2026-07-19 Observatory briefing¶

US hits Iran for eighth consecutive night, Iran returns fire on Gulf bases¶

The eighth consecutive night of US strikes on Iran marks a shift in the character of the conflict. What began as a battle over the Strait of Hormuz — a strategic chokepoint for global oil flows — has widened into strikes on desalination plants and bridges. The Bonji plant destroyed, cutting water to 10,000 people; Qeshm Island’s plant damaged. This is not mission creep born of tactical necessity. It is the logic of a war that cannot achieve its stated aim — degrading Iran’s ability to threaten shipping — through military means alone, so it expands the target set to include the civilian infrastructure that sustains the society hosting the adversary.

The US frames the strikes as retaliation for Iranian attacks that killed two service members in Jordan. But 16 US service members have now been killed since February, and the strikes have killed at least 50 Iranians. The proportionality is inverted, and the escalation is self-sustaining: each round of US strikes generates the domestic pressure in Iran that the IRGC can exploit to justify its own retaliatory fire on Kuwaiti bases. The two sides are locked in a reciprocal dynamic that neither can break without appearing weak, and neither can win without risking a wider war.

Iran’s deputy foreign minister notes the US is violating the interim deal. This is the diplomatic corpse being dragged out to prove a point. The deal was already hollowed out; now it serves only as a rhetorical cudgel. Meanwhile, the population in Tehran experiences the war as a fog — no sense of duration, no visible endpoint, no diplomatic off-ramp. That uncertainty is itself a weapon, and it is being wielded by both sides.

America’s Trillion-Dollar Failure¶

Source: Foreign Affairs

The piece, written by a sitting Democratic congressman, performs a useful service by naming the 1993 "Last Supper" as the moment the US defence industry was deliberately consolidated into five oligopolistic firms. What Khanna presents as a policy failure is better understood as a class strategy that has now exhausted its usefulness for the ruling class. The consolidation was not a mistake; it was a solution to falling profit rates in the 1980s, achieved by concentrating state-guaranteed demand into fewer hands. The result—51 firms collapsing into five, the defence workforce shrinking from three million to 1.1 million—was the intended outcome, not a bug.

The contradiction Khanna cannot name is that the Pentagon's dysfunction serves capital perfectly well. Over-budget, behind-schedule programmes like the F-35 or Golden Dome are not failures; they are mechanisms for transferring public money to Lockheed Martin and Northrop Grumman with minimal accountability. The $24.4 billion for a missile defence system that cannot win a modern conventional war is not waste—it is a subsidy. Khanna's proposed remedies—banning stock buybacks, capping executive compensation, antitrust action—would require the state to discipline the very firms it spent three decades empowering. Congress has the power of the purse but not the will to use it against the defence oligopoly, because that oligopoly is not external to the state but fused with it through the revolving door and campaign finance.

The real question the article sidesteps is why China can produce a million attack drones this year while the US struggles to produce 300,000. The answer is not that China has better procurement rules. It is that China's state-owned defence sector is not organised around shareholder value. The US defence industry is. Until that changes, no amount of congressional oversight will fix the underlying dynamic: private profit extracted from public necessity, with the bill paid in strategic vulnerability.

Subsidies Do Not Explain China’s Competitiveness¶

Source: Project Syndicate

The OECD report that Kai Guo is pushing back against has clearly struck a nerve, because the argument he makes in response is revealing in what it concedes. He does not deny that Chinese industrial policy exists or that subsidies have been substantial. Instead, he argues that subsidies are no longer the most convincing explanation for Chinese firms’ global dominance. The shift is telling: the terrain of debate has moved from whether the state intervenes to whether state intervention is the decisive factor.

Guo’s counter-explanation is that Chinese firms now compete on genuine advantages: scale, supply-chain density, rapid iteration, and a vast domestic market that allows cost curves to be driven down before firms even look abroad. This is not wrong, but it smuggles in a premise that deserves scrutiny. Those advantages did not fall from the sky. The domestic market was built through deliberate state-directed urbanisation and infrastructure spending. The supply chains were assembled through decades of targeted investment and technology transfer policies that were themselves a form of subsidy — just not always the direct production subsidy the OECD fixates on.

The real weakness in Guo’s piece is that he treats competitiveness as a technical achievement rather than a social one. Chinese firms are competitive because the state has socialised many of the costs of production — land, energy, research, logistics — while allowing profits to be privately captured. That is not a refutation of the subsidy thesis; it is a more sophisticated version of it. The OECD may overstate direct handouts, but Guo’s alternative — that Chinese firms are simply better at organising production — leaves the political economy of how that organisation was funded entirely unexamined.

UK aid cuts 'reduce bilateral support to some African countries by 90%'¶

Source: The Guardian

The Labour government’s defence-driven aid cuts are not simply a matter of broken promises or moral failure, though they are both. The Foreign Office figures show bilateral support to Mozambique and Malawi dropping 90% by 2029, with Rwanda and Sierra Leone losing 80%. The stated rationale — that channelling funds through multilateral institutions like the World Bank is more efficient — masks a deeper political logic: the shift from direct country-to-country grants to multilateral lending transforms aid from a political relationship into a financial one, where the UK retains influence through institutional leverage rather than direct obligation.

The timing is instructive. The UK takes up the G20 chair next year, and the foreign secretary’s language of “modernised partnerships” and making “every pound work harder” echoes the World Bank’s own turn toward private capital mobilisation and risk-sharing instruments. This is not efficiency; it is the subordination of development policy to the demands of fiscal consolidation and inter-imperialist positioning. The cuts free up resources for defence spending while outsourcing the political costs of austerity to multilateral bodies where the UK’s voting share still carries weight.

The charities’ objections — that the cuts “send a global message” — are accurate but incomplete. The message is not about Britain’s moral standing but about the material reality of a declining imperial power choosing to concentrate its shrinking resources on military capacity rather than the social reproduction of peripheral economies. The incoming prime minister inherits a development apparatus hollowed out not by accident but by a deliberate reordering of state priorities in which African populations are simply the most disposable variable.

US manufacturing employment is down, but each state has its own story¶

Source: FRED Blog

The FRED Blog's map of state-level manufacturing employment for May 2026 is a useful corrective to the national aggregate, but it presents the data as a collection of local accidents — a factory opening here, a closure there — rather than as the uneven expression of a single structural process. The national decline of 0.4% is not simply the sum of fifty independent stories; it is the surface of a deeper reorganisation of productive capital.

Connecticut gains 3% while Virginia loses nearly 5%. These are not random. Connecticut’s manufacturing base is heavily oriented toward aerospace and defence — sectors sustained by state contracts and military demand, not by competitive accumulation in civilian markets. Virginia’s losses, by contrast, reflect the continued hollowing out of its traditional manufacturing alongside the expansion of logistics and data-centre infrastructure, which absorbs land and labour without producing the same density of industrial employment. The map captures a spatial fix in motion: capital is not fleeing manufacturing altogether but concentrating in politically protected niches, while abandoning regions where the rate of profit has fallen below the threshold required for reinvestment.

The District of Columbia’s 9.1% decline is the most revealing figure. DC has almost no manufacturing to lose; the percentage drop signals the evaporation of a handful of remaining print shops, food-processing plants, or construction-material fabricators. That a capital city — the nerve centre of the world’s largest economy — cannot sustain even a token industrial base is not a quirk of local conditions. It is the logical endpoint of a half-century in which the US economy has subordinated productive industry to the circulation of fictitious capital, financial services, and the administrative apparatus of empire. The map’s real story is not that some states are doing better than others, but that the entire territory is being sorted into zones of extraction, logistics, and political rent-seeking, with manufacturing reduced to a residual category.

2,000-Jet Shortage: Boeing Warns Global Supply Squeeze Will Last Into The 2030s¶

Source: Simple Flying

Boeing’s forecast of a 2,000-aircraft undersupply in 2026, with shortages persisting into the 2030s, is less a prediction of scarcity than an admission that the industry’s production apparatus has structurally failed to keep pace with the accumulation demands of its major customers. The numbers are telling: 43,625 deliveries projected over two decades, yet current output remains below 2018 levels despite passenger traffic having recovered to pre-pandemic benchmarks. This is not a temporary bottleneck but a systemic lag between the financialised expectations of airline expansion and the material limits of aerospace manufacturing.

The split between replacement (21,475) and growth (22,150) deliveries reveals the dual pressure on Boeing. On one side, airlines need new planes to retire older, less fuel-efficient fleets — a sustainability imperative that also serves to maintain fleet values and access to capital markets. On the other, they need expansion to capture projected 4% annual passenger growth, particularly in China (21% of deliveries) and Eurasia (20%). Yet Boeing cannot deliver because its own production has been hobbled by certification delays on the MAX 7, MAX 10, and 777-9 — delays rooted in the same cost-cutting and design shortcuts that produced the 737 MAX crisis.

The contradiction is concrete: the very financial and regulatory pressures that drove Boeing to maximise shareholder returns through the 2010s have now frozen its ability to supply the aircraft that airlines need to realise their own growth projections. The 2,000-jet gap is not a market imbalance to be corrected by price signals; it is the material expression of a production system that prioritised fictitious capital over productive capacity. Until Boeing can resolve the tension between its financial structure and its factory floor, the shortage will remain a structural feature, not a temporary one.

FAA returns to Boeing full authority to issue airworthiness certificates¶

Source: FlightGlobal

Seven years after the 737 Max crashes exposed the fiction of self-regulation, the FAA has handed Boeing back the power to certify its own aircraft as airworthy. The regulator’s own language is telling: it took “months of thorough data and safety review” to conclude that Boeing’s quality was “consistent” enough to trust again. But the data in question was produced by the same production system that, in 2019 and 2022, had to be stripped of certification authority precisely because it could not be trusted.

The FAA’s logic is circular. It restored partial authority in September 2025, then alternated weeks with Boeing, comparing the quality findings from each period. Finding them comparable, it concluded Boeing was safe to self-certify. But comparability between Boeing’s output and the FAA’s own is not evidence of adequate oversight — it is evidence that the FAA’s own inspectors, embedded in the same production apparatus, were already working within the same degraded quality regime. The comparison benchmarks nothing.

What has changed is not Boeing’s quality system but the pressure to ramp production. A new 737 Max line is opening. The 777-9 is in flight test. The company is rebuilding momentum ahead of Farnborough. Returning certification authority removes a bottleneck that would otherwise slow deliveries as output increases. The FAA’s decision is a production decision dressed as a safety decision.

The Organisation Designation Authorisation system was always a delegation of state power to capital in the name of efficiency. The crashes did not abolish it; they merely interrupted it. Now that the political heat has subsided and the production targets loom, the delegation is restored. The FAA will “continue inspections, audits and monitoring” — the same promise that preceded the last disaster.

How American & United Airlines Are Turning 19-Year-Olds With Zero Flight Hours Into Branded Airline Pilots¶

Source: Simple Flying

The article describes a structural shift in how two major US airlines reproduce their labour force. United and American are no longer passive buyers on a pilot labour market; they are vertically integrating training into their own operations, financing candidates from age 19 and guiding them through a standardised pipeline from zero hours to the right-hand seat of a mainline jet. The stated driver is demand — 600,000 new pilots needed globally over two decades — but the material logic runs deeper.

What is being solved is not simply a numerical shortage but a crisis of predictability. The traditional model left airlines exposed to the vagaries of individual career decisions, instructor turnover, and regional airline churn. By absorbing the cost and risk of training, the carriers gain something more valuable than pilots: control over the timing, quality, and cultural formation of their future workforce. The 1,200-hour command requirement for United's Aviate participants is not just a safety threshold; it is a mechanism to ensure pilots arrive already socialised into company norms, having never worked for a non-affiliated operator.

The financing arrangements — American's credit union loans of up to $148,000 with 45-month payment deferrals — remove the traditional barrier of upfront cost, but they also bind the trainee to the carrier through debt. This is not a gift; it is a form of indentured human capital development. The candidate trades financial uncertainty for career certainty, while the airline trades capital outlay for a reliable, brand-loyal labour supply whose entire professional identity has been shaped within the company's ecosystem from the first lesson. The pilot shortage is real, but the response reveals how capital responds to labour market tightness: not by raising wages to attract existing workers, but by restructuring the production of workers themselves.

Kimi: Threat or menace?¶

Source: TechCrunch

The panic around Kimi K3 is not really about technology. It is about who controls the conditions under which value is extracted from it. Moonshot AI released an open-weight model that independent benchmarks place near the frontier, and the Nasdaq dipped 1% as chip stocks sold off. That is a market responding not to a new capability but to a shift in the terrain of competition — specifically, the prospect that the monopoly rents Nvidia and its customers have been collecting on frontier AI might now be eroded by a freely available substitute.

The discourse that follows is instructive. David Sacks blames US regulation for tying American capital’s hands. Travis Kalanick accuses the Chinese of “distilling” from US models — a complaint that only makes sense if one assumes American training data is private property by natural right, rather than scraped from the same public internet. Dean Ball, formerly of the Trump administration, openly proposes manufacturing regulatory risk — FUD — around Chinese open models, not because they are dangerous but because their existence threatens the business model of proprietary AI. The state would be deployed not to protect citizens but to protect a rate of profit.

The real contradiction is not between open and closed source. It is between the social character of AI development — which is increasingly collaborative, distributed, and reliant on shared datasets and architectures — and the private appropriation of its most valuable outputs. Open-weight models expose this tension: they make it harder for any single capital to capture the full surplus. Ball’s “AI communism” nightmare is simply the logical endpoint of a technology that resists enclosure. The Chinese state, for its part, permits open release because it serves its own accumulation strategy — global influence, talent attraction, ecosystem dominance. It will restrict the same models the moment they threaten its own control. Neither side is principled. Both are manoeuvring within the same contradiction.

All the EVs that were discontinued or killed off in the U.S. this year¶

Source: TechCrunch

The Honda Prologue’s death certificate reads like a case study in how state policy reshapes industrial geography. The $7,500 federal tax credit’s removal in autumn 2025 didn't just dent demand — it reconfigured the viability calculations of every automaker selling into the US market. Prologue sales halved from 39,000 units in 2025 to a free fall afterward, a trajectory that tracks almost perfectly with the credit's expiration. But the credit is only the most visible lever.

Tariffs on Chinese components and finished vehicles are doing the heavier structural work. Polestar’s effective expulsion from the US market under the Chinese-connected vehicle ban shows how geopolitical competition now directly determines which commodities can circulate where. Geely-owned Polestar couldn't get Commerce Department authorisation; Volvo, its sibling under the same Chinese parent, did. The distinction isn't technological — it's political, and it fragments what was once a single global supply chain into nationally-bounded circuits.

The list of discontinued models reads less like consumer preference shifts and more like a forced rationalisation of overcapacity. Honda killed three planned EVs and the Prologue simultaneously, blaming both tariffs and Chinese competition — two pressures that push in the same direction: make fewer cars, concentrate production in tariff-sheltered sites. Hyundai dropped the Ioniq 6 (made in South Korea) while keeping the Ioniq 5 and 9 (made in Georgia). The factory location, not the product, determined survival.

Tesla’s decision to end the Model S and X is the exception that proves the rule. Those vehicles weren't killed by tariffs or tax credits — they were cannibalised by Tesla’s own cheaper models and by a strategic pivot toward autonomy as a revenue stream. The assembly lines are being ripped out for Optimus robots. Here, the contradiction is between two forms of fictitious capital: the old model of selling luxury hardware at a margin, versus the new promise of selling software-defined transport services. The S and X lost because they belonged to the wrong accumulation strategy.