more expensive equipment. Every quarter of record profits is also a quarter of ageing fan blades spinning closer to their fatigue limit. The industry will tighten inspection regimes, costs will rise, and the pressure will simply move to the next component that can be safely — or not so safely — stretched.

Gol appoints new chief executive and president¶

Source: FlightGlobal

The appointment of Rafael Luiz Canossa as Gol’s chief executive is less a personnel story than a signal that Abra Group is preparing its Brazilian subsidiary for a different kind of competition. Gol has spent years as a domestic carrier, its network built around the density of Brazilian trunk routes and the cost discipline of a single fleet type. The move toward long-haul international flying changes that calculus entirely.

Abra’s ambitions for Gol collide with the actual structure of the Brazilian market. Domestic aviation there has consolidated into a duopoly, with Latam and Gol splitting the profitable routes while Azul scrapes the edges. That arrangement has kept yields tolerable but has also capped growth. International long-haul from Brazil is a different game: it means going head-to-head with Latam’s established transatlantic and North American operations, and with the US majors on the densest routes. Gol would be entering those markets without the feed, the frequent-flyer partnerships, or the widebody experience that incumbents have spent decades building.

The timing is telling. Brazilian carriers have been squeezed by aircraft financing costs and a weak real, which makes dollar-denominated leases and fuel more expensive in local terms. Expanding internationally is a hedge against domestic currency risk — revenue in dollars or euros offsets costs that are already dollarised. But it also requires the very capital that is scarce in this environment. Canossa’s mandate is presumably to find that capital, or to make the case that Gol can grow its way out of a cost structure that domestic yields no longer justify.

Whether this is ambition or necessity depends on how much pressure Abra is under from its own creditors. A management change at this stage suggests the latter.

The AI Growth Paradox¶

Source: Foreign Affairs

The $40 trillion figure is doing more than measuring Washington's borrowing; it measures the collapse of the post-2008 consensus that debt could be treated as a free input. Rogoff's argument is essentially that the state has spent two decades assuming the future would deliver growth cheaply, and AI is now being drafted as the latest guarantor of that promise. The secular stagnation orthodoxy of the 2010s was always a convenient fiction for a ruling class that wanted stimulus without taxation; now the bill arrives just as the putative solution — an AI-driven productivity surge — threatens to raise the interest rates that make the debt unserviceable.

The real tension here is between fictitious capital and the state's fiscal base. If AI genuinely transforms productivity, it should expand the surplus available for taxation. But the boom itself is being financed by debt, and the interest payments on that debt now exceed defence spending. The state is betting on growth it cannot control to service obligations it refuses to address through redistribution. Rogoff is right that inability to pay is rarely the trigger for sovereign crises; the trigger is a loss of confidence among creditors. But he stops short of asking why the US state cannot simply tax the enormous concentration of wealth that the AI boom is generating. The answer is political: the same class that owns the AI firms owns the debt, and it will not tax itself to save the state.

The global implication is that the dollar's exorbitant privilege is now a hostage to the AI trade. If the boom falters, the debt spiral accelerates; if it succeeds, higher rates squeeze every economy that borrowed in dollars. Either way, the periphery pays for the core's fiscal irresponsibility.

Databricks wanted to raise $1B, investors wanted $15B. It settled on $5B at a $190B valuation.¶

Source: TechCrunch

The mechanics of this raise are worth pausing on because they invert the usual power relation between issuer and buyer. Databricks wanted $1 billion, found itself facing $15 billion of demand, and settled on $5 billion — not because it needed the capital, but because refusing long-term backers would have soured relationships. That is the logic of the late-stage private market operating at full maturity: the company is not selling equity to fund operations so much as managing a constituency. Ghodsi's own framing — that a press leak forced his hand — is a convenient fiction. The real constraint is that when a firm has raised $20 billion in twenty months, its investor base is not a source of funding but a claim structure that must be serviced with allocations.

The valuation itself does the analytical work. A $190 billion price tag on $7 billion of run-rate revenue, growing at 80%, cash-flow positive, is not absurd by current AI-market standards. But the gap between the $1 billion the company wanted and the $15 billion on offer measures something other than the quality of the business. It measures the pressure on institutional capital to find homes for accumulated liquidity in a moment when the public markets offer few comparable AI exposures. Databricks is a private proxy for the AI trade, and the investors piling in are not making a company-specific bet so much as buying a position in the sector's continued overvaluation. The fact that the company is cash-flow positive and still raising — with multibillion-dollar cloud commitments to all three hyperscalers — shows where the money actually goes: not to survival, but to securing compute capacity and buying up smaller firms in an arms race where the main competitors are the very cloud providers being paid.

The joke about running out of alphabet letters for funding rounds is sharper than it appears. Databricks' repeated private raises, rather than an IPO, let it set terms unilaterally while keeping financials out of public scrutiny. But the roster of investors — Coatue, Blackstone, T. Rowe Price — will eventually want liquidity, and at a $190 billion valuation, the exit pressure is enormous. Ghodsi's stated preference for staying private is rational for now: the moment this company touches public markets, the fiction of its valuation meets quarterly disclosure. The question is not whether Databricks will IPO, but whether the AI sector's current pricing can survive contact with actual public-market scrutiny when it does.