2026-07-31 Observatory briefing¶

UK economy to flatline, warns Bank of England¶

Stagnation in a core imperialist economy confirms the organic crisis — growth is absent without speculative stimulus, validating the thesis of exhausted productive forces.

Tanker Market: Saudi Arabia's New Strategy Shifting Crude Volumes¶

Source: Hellenic Shipping News

The Houthi threat against Saudi port calls has exposed the circular logic at the heart of Riyadh’s energy strategy. Having spent two years rerouting crude through Yanbu to escape the Strait of Hormuz, Saudi Arabia has simply relocated its vulnerability rather than resolved it. The East-West Pipeline was a geopolitical hedge that has now become a target in its own right; the Red Sea terminal’s rise from alternative outlet to cornerstone of export policy means the Houthis can threaten the kingdom’s entire Asian customer base with a single declaration, no blockade required.

The traffic data confirms the mechanism: a 30% single-day drop in Bab el-Mandeb crossings shows how quickly the threat premium reasserts itself. The $1 million per transit cost of Cape diversions is not a market inefficiency but the price of doing business when the security of the route depends on the forbearance of a non-state actor. Saudi Arabia’s shift has not eliminated risk; it has merely changed which chokepoint the global oil trade must price in.

What is striking is how little this reshuffling alters the underlying distribution of power. The kingdom’s customers in China, India and South Korea now absorb nearly 70% of Yanbu’s exports, up from under a quarter when Egypt dominated. The geography of the trade has changed, but the dependence on a single, fragile corridor remains — only now it is one where the threat actor has explicitly named Saudi infrastructure as legitimate. The diversification is real, but it is diversification within a system where the ultimate guarantee of supply remains military and political, not commercial. For shipowners and charterers, the calculus is simple: the risk premium has not disappeared, it has just found a new address.

Can capitalism save Cuba's revolution?¶

Source: Le Monde Diplomatique

The care home in Vedado is the revolution’s new face: a former paediatrician and an IT specialist running a private business because the state can no longer provide. That this is presented as progress tells you how far the ground has shifted. Havana’s liberalisation of 2021 was not an ideological opening but an emergency measure, a bid to prevent collapse under the combined weight of the US embargo, Trump’s coercive measures, and the aftermath of currency reform. The result is a hybrid that satisfies no one: private firms now employ a quarter of the workforce, yet the political structure remains untouched, and the visible fruits of the market — the renovated turquoise house, the minimart — sit alongside children begging and empty luxury hotels.

The contradiction is not between socialism and capitalism as abstract systems, but between the state’s need to survive and its need to control. Every concession to the market is a concession to a class that the revolution was built against. The new entrepreneurs are not a bourgeoisie in waiting; they are a survival mechanism, licensed and revocable. But the licence cuts both ways. Washington’s tightening vice — the demand that Venezuela cease oil deliveries, the threat of tariffs on any country supplying hydrocarbons — is designed to force a choice between economic opening and political opening. The regime has bet that it can have the former without the latter. That bet may be the most fragile thing on the island.

Donald Trump's Latin American proconsuls¶

Source: Le Monde Diplomatique

The relationship Trump cultivates with Latin America’s new far right is not the old gunboat diplomacy, though Espriella’s demand for "Plan Colombia 2" and the return of US bases suggests the White House would happily resurrect its trappings. What has changed is the mechanism of subordination. Washington no longer needs to orchestrate coups against a threatening left; it can simply endorse the electoral victory of men who hold US passports and pledge fealty before taking office. Espriella’s acquisition of US citizenship in 2023, after a decade in Florida, is not a biographical curiosity but a structural fact: the region’s ruling class is now so integrated into the imperial centre that formal loyalty replaces the need for coercion.

The article’s framing of a "weaker left" is the crux. Petro’s failed "total peace" strategy was not merely a policy error; it was the political expression of a left that had absorbed the state’s security apparatus without transforming it, leaving a vacuum that organised crime and the far right have filled. The fivefold increase in coca cultivation since 2013 is the material index of this failure — a crop economy that thrives precisely because the state’s anti-drug war has become a revenue stream for armed groups, not a suppression of them. Espriella’s promise to crush cartels with US and Israeli help is a fantasy, but a functional one: it justifies the militarisation that will deepen the very conditions of violence it claims to end.

The deeper point is that inter-imperialist rivalry has collapsed into a single, uncontested hegemony in the hemisphere. With no countervailing power, Washington’s "neocolonial policy" operates through the willing collaboration of local elites who have internalised its interests as their own. The far-right tide is not a spontaneous eruption but a managed transition, one that suits the White House precisely because it no longer needs to intervene to secure its backyard.

The Limits of Asia's Two Major Growth Models¶

Source: Project Syndicate

The argument that China and India’s growth models are sui generis, and therefore unrepeatable, quietly concedes more than it states. Both economies are treated as discrete experiments in industrial policy, with their respective successes and bottlenecks attributed to internal coordination problems. But the framing of emulation misses the material basis of their ascent: neither model functioned in isolation, and both were profoundly conditioned by their insertion into a global division of labour that has since shifted.

China’s model rested on a unique fusion of state-directed credit, massive infrastructural absorption, and access to Western markets at a specific historical juncture. India’s service-led path, by contrast, was always more dependent on global capital flows and a domestic elite whose consumption patterns were tethered to imported technology. The authors’ focus on structural limitations — the coordination failures, the demographic pressures — is accurate but partial. What they do not name is that the very fragmentation of the global economy they cite as context is itself the product of overaccumulation in the advanced economies and the resultant inter-imperialist rivalry, particularly between Washington and Beijing. The window for a late developer to replicate China’s export-led industrialisation has closed not because of poor policy design, but because the absorptive capacity of the core economies has contracted.

The real lesson for other countries is not that industrial policy is futile, but that the terrain on which it must be executed has changed. The coordination problems are real; they are also secondary to the question of whether any peripheral state can now secure the kind of developmental space that China and India enjoyed. That space was a product of a specific balance of forces, not a template.

What France's far right has learned from Israel's apartheid system¶

Source: Le Monde Diplomatique

The comparison to Mussolini flatters the French far right. What Bréville and Rimbert sketch instead is a model of permanent, legalised subordination — the Israeli system — which offers something fascism never could: a stable architecture of exclusion that does not require the suspension of democratic forms. The RN’s trajectory under Marine Le Pen has been precisely this: a drive to govern, not to seize power in a rupture. The party has learned that segregation can be administered.

The argument hinges on a distinction between spectacular authoritarianism and bureaucratic domination. Israel’s regime, as described here, does not need to declare a state of emergency to maintain Jewish supremacy; it operates through a dense web of residence permits, military courts, zoning laws and differentiated citizenship. This is the template that attracts a far right seeking legitimacy. The French version would not be a police state in the classical sense but a state that classifies its population by origin and disciplines the "alien" part through administrative means — the citizenship-stripping proposal under Hollande being an early, bipartisan rehearsal of exactly this logic.

The deeper point is that this slide is compatible with, even functional for, capital. The left’s fixation on fascism as an imminent rupture obscures how the RN’s project dovetails with the existing order: a reserve army of labour kept precarious and legally segmented suits employers as much as it suits the party’s base. The warning is not that France is about to become 1930s Italy, but that it is quietly assembling the administrative machinery for a two-tier society — and that the left, by aiming at the wrong target, is helping to normalise the real one.

WestJet flight attendants threaten to strike¶

Source: FlightGlobal

The dispute turns on a narrow technicality with a wide material base. Credit-hour pay — compensation only for airborne time — is the industry standard precisely because it shifts the cost of ground operations onto the workforce. Boarding, turnaround, delays: all unpaid labour that the carrier pockets as margin. WestJet’s claim that higher hourly rates compensate is an accounting fiction; the union’s demand for boarding pay simply asks the airline to price labour that already exists.

The timing is the tell. A strike threat at the height of the summer season is leverage, and WestJet’s immediate lockout notice is the counter-move. The airline is not negotiating over wages so much as over who absorbs the cost of the industry’s operational rhythm — the idle hours that are structurally unavoidable but conveniently unremunerated. That WestJet frames its position as “sustainable for our company” while demanding attendants work unpaid is the contradiction stated plainly.

The Air Canada precedent matters. Its 2025 strike won ground pay at 70% of hourly rates by 2028 — a phased concession that shows the employers’ front is not monolithic. When one major carrier breaks rank, the standard shifts for the whole sector. WestJet’s resistance is a rear-guard action against a norm already in motion across North American carriers. The lockout notice, then, is less about this contract than about holding the line before the next one.

For passengers, the flexibility offered through 4 August is a hedge, not a solution. For the airline, every day of disruption during peak revenue season is a calculation: whether the cost of conceding boarding pay exceeds the cost of a grounded fleet. The arithmetic will decide, as it always does.

Boeing completes 737 Max 10 certification flight testing¶

Source: FlightGlobal

The completion of Max 10 certification flight testing is a milestone Boeing badly needed, but the framing obscures the more telling detail: this certification is happening nearly a decade after the Max was grounded, and it is being delivered through design patches rather than a clean-slate fix. The anti-ice turbulators and the simplified AoA fault display are retrofits bolted onto a frame that was already in crisis. The AoA system, in particular, is a direct admission that the original architecture — the one that contributed to two fatal crashes — was fundamentally flawed in how it presented information to pilots. Boeing is not certifying a new aircraft; it is certifying a series of corrections to an old one.

The commercial logic is equally revealing. Boeing has been bleeding market share to Airbus and, increasingly, to Comac's C919, which just completed its own high-altitude maiden sortie. The Max 10 is not a response to demand so much as a defensive play to hold the narrowbody duopoly together. Certification before year-end matters less for the aircraft's technical merits than for Boeing's ability to convert its backlog into deliveries and cash flow. The 2,060 flight hours across three test aircraft represent a significant capital outlay that only pays off if the type enters service and starts generating revenue. Every quarter of delay compounds the cost of capital already sunk into a programme that has been running at a loss since 2019.

The real tension is between Boeing's need to project stability to regulators and customers, and the structural reality that its production system is still recovering from the quality-control failures that halted deliveries. A certified Max 10 does not resolve the underlying question of whether Boeing can build these aircraft at scale without another incident. The FAA's involvement in the certification process, after years of delegated authority, signals that the regulatory relationship has shifted — but whether that shift is durable or merely performative remains to be seen. For airlines, the Max 10's entry into service next year will be a test of whether Boeing's promises of reliability can survive contact with actual operations.

Who will pay for the FAA's $7bn altimeter mandate?¶

Source: FlightGlobal

The FCC’s rebate programme is a rare moment of regulatory clarity about who bears the cost of a technological transition, and the answer is revealing: the cellular industry will hand over up to $5.71 billion to US operators, but foreign carriers flying into the country are simply out of luck. That asymmetry is not an oversight. The FCC has jurisdiction over US airwaves and the companies that lease them, so it can compel AT&T and Verizon to fund modifications for American operators. It has no such lever over, say, Lufthansa or Singapore Airlines, which will have to absorb the $80,000–120,000 per unit cost themselves or pass it to passengers on US routes.

The deeper structure here is the state mediating between two fractions of capital that both claim a public resource — spectrum — but have wildly unequal bargaining power. The cellular providers are being forced to pay because they are the newcomers seeking access to the Upper C-Band, and the FAA has effectively granted the aviation industry a veto over their expansion plans. The rebate is the price of admission. Yet the burden is not evenly distributed within aviation either: the FCC’s cap of $3.5 billion for “all other US operators” suggests general aviation and cargo carriers may face shortfalls, while the airlines’ $2.2 billion allocation is more secure. The 2030 deadline for Part 121 carriers versus 2034 for everyone else further entrenches the primacy of scheduled passenger aviation.

What remains unsettled is the auction timeline. The FCC must sell the bandwidth by July 2027, and rebates flow only after that. If the auction underperforms — a real possibility given the capital intensity of 5G buildouts — the funding could shrink, leaving operators with a mandate but no reimbursement. The FAA’s rule is absolute; the money is contingent. That gap between regulatory obligation and financial provision is where the real politics of this transition will play out, and it is unlikely to resolve in favour of the smaller operators who lack the lobbying weight of the major airlines.

The AI Boom and the Future of Finance¶

Source: Project Syndicate

The framing of the AI bubble as a moral outrage — mass layoffs, copyright theft, environmental destruction — obscures the more uncomfortable truth that its collapse would not resolve any of these harms. The $27 trillion in market value accrued since November 2022 is not a measure of productive capacity but of fictitious capital: a claim on future revenues that remain speculative, with investments increasingly circular and debt-financed. When the claims are called in, the pain will be real and concentrated, but the underlying transformation will continue because it is not driven by the bubble itself.

The distinction matters. A bubble is a financial phenomenon; the AI build-out is a material one. Data centres, chips, and energy infrastructure are being laid down regardless of the valuations attached to them. The circularity of investment — AI companies buying each other's compute, cloud providers capitalising each other's revenue — is a hallmark of overaccumulation, where capital circulates within the financial sphere without ever grounding itself in expanded reproduction. But the physical assets being installed will outlast the financial instruments that financed them. When the credit cycle turns, the infrastructure remains, often at fire-sale prices, ready to be absorbed by whoever survives.

This is where the conventional "bubble versus revolution" dichotomy fails. The dot-com comparison is instructive precisely because it is not analogous in the way the optimists claim. The internet's productive transformation was real, but it was also a vehicle for massive capital destruction that reshaped the ownership structure of the new economy in favour of a handful of survivors. The same dynamic is likely here: the collapse will not end AI, but it will determine who controls it. The debt-financed nature of the current boom means the losses will be socialised through the financial system while the assets are concentrated in fewer hands.

The environmental damage is not an externality to this process but a condition of it. Data centres consume water and power at a scale that assumes indefinite growth, and the destruction of books and copyright violations are not collateral damage but the raw material of the model itself. A crash that halts new investment would slow the damage, but the existing infrastructure continues to consume and extract. The question is not whether the bubble bursts, but whether the underlying accumulation strategy survives it — and in whose interest.

Investors love AI, as long as you're a cloud host¶

Source: TechCrunch

The market has settled on a peculiar division of labour. Amazon, Microsoft and Google are rewarded for pouring hundreds of billions into data centres, while Meta is punished for doing the same without a commensurate revenue line. Investors are not actually betting on AI's ultimate utility; they are betting on the tollbooth. AWS's 37% revenue growth and the 10% share pop suggest the financialised logic is simple: whoever controls the physical substrate can charge rent on every speculative venture built atop it, regardless of whether those ventures ever turn a profit.

But this is a circular arrangement, and the article half-acknowledges it. Amazon's hosting revenue is Anthropic's operating cost — the same money circulating through the system, extracted at one point and injected at another. The $173 billion in property and equipment spending, financed partly by drawing down cash reserves, represents a bet that the demand curve will keep shifting right faster than the depreciation schedule. That is the crux of the "$3 trillion question": the entire edifice rests on whether end-user demand materialises to justify the intermediate layer of AI labs and startups who are, for now, the cloud hosts' actual customers.

The contradiction is not between spendthrift tech firms and cautious investors. It is temporal. Data centres take years to build and monetise; capital markets demand quarterly vindication. Amazon's negative free cash flow is the moment where those two clocks diverge. The company is effectively borrowing from its own balance sheet to finance infrastructure whose payoff depends on a demand profile that has not yet been proven. If the AI bubble deflates, the tollbooth loses its traffic. If it inflates further, the hosts capture outsized rents. Either way, the risk has been socialised upward into the capital markets while the rewards accrue to a handful of vertically integrated firms. The real question is not whether investors love AI, but how long they will keep mistaking a circular flow of capital for genuine value creation.

AI hedge fund Situational Awareness may have sold its public portfolio, but it still has its Anthropic shares¶

Source: TechCrunch

The collapse of Situational Awareness is a familiar story wearing an unfamiliar costume. A 25-year-old with no trading experience raises hundreds of millions on the strength of a thesis — that AI scaling demands a physical build-out of chips, power and data centres — and turns it into $45 billion in assets through leverage and a frothing market. The thesis was correct enough to make him rich; it was also correct enough to destroy him when the market decided that capital expenditure without near-term revenue was a problem. The AI infrastructure trade was never a bet on fundamentals. It was a bet on other people continuing to make the same bet, and leverage was the mechanism that converted that collective conviction into a forced unwind when it broke.

What is more interesting than the fund's demise is the shape of its aftermath. Citadel, which had already been accumulating similar AI infrastructure positions, bought the distressed holdings at a discount. This is not vulture investing so much as the consolidation phase of a speculative cycle: the leveraged true believers sell to the unleveraged patient capital that can afford to wait for the recovery they both expect. Griffin gets the assets cheaper; Aschenbrenner gets to keep his private stake in Anthropic, now valued at $5 billion against a fund that has shrunk to roughly $10 billion. The public markets punished him for overreach, but the private markets — where valuation is a matter of narrative rather than daily mark-to-market — still hold his windfall. The distinction between the two is the real story. Public equity markets forced a reckoning with the gap between AI capex and revenue; private markets, where Anthropic sits at a $965 billion valuation ahead of a rumoured IPO, have yet to hold that conversation. The losses were real, but they were realised in the public sphere. The gains remain notional, parked in an asset that may or may not convert when the IPO window opens. Aschenbrenner's fate will ultimately be decided not by his trading acumen but by whether Anthropic's public debut validates the private valuation — at which point the same market that just punished him will be asked to price the same thesis all over again.