2026-06-13 Observatory briefing¶

The Real Problem With Global Trade¶

The Foreign Affairs piece by Setser and Vallée performs a useful service by naming what mainstream trade discourse routinely obscures: that China’s export surge is not simply a story of industrial policy or comparative advantage, but of active currency suppression. The renminbi’s depreciation since the property collapse of 2021 has been a deliberate mechanism to offload overcapacity onto global markets, allowing Beijing to meet growth targets that domestic demand—shattered by the bursting of a vast speculative bubble—can no longer sustain.

What is revealing is the authors’ diagnosis of the G-7’s paralysis. They note that Washington has abandoned any serious demand for rebalancing, content instead to celebrate AI-related capital goods imports that will push the US trade deficit higher. This is not incompetence but a reflection of deeper contradictions. American capital benefits from cheap Chinese inputs and the dollar’s reserve status; tariffs on final goods merely redirect the surplus through intermediate supply chains in Southeast Asia. The US state is unwilling to challenge the financial architecture that makes the whole system function, even as it decries the results.

The real problem, then, is not that policymakers lack a solution. It is that the solution—forcing renminbi appreciation through coordinated currency diplomacy—would require confronting the structural dependence of both US and European economies on cheap Asian exports and the financial flows that recycle the resulting surpluses. The G-7’s studied avoidance of exchange rates is not a blind spot. It is a political choice that reveals whose interests are served by the current disorder.

China’s Edifice Complex¶

Source: Foreign Affairs

The Chinese Communist Party’s inability to curb "visibility projects" — grandiose, loss-making infrastructure built to impress superiors — is not a failure of discipline but a structural feature of its political economy. As Ning Leng argues, the top-down cadre evaluation system rewards spectacular, photographable achievements over the invisible, unglamorous foundations of sustainable growth: sewage pipes, recycling infrastructure, genuine R&D capacity.

This is a contradiction internal to the party-state’s accumulation model. The state directs massive investment to maintain growth, legitimise the party’s developmentalist claim, and absorb surplus labour and capital. But the political imperative to demonstrate output systematically distorts the content of that investment. A $1.6 billion bridge that operates at a loss is not a miscalculation; it is a rational response to the bureaucratic incentive structure. The result is a growing mass of fixed capital that is either underutilised or functionally useless — a form of waste that is not accidental but systemic.

Xi Jinping’s rhetorical crackdowns will prove futile because the problem is not individual corruption but the logic of the system itself. As growth slows and local government revenues shrink, this waste becomes more acute. The party cannot resolve it without undermining the very mechanisms — top-down target-setting, performance competition, rapid visible delivery — that have sustained its developmental state.

The article does not mention it, but this dynamic has implications for global supply chains: China’s overbuilt airports, tech parks, and industrial zones represent a vast overhang of capacity that can be dumped onto world markets at below-cost prices, exacerbating overaccumulation crises elsewhere. The "visibility trap" is also a trap for the global economy.

Dangote ramp-up slashes WAF product imports, reshapes clean tanker routes: analysts¶

Source: Hellenic Shipping News

Dangote refinery: a disruption in the geography of fuel flows¶

The Dangote refinery’s ramp-up is not merely a Nigerian industrial success story — it is a material reconfiguration of Atlantic basin tanker routes, with measurable consequences for shipping capital. West African clean product imports fell 23% month-on-month in May; BIMCO reports a 44% drop. The Rotterdam-to-Lagos MR tanker run, long a staple of the Atlantic clean products market, has been gutted. MR ton-mile losses were limited only by a 34-fold surge in volumes from the Americas — a compensatory flow that reveals how quickly trade patterns re-form under pressure.

What is unfolding is a displacement of import dependency by regional export capacity. Dangote now supplies over 90% of Nigerian gasoline demand and is exporting to Ghana, Ivory Coast, South Korea, the US, and Europe. The refinery’s shift from destination to source has created a surplus of MR tankers in the Atlantic Basin, but this has not produced a straightforward collapse in rates. The refinery’s own export volumes and the emergence of new regional hubs have partially absorbed the excess fleet capacity.

The deeper dynamic is spatial: long-haul imports are being replaced by shorter shuttle voyages, reducing overall ton-miles even as vessel call frequency increases. Lome’s role as a storage and redistribution hub is threatened. Meanwhile, Dangote’s planned capacity doubling to 1.4 million b/d by 2028, combined with new marine jetty infrastructure and automated port systems at Lekki, signals an ambition to become a global export hub — one that competes directly with European refineries and Middle Eastern exporters.

This is not a crisis of overaccumulation. It is a crisis of location — a sudden obsolescence of established shipping routes and the capital tied to them. The MR tankers idled by the loss of the Nigerian run are not yet worthless, but their employment has become contingent on new, thinner margins. The real question is whether the Atlantic Basin can absorb this fleet without a downward spiral in rates, or whether the surplus will eventually force consolidation and scrapping.

Trade stands at a crossroads¶

Source: Hellenic Shipping News

The OECD’s latest Economic Outlook presents global trade as a system whose fate hinges on the duration of Middle East disruptions. This framing is revealing: it treats the crisis as an external shock to an otherwise functional order, rather than a symptom of deeper structural fragilities.

The report’s two scenarios — time-limited versus prolonged disruption — share a common trajectory of deceleration. Even in the optimistic case, trade growth falls from 5% to 2.9% by 2027. The slowdown is driven not by demand collapse but by input shocks: energy costs, maritime chokepoints, and cascading supply shortages. This is a supply-side crisis, rooted in the geographical concentration of critical inputs — hydrocarbons, fertilisers, petrochemicals — in a single volatile region.

What the OECD describes as “indirect linkages” is a polite term for the vulnerability built into global production networks. Semiconductor fabrication, agriculture, chemicals — all depend on Gulf-sourced inputs with no ready substitutes. When the report warns that shortages would force “enforced rationing of energy for businesses”, it describes a situation where capital cannot simply price its way out of a physical bottleneck.

The prolonged scenario implies a reordering of trade geography. Asian manufacturing hubs, central to global accumulation, would be hit hardest. Energy exporters might gain temporarily, but only if production itself is not disrupted. Import-dependent emerging markets face deteriorating terms of trade and rising financing costs — a classic mechanism by which crisis in the core is transmitted to the periphery.

The turn toward “resilience” and “friend-shoring” is not a policy preference but a necessity imposed by the contradiction between globally integrated production and nationally fragmented political authority. The OECD’s call to maintain openness while diversifying supply chains is a wish, not a strategy. The real question is whether the system can absorb this shock without a more fundamental restructuring — or whether 2026 marks the beginning of a permanent fragmentation of trade networks.

Tanker Market Stays Elevated in May¶

Source: Hellenic Shipping News

The tanker market remains elevated well into 2026, with freight rates on several key routes still showing triple-digit year-on-year increases. OPEC’s latest report confirms that while rates have eased from March’s record peaks, they are not normalising — they are settling at a higher plateau. VLCC rates on the Middle East-to-East route are up 878% year-on-year. Suezmax and Aframax rates, though down month-on-month, remain 129% and 76% above 2025 levels respectively.

The surface explanation is supply disruption and long-haul rerouting. But the underlying dynamic is more structural. The sustained elevation of freight rates — across vessel classes and despite month-on-month declines — points to a shipping sector absorbing the costs of a fragmented global oil market. US Gulf exports remain near record levels, sustained by Strategic Petroleum Reserve releases, while Atlantic Basin crude is being pulled to Asia to replace disrupted supply. This is not a temporary spike; it is the logistical expression of a world in which energy trade routes are being forcibly lengthened and multiplied.

The divergence between East and West of Suez in the clean tanker market is revealing. East of Suez rates rose; West of Suez fell sharply. This suggests that Asian demand is the primary driver of tightness, while Atlantic markets are experiencing a relative glut of available tonnage. The repositioning of vessels from Asia to the Atlantic — noted in the report — is a rational response, but it also confirms that the system is being stretched unevenly.

What this amounts to is a permanent increase in the cost of moving oil, baked into the structure of trade itself. The shipping industry is not experiencing a crisis of overaccumulation — it is enjoying a profitability boom driven by geopolitical fragmentation. The question is how long the real economy can absorb these costs before demand destruction sets in.

Modi Is Rigging Indian Democracy¶

Source: Project Syndicate

Jayati Ghosh’s account of India’s electoral manipulation describes a process that is less a sudden rupture than a logical extension of the BJP’s political strategy. Under the cover of routine voter-roll maintenance, the Election Commission has purged tens of millions of names, disproportionately targeting opposition strongholds, poor voters, and Muslim minorities. The effect is an electoral system increasingly designed to select the electorate rather than be selected by it.

What is striking here is not the fact of manipulation — every bourgeois democracy has its tricks — but the scale and the institutional complicity. The Election Commission, formally independent, has become an instrument of partisan consolidation. This is not a breakdown of democracy but its reconfiguration: the state apparatus is being used to narrow the political terrain so that popular discontent, which Ghosh notes is rising, cannot find electoral expression.

The deeper contradiction is that this rigging occurs alongside growing economic strain. India’s growth model, reliant on a narrow base of consumption and services, has failed to absorb the vast reserve army of labour. The BJP’s political response has been to foreclose democratic channels for that frustration. The result is a system that suppresses the very class forces it claims to represent — a recipe not for stability, but for a more volatile form of rule.

Air Canada operates first domestic A321XLR flight ahead of transatlantic debut on 15 June¶

Source: FlightGlobal

Air Canada has finally put its first A321XLR into commercial service, four years after ordering and two years behind schedule. The aircraft will fly domestically for a few days before launching Montreal–Toulouse on 15 June — a route that neatly symbolises the XLR’s commercial logic: connecting mid-sized cities across the Atlantic without the capacity or cost of a widebody.

The delays matter. They reflect a broader pattern across aerospace manufacturing where supply chains, engine problems (Pratt & Whitney’s GTF issues are well documented), and labour constraints have stretched delivery timelines across the board. Air Canada’s wider fleet plan — 787-10s, A350-1000s, A220s, 737 Maxes — suggests a carrier hedging aggressively, unwilling to bet on a single OEM or engine type.

The XLR itself is a response to a structural tension in network aviation. The old model assumed that long-haul meant high-density hubs and large aircraft. The XLR cracks that assumption open: it allows airlines to bypass hubs entirely, linking secondary cities directly. This is not a niche product. It threatens the economics of hub-and-spoke carriers and, by extension, the value of airport slots at major connecting airports. For labour, the implications are mixed — more routes, but thinner crews and potentially less job security at legacy carriers whose networks are being unbundled.

Air Canada’s 30 XLRs will not transform global aviation on their own. But the type’s entry into service signals that the industry is shifting toward a more fragmented, point-to-point long-haul structure — one that rewards fleet flexibility over network scale. That is a genuine shift in the competitive terrain, not just a new plane.

Greece poised to become latest C-390 customer¶

Source: FlightGlobal

Greece is set to acquire three Embraer C-390s via Portugal’s existing purchase options, replacing six ageing C-130 Hercules — the oldest of which is 64 years old. The deal, valued at roughly €600 million, follows a pattern already established in Europe: Sweden and Austria ordered through a Dutch master contract, while Portugal acts as intermediary for Greece. Embraer is not directly involved in negotiations; the arrangement is government-to-government.

This is not a story about a sudden leap in military capability. The C-130s Greece operates are decades past their intended service life. Replacement is a necessity, not a strategic choice. What is more revealing is the structure of the sale. By routing the order through Portugal’s options, Greece avoids a direct tender or open competition. The transaction is opaque, mediated by another state’s contractual position. This is how smaller NATO members manage procurement within the alliance’s informal hierarchy — not through market competition, but through bilateral state arrangements that bypass transparency.

For Embraer, the sale does not immediately affect its backlog of 60 firm orders plus 29 options. The company plans to increase production from six deliveries this year to ten annually by 2030. The Greek order shifts three options to firm sales, but the real significance is the growing European footprint of a Brazilian manufacturer. The C-390 is now operated by Portugal, Hungary, and soon Austria, Sweden, and Greece. This is not inter-imperialist rivalry in the classic sense — Brazil is not challenging US or European defence primes on their own terrain. But it does indicate a fragmentation of the military aerospace market, where medium powers seek alternatives to Lockheed Martin and Airbus without fully exiting the NATO framework. The C-390 is a niche product filling a gap left by the withdrawal of US and European manufacturers from the medium transport segment. That gap is itself a product of overaccumulation in high-end fighter and bomber programmes, which concentrate capital and R&D while leaving less glamorous but essential roles to second-tier suppliers.

Algerian regulator under pressure as first carrier put on European blacklist¶

Source: FlightGlobal

The blacklisting of Air Express Algeria exposes a contradiction at the heart of aviation safety regulation: the formal sovereignty of national regulators versus the real subordination of their capacity to the demands of international capital circulation.

Air Express Algeria, linked to the energy sector, expanded into medical evacuation flights to the EU under general aviation rules, bypassing commercial air transport authorisation. This is not merely a compliance failure. It reflects a structural pressure on carriers in resource-exporting states to extend operations into higher-value markets without the supervisory infrastructure such expansion requires. The airline's corrective action plan relied on a "deficient" root-cause analysis — a bureaucratic symptom of a deeper inability to internalise the regulatory standards of the core economies it sought to serve.

More revealing is the exposure of ANAC Algeria, the regulator established in 2023. An ICAO audit found "serious deficiencies" in its core functions. During the EU hearing, ANAC claimed the carrier's deficiencies were nearly resolved; the carrier itself said they would take until year-end. This contradiction is not incompetence alone. It is the material gap between the formal adoption of international standards and the actual institutional capacity to enforce them — a gap widened by the uneven development that defines the global aviation hierarchy.

The EU blacklist functions here as a disciplinary mechanism, enforcing the real subsumption of peripheral regulators into a system whose standards are set by the dominant blocs. For global supply chains, the implication is clear: as carriers from the periphery attempt to integrate into European networks, the regulatory friction will intensify, not diminish.

Silicon Valley’s Bad Bet on the Gulf¶

Source: Foreign Affairs

The Foreign Affairs piece describes a strategic miscalculation dressed as a deal. Trump’s $2.2 trillion Gulf AI infrastructure package was sold as a win-win: cheap energy and lax regulation for US tech, geopolitical alignment for Washington. But the March 2026 drone strikes on AWS data centres in the UAE exposed the underlying contradiction. The infrastructure was built as though the Gulf were a neutral energy depot, not a theatre of war shaped by decades of US military intervention.

The analysis is strongest on the web of dependencies. The attack didn’t just disrupt regional ride-hailing apps. It knocked out cloud workloads routed through Gulf servers for companies that had no idea their data passed through the zone. This is not a supply chain story in the usual sense — no raw materials were delayed — but a revelation of fictitious geographical neutrality. Capital flows treated the Gulf as frictionless space; the missiles proved it is still territorial, contested, and combustible.

The Pax Silica framework, designed to manage chip diversion and Chinese competition, had no provisions for physical defence. This is a classic case of planning for the last war — or rather, for the war that never came. The real risk was not that Gulf states would re-export Nvidia chips to Huawei, but that the data centres themselves would become lawful military targets. Legal scholars had flagged this. The US simply assumed its own escalation dominance would protect the assets.

The piece ends with a recommendation to bring projects back to the US. That may be sensible, but it misses the deeper point: the overaccumulation of capital in AI infrastructure has driven a frantic search for any site with cheap power and permissive regulation. The Gulf was never a safe bet; it was a desperate one. The contradiction is not that the bet failed, but that it was ever considered prudent.

Anthropic’s safety warnings may have just backfired — the government has pulled the plug on its most powerful AI¶

Source: TechCrunch

The US government has ordered Anthropic to disable its two most powerful AI models, Claude Fable 5 and Mythos 5, for all users worldwide, citing national security concerns over a claimed jailbreak. Anthropic argues the evidence is thin — a verbal description of a narrow vulnerability that, it insists, is already replicable with publicly available models including OpenAI's GPT-5.5.

The irony is sharp and material. Anthropic built its brand and its IPO narrative around safety. It restricted Mythos precisely because it was too capable, marketing the restriction as responsible stewardship. That caution has now handed the state a justification to shut down the commercial product derived from it. The company's own fear-based differentiation has been turned into a regulatory weapon against it.

Sam Altman's earlier jibe — that Anthropic was selling bomb shelters after claiming to have built a bomb — now reads less as competitive sniping and more as a prediction of this exact contradiction. When you tell the state you have created something uniquely dangerous, the state may eventually act on that claim, regardless of the technical merits.

The broader dynamic is instructive. AI capital needs state permission to scale, but the state's threshold for intervention is shaped by the very hype the industry generates to attract investment. Anthropic's predicament is not a case of overreach by a rogue regulator. It is the logical outcome of an industry that must simultaneously terrify and reassure the same audience — investors, customers, and the state — to survive.

Meta’s months-old AI unit is a soul-crushing gulag, say the engineers stuck inside it¶

Source: TechCrunch

Here, the contradiction is not between capital and labour in the abstract, but between the form of labour and the function of the worker under a specific stage of accumulation.

Meta has spent billions on AI infrastructure and acquisitions — notably the $14.3bn purchase of Scale AI — only to discover that its models still cannot outperform humans at basic technical tasks. The solution is not more capital, but more cheap, compliant human labour of a particular kind: generating puzzles and coding problems to train the very systems designed to replace them. This is the real content of "applied AI."

The problem is that Meta’s workforce is not structured for this. These are not gig-economy data-labourers in the global south, but salaried engineers accustomed to the ideological promise of "impact" and autonomy. The company’s response — forced transfers, 50-to-1 manager ratios, keystroke monitoring — strips away the fiction that these workers are creative professionals. They are being treated as what they have become: inputs in a production process that devalues their specific skills even as it extracts them.

The "gulag" language is hyperbolic, but it captures a real degradation: the reduction of high-cost technical labour to the raw material of a machine that will eventually make that labour superfluous. Zuckerberg’s admission that Meta employees are "significantly higher" intelligence than contractors is not a compliment — it is a statement of comparative surplus value. The company needs better data, not cheaper data. For now.

The broader implication is that the current AI boom, far from resolving capital’s need for labour, has created a new bottleneck: the production of training data itself. And the most efficient source of that data, for now, is the very workforce being prepared for obsolescence. That is not a management problem. It is a structural trap.