2026-08-27 Observatory briefing

Export disruptions increasingly hurting tanker demand

Ref: O1 Item-ID: article-71ec87eb7e Source: Hellenic Shipping News

BIMCO’s Niels Rasmussen is describing a market where the physical commodity and the price mechanism have come unstuck from each other. Export volumes are down sharply — oil and heavy products off 5.7% year-on-year, clean products down 11.2% — yet dirty tanker freight rates have spiked and stayed high. The war-risk premium and the effective withdrawal of ships from the pool, through stranding and delay, have done what cargo volumes could not. Supply of tonnage, not demand for it, is setting the price.

The stock drawdown of over 500 million barrels is the hinge. Releases have been compensating for lost production, which is why the oil price has not yet run away. But stocks are finite, and OECD cover could fall to 70 days by late 2027. When the buffer is gone, the adjustment will be violent: higher prices, demand destruction, and then tanker demand falls anyway. The industry is being paid handsomely today for a disruption that is eroding the conditions of its own future earnings. Freight rates are high because the system is broken, and the repair of that system would normalise rates downward.

The two scenarios BIMCO offers are not symmetrical. In the “SoH open” case, demand recovers but fleet growth and restored productivity cap the upside for owners. In the “SoH closed” case, demand weakens as stocks deplete while new tonnage keeps arriving. Either way, the current elevated rates are a temporary artefact of friction. The political economy here is stark: the US-Iran war has not resolved the underlying overaccumulation in tanker capacity, it has merely deferred its expression. When the strait reopens or the stocks run dry, the surplus tonnage will reassert itself. The MoU signed on 17 June was always a fragile instrument, and its stalling leaves the market suspended between a peace that would depress rates and a war that eventually destroys its own demand base.

India Is Sticking With America—For Now

Ref: O2 Item-ID: article-f915e588c2 Source: Foreign Affairs

Modi’s restraint is not the product of a strategic epiphany about the virtues of the American alliance. It is the behaviour of a state that has spent three decades engineering a structural dependence on Washington and now finds the patron volatile. The "America plus" formulation the author offers is a euphemism for hedging under duress: New Delhi is diversifying its portfolio of partners because it cannot diversify away from the United States itself. The Quad is kept alive by Indian initiative, not American commitment; trade negotiations continue because India’s export model is hitched to the US market; the defence framework deepens even as Trump courts Pakistan and flirts with a "G-2" with China.

The material basis of this asymmetry is stark. China and Russia cannot absorb Indian goods and services at the scale the US market does, and neither can substitute for American technology transfers or the implicit security guarantee against Beijing. India’s ruling class understands that its accumulation strategy, built on services exports and a cautious opening to foreign capital, requires stable access to the American economy. Trump’s tariffs and secondary sanction threats attack that access, yet Modi cannot retaliate without undermining the very relationship his growth model depends on.

What could break this calculus is not elite assessment but mass politics. The article notes public opinion turning against Trump in 2025, with Rahul Gandhi weaponising the charge of surrender and farmers mobilised against agricultural concessions. Modi’s political capital is being spent to absorb American coercion, and that capital is finite. If Trump continues to humiliate India publicly, the domestic cost of restraint may exceed the cost of rupture. For now, the Indian state absorbs the shocks; the question is how many shocks the political system can absorb before the strategy becomes untenable.

The Impossible Middle East

Ref: O3 Item-ID: article-7a09ec7bcf Source: Foreign Affairs

The February 28 war has done what decades of regional rivalry could not: it has made the Gulf's dependence on Washington feel like a liability rather than an insurance policy. Trump's decision to prioritise Israel's campaign against Iran over Arab stability has forced Saudi Arabia, the UAE and Qatar to confront the uncomfortable arithmetic of their security arrangements. American bases make them targets; American protection no longer guarantees their safety. The authors note the Gulf's diversification efforts since 2011 were largely coercive gestures aimed at Washington. This time the shift is substantive, though constrained: no alternative security guarantor exists, and China's willingness to supply Iran with defence-industrial material while trading heavily with the Gulf exposes the limits of economic partnership as a substitute for military protection.

The war's uneven economic damage threatens to destabilise Gulf domestic politics. States hit harder by Iranian strikes on desalination plants and ports will face different pressures than those spared, and the UAE's resentment of its neighbours' muted condemnations suggests the conflict is already fraying intra-Gulf relations. OPEC+ faces strain as members' interests diverge along the axis of who can still export and who cannot.

The authors' proposed remedy has Gulf states engaging Iran directly, offering it a stake in regional security in exchange for restraint. This is not appeasement but recognition that containment has failed and Washington's reliability has diminished. The region's ruling classes are adapting to a multipolar security environment where the old patron-client relationship no longer functions. Whether Tehran accepts such overtures while fighting the United States remains the open question, but the very fact that Gulf monarchies are contemplating this path marks a structural shift in the region's political economy of security.

Households warned of ‘worse to come’ as energy bills rise

Ref: O4 Item-ID: article-9d709b453d Source: The Telegraph

Energy price rises squeezing household consumption are a concrete manifestation of the cost-price scissors phase of the crisis, driving the political superstructure's decay.

New! Chris dePloeg’s “The Exterminating Empire”

Ref: O5 Item-ID: article-ca74763851 Source: Monthly Review

The exterminatory impulse has never been an aberration of empire, a lapse into excess that sober statecraft would correct. Chris dePloeg's new Monthly Review Press book takes the genocide in Palestine as its starting point to argue that the drive to kill everything, including the imperial core's own conditions of existence, is constitutive of capital's relation to life. The slave plantations and scorched-earth campaigns were not primitive preludes superseded by more rational forms of exploitation; they were the template.

What gives the argument its edge is the insistence that omnicide operates at different scales simultaneously. The same logic that razed Vietnam's forests and depopulated Haiti's countryside now produces climate apartheid, where the ecosystems sustaining the majority are sacrificed so that accumulation can continue for the few. DePloeg's provocation is that this tendency cannot be contained. The drive to annihilate, once unleashed as a structural feature rather than a contingent policy choice, threatens the core along with the periphery. Imperial violence is not a tool that can be wielded selectively; it is a metabolic process that eventually consumes its host.

The interview with Adnan Husain frames this as a war against life itself, a formulation that refuses the comfort of economistic readings. This is not merely about primitive accumulation or resource extraction, though it includes both. The exterminating empire names a stage where the reproduction of life, human and non-human, has become an obstacle to be managed rather than a precondition to be secured. For those tracking the current crisis, the book's timing is pointed: Gaza functions not as an exception that proves humanitarian rules, but as the revealed form of what imperial management always was.

Qantas brings forward A380 phase-out to 2028

Ref: O6 Item-ID: article-fbef21287e Source: FlightGlobal

Qantas has decided its A380s are done four years earlier than planned, with the last of the superjumbos due out by 2028 and the older A330s beginning their exit this year. The airline frames this as fleet renewal, but the timing is doing more work than the announcement lets on. A380s were already hard to justify outside a handful of trunk routes; the pandemic-era grounding turned them into expensive monuments to a pre-2019 bet on hub-and-spoke density that never fully returned. Bringing the phase-out forward is less a strategic revelation than an admission that the aircraft's residual value is falling faster than the cost of keeping it flying.

The interesting move sits underneath: Qantas is compressing its widebody transition into a narrow window, which means the replacement orders have to land on time or the airline faces a capacity gap it cannot paper over with A321s. That dependency is the real exposure. Every major carrier is chasing the same next-generation widebodies from Boeing and Airbus, and the duopoly's delivery slots are already stretched by years of supply chain disruption. Qantas is betting its schedule on someone else's production line.

The A380's early retirement also quietly settles an argument about second-hand values. The superjumbo was always a bespoke asset with a shallow resale market, and as more airlines accelerate their exits, the scrap value and part-out economics will set the floor. For the leasing companies holding A380 paper, this is another markdown in a portfolio that has already absorbed several. The aircraft's fate was sealed when the secondary market failed to materialise; Qantas is just acknowledging the invoice.

Why A 20% Pay Raise Still Isn't Enough To Stop Royal Canadian Air Force Pilots From Leaving For Air Canada

Ref: O7 Item-ID: article-50aab94e56 Source: Simple Flying

The 13% Military Factor adjustment, pensionable and retroactive to April 2025, lifts a senior RCAF Captain to roughly $180,000 in total annual compensation. A senior Air Canada 777 captain now clears $360,000, and that gap has widened from $56,000 to $180,000 in five years. The headline 20% raise was never for pilots anyway; it applied to entry-level Private/Aviator pay, a sleight of hand that tells you everything about how the state frames this crisis.

The structural bind is the officer corps itself. Military pay is rigidly tied to rank, so raising pilot base pay means raising pay for every army logistics officer and naval administrator, most of whom have no private-sector suitor offering to triple their salary. Canada's workaround, a 2022/2023 restructuring that folded flying allowances into pensionable base pay, produced its own absurdity: Captains now out-earn Majors, and pilots refuse promotions because a stripe would cost them money. The state cannot even rationalise its own compensation hierarchy without colliding with the logic of the market it is trying to resist.

What the RCAF is really competing against is not Air Canada's balance sheet but the broader social settlement that says public servants should be decently paid, never enriched. A senator earns $138,867; a widebody captain earns three times that. Governments cannot break this norm without political scandal, so they tinker with allowances and service pay, hoping patriotism plus a pension covers the difference. It does not, and the $15 million sunk into training each fast jet pilot walks out the door after a decade.

The Gripen order adds a further twist: a new airframe will demand experienced pilots just as the retention pipeline leaks. Canada is not losing a wage war. It is losing a class war it refuses to acknowledge, where the state's legitimate monopoly on violence cannot compete with the private sector's monopoly on the pilot's labour power.

Why Airlines Can’t Get Enough Widebody Jets

Ref: O8 Item-ID: article-67f8a8be69 Source: Simple Flying

Boeing’s 777X programme has slipped again, and the ripple effect is now measurable in years, not quarters. Airlines that ordered the twin-aisle jet in the mid-2010s were promised delivery by 2020; the current timeline pushes first handovers to 2026 at the earliest. Every month of delay compounds a structural shortage that predates the pandemic but was sharpened by it. When carriers grounded fleets in 2020, they deferred orders and retired older widebodies early. The recovery in long-haul demand arrived faster than the production lines could respond, and the gap between the two is now the central fact of the market.

The shortage is not evenly distributed. Airbus has sold out its A350 production slots into the early 2030s, and its A330neo is effectively a stopgap that airlines are taking because they cannot wait for the next generation. Boeing, meanwhile, is still working through the certification and supply-chain problems that have dogged the 777X, and its 787 output remains constrained by the quality-control issues that halted deliveries in 2023. The two manufacturers, duopolists in a market with enormous barriers to entry, are both operating at the limits of their capacity, yet neither can simply build more. The bottleneck is not demand, which is abundant, but the physical and regulatory capacity to convert orders into aircraft.

For airlines, the consequences are concrete. Lease rates for used widebodies have climbed sharply, and carriers are extending the service lives of aircraft they had planned to retire. The secondary market has become a seller's market, with lessors holding significant pricing power. This is a classic overaccumulation problem in reverse: capital poured into new production capacity during the boom years, but the crisis interrupted the realisation of that investment, and now the capital is stuck in unfinished aircraft and unfulfilled orders. The airlines that hedged by ordering early are not necessarily better off, since their delivery dates have slipped along with everyone else's.

The wider implication is that the aviation industry's recovery is being shaped less by demand than by the productive forces that supply it. For the next several years, the constraint on long-haul growth will be manufacturing, not passenger numbers. That is an unusual position for an industry that has spent two decades worrying about overcapacity.

In the AI Gold Rush, the Cloud Wins

Ref: O9 Item-ID: article-4e1b8aaa29 Source: Project Syndicate

Amazon, Microsoft and Google have positioned themselves so that the speculative mania around AI becomes, whatever its eventual fate, a transfer of wealth into their balance sheets. The distinction between the model-makers and the infrastructure-owners is the decisive one. OpenAI and Anthropic burn cash chasing a business model; the cloud giants sell shovels, and they sell them to the miners and to everyone else alike. Even a collapse in AI valuations would leave the data centres standing, depreciated perhaps but still operational, still the physical substrate that any future iteration of the technology would require.

The investor panic over Google's negative free cash flow misses this. Capital expenditure on data centres is the construction of a toll road. The market treats it as a cost, but for the firms that can absorb it, it is the acquisition of a permanent rentier position. Meta and Nvidia, for all their profits, lack this structural advantage. Nvidia sells the picks, but it must continually innovate to keep selling them; the cloud firms own the mine itself. This is overaccumulation in its most literal form, capital sunk into fixed infrastructure that will outlast the boom that justified it, and the firms that own it will collect the rents long after the speculative froth has evaporated.

The realignment is geopolitical as much as economic. Whoever controls the physical infrastructure of computation controls the terms on which every other actor, state or corporate, can participate in the AI economy. The European Union's attempts to build sovereign cloud capacity are a political project, an effort to resist a dependency that is already largely locked in. The window for that resistance is closing, if it is not already shut.

Amazon just tripled its order of Nvidia chips over ‘surging demand’

Ref: O10 Item-ID: article-80b256b102 Source: TechCrunch

Huang's own framing on the earnings call does the analytical work for us. "If we had more compute, we could generate more profitable tokens, which results in more profit for all of the services." That is the logic of overaccumulation stated in its purest form: the circuit of capital can only be sustained by pouring ever more fixed capital into production, regardless of whether the underlying demand for the output justifies it. Amazon's tripling of its order, from 1 million to 3 million GPUs, is not a response to measured need but a defensive move within a competitive race where every hyperscaler fears being the one left without capacity when the next wave of model training arrives.

The $279 billion Nvidia has committed to secure supply and manufacturing capacity, up from $119 billion last quarter, shows the contradiction operating at both poles. Nvidia must lock in fabrication capacity years ahead, while Amazon must lock in chips years ahead, each treating the other's commitment as proof of demand. The result is a self-referential loop: the more both sides invest, the more credible the story of "surging demand" becomes, and the more both sides must invest to stay in the game. Amazon's simultaneous development of its own Trainium and Graviton chips, a $25 billion annualised run rate, is the hedging strategy of a buyer that knows it is overpaying but cannot afford to stop.

The $96.2 billion quarterly revenue and 117% data centre growth are the material expression of this dynamic. Whether the "profitable tokens" Huang promises will materialise at the scale required to service this debt-financed build-out is the question investors are circling, and the answer will determine which of these commitments become stranded assets.

Anthropic continues compute-gobbling streak in $45B deal with Nscale

Ref: O11 Item-ID: article-859f8c4f4e Source: TechCrunch

Anthropic has committed to roughly $45 billion in compute rentals from Nscale, a British infrastructure firm founded in 2024 that will supply Nvidia's Vera Rubin systems from a West Virginia data centre. The six-year deal follows a $10 billion agreement with Volta in Norway, a $5 billion deal with AMD, a $1.25 billion-per-month arrangement with SpaceX, and expansions of its Amazon, Google, and Broadcom partnerships. The company is spending its way toward parity with OpenAI, and the sums are staggering for a lab whose revenue remains a fraction of its capital burn.

The shape of the spending matters more than the total. Anthropic is renting access to chips from startups that have existed for months, while the data centres and hardware belong to third parties who are themselves borrowing heavily to build. The lab converts venture capital and strategic investment from Amazon and Google into long-term rental obligations, pushing fixed capital costs off its own balance sheet and onto infrastructure firms that must service debt with future compute sales. Capacity is being built not because demand is proven but because the competitive logic of the sector demands that no rival secure an exclusive claim on the next generation of chips.

The Vera Rubin system, six chips working in concert, is the material expression of this dynamic. Nvidia's product cycle now dictates the investment schedules of the world's most valuable companies, and every new generation renders the previous one obsolete, forcing labs to sign ever larger deals to avoid falling behind. The infrastructure providers are the transmission belt for this pressure, borrowing against contracts that may not be honoured if the AI bubble deflates before 2027, when the Nscale capacity comes online.