For most of the last three decades, the answer to the most dangerous question in Gulf energy security was a pipeline.

The question was simple enough to state and almost impossible to answer: what does Saudi Arabia do if the Strait of Hormuz closes? Roughly a fifth of the world's oil passes through a channel thirty-three kilometres wide at its narrowest, with Iranian territory along its entire northern shore. Every serious planning exercise since the Iran–Iraq war has treated its closure as the scenario that matters, and the Kingdom's answer was to build a second door. The East–West Crude Oil Pipeline, known as the Petroline, runs 1,201 kilometres from the Abqaiq processing complex in the Eastern Province across the peninsula to Yanbu on the Red Sea. Commissioned in 1981, expanded to five million barrels a day in 1992, thirteen pumping stations along its length, it exists for one purpose: to move crude to water that is not the Gulf.

That answer held for thirty-four years. It stopped holding on 11 September 2026.

In the eleven days between 1 and 11 September, both halves of Saudi Arabia's export system were disabled at once. The eastern route had already been shut since March, when Iran declared the Strait of Hormuz closed in the opening weeks of the war. The western route — the pipeline, the Red Sea, the port of Yanbu, the passage out through Bab al-Mandab — came apart in two stages. Houthi forces took the shore on both sides of Bab al-Mandab in the first ten days of the month and announced that navigation was safe for everyone except Saudi shipping. Then, on the night of 10–11 September, drones launched from Iraqi territory struck the Petroline in the Riyadh and Madinah regions, and Aramco shut the pipeline as a precaution.

A country whose entire energy security architecture rested on having two ways out now has neither.

The logic of the Petroline was never really commercial. Shipping crude from Abqaiq to Yanbu and loading it on the Red Sea coast costs more than loading it at Ras Tanura on the Gulf, and for most of the pipeline's life it ran well below capacity. It was insurance, and like most insurance it looked expensive right up until the moment it was needed.

The 2019 drone and cruise missile attack on Abqaiq sharpened the thinking considerably. That strike removed more than half of Saudi production for a matter of days and demonstrated that the Kingdom's vulnerability was not only the strait but the processing chokepoint behind it. The response was to harden Abqaiq, disperse storage, and lean harder on the western route as the answer to eastern risk.

When the war began in late February 2026 and Iran closed Hormuz in the opening weeks, the architecture did what it was designed to do. The Petroline was attacked in the spring and restored to full capacity by 12 April. Aramco then went further than the original design, converting parallel lines previously used for natural gas liquids to crude service and pushing throughput towards seven million barrels a day — an emergency adaptation that says a great deal about how completely the Kingdom's exports had been pivoted west.

For roughly five months, that worked. Saudi Arabia lost volume, lost margin and lost the Gulf, but it kept exporting. The hedge was doing its job.

The eastern door: closed, and quietly staying closed

The scale of the Hormuz closure has become easy to underestimate, because it has gone on long enough to stop being news.

Transits through the strait have fallen by roughly 95 per cent against pre-war norms. On 6 September, six ships made the passage. The normal figure is around eighty-five a day. Crude exports from the Gulf as a whole have dropped from about seventeen million barrels a day in 2025 to roughly nine million in August 2026, a fall of some 47 per cent. Iran declared the strait closed in early March, permitted a narrow category of vessels it considered friendly, agreed a memorandum with Washington on 17 June that briefly reopened it, and saw that arrangement collapse in early July after attacks on commercial shipping resumed.

Gulf states and Iran are reported to be weighing talks on reopening. Nothing has been agreed. The working assumption in the market — and, on the evidence of its own behaviour, in Riyadh — is that the eastern door stays shut for the foreseeable future.

Confidence in that assessment is high. The transit data is independently observable through vessel tracking and does not depend on any party's account of itself.

The Houthi advance to Bab al-Mandab was covered in this publication ten days ago and needs only summary here. Between 3 and 10 September, Ansar Allah forces took roughly 2,600 square kilometres of Yemen's Red Sea coast, moving through Hays, Al-Khokha, Mawza and Al-Wazi'iyah before entering Mocha on the morning of 10 September and taking the port and airport without a fight, the Saudi-backed forces holding the city having withdrawn overnight. They then landed on Zuqar, Hanish and Mayyun, the island inside the strait itself.

What matters for Saudi Arabia is not the territory but the sentence that followed it. Sanaa announced that navigation in the Red Sea was safe for all shipping except Saudi shipping.

That is a precise and deliberate formulation. It is not a blockade of the Red Sea, which would unite half the world against the movement holding the shore. It is a blockade of one country, framed as an act of restraint towards everyone else. And it is enforceable in a way that anti-ship missiles are not: Bab al-Mandab is about twenty kilometres across at its narrowest, and ordinary artillery on a headland reaches across it. Artillery on a headland is not an incident that can be intercepted, insured against or absorbed. It is a condition.

So the western route's maritime exit was compromised before the pipeline was touched. Yanbu could still load, but what loaded there had to run a strait held on both banks by a movement that had publicly exempted everyone except the cargo's owner.

On the night of 10–11 September, multiple drones struck the Petroline in the Riyadh and Madinah regions. The Saudi Ministry of Foreign Affairs condemned the attack, confirmed injuries and damage, and said repairs were in hand. Aramco shut the line as a precautionary measure. The Iraqi Prime Minister's office subsequently dismissed a military commander responsible for operations in Maysan province after investigations established that the drones had been launched from within that governorate. The umbrella grouping of Iran-aligned militias in Iraq denied involvement.

This was not the first time. On 27 July, drones struck both the Abqaiq facility — the world's largest crude stabilisation plant — and the East–West Pumping Station. The pipeline had been hit in the spring and repaired. The pattern across the war is consistent: the western route has been under sustained attack from the moment it became the only route.

Two points deserve emphasis, because they are what turn an incident into a strategic fact.

The first is the launch geography. These drones did not come from Yemen or from Iran. They came from southern Iraq, across a land border, against a pipeline running through the Saudi interior. The Kingdom's air defence architecture is built around threats from the south and the east. A third axis, from the north, against a linear target 1,201 kilometres long with thirteen pumping stations, is not a problem that can be solved by buying more interceptors. A pipeline cannot be hardened along its length. It can only be defended at the point where the attack is launched, which is territory Riyadh does not control.

The second is the choice of aim point. Striking a pumping station is not a spectacular attack. It does not burn for days on international television the way Abqaiq did in 2019. It is instead the cheapest available way to impose a precautionary shutdown on a system that has no substitute — and a precautionary shutdown is the objective. The attacker does not need to destroy the pipeline. He needs only to make operating it look reckless.

Saudi Arabia held two export routes, and treated them as independent. They were not independent, and the war has revealed why in three separate ways.

They shared a common origin. Both the Gulf route and the western route begin at Abqaiq. The stabilisation complex that feeds Ras Tanura also feeds the Petroline. A successful strike on Abqaiq — attempted on 27 July, achieved in 2019 — does not close one door or the other. It closes both, because it removes the crude before the crude reaches either.

They shared a common adversary with distributed reach. The closure of Hormuz, the seizure of the Bab al-Mandab shore and the drone launches from Maysan were executed by three different actors — the Islamic Revolutionary Guard Corps, Ansar Allah and Iraqi militias — operating in three different countries under no unified command that anyone has demonstrated. But they are aligned, and alignment is sufficient. Redundancy protects against independent failures. It does nothing against a single adversary who can reach every leg.

And they shared a common assumption: that the threat was maritime. The Kingdom planned for a strait, and then for a second strait, and built a land bridge between them. What it did not plan for was the land bridge itself being the target, struck from a direction that neither strait faces.

An engineer would call this a common-mode failure: two systems that appear redundant but fail together because they share a dependency the design did not account for. It is the reason nuclear plants do not site both emergency generators in the same basement. Saudi Arabia sited both of its emergency generators in the same basement, and the basement is the Arabian Peninsula's exposure to Iranian-aligned force projection.

What the numbers say, and what they say about time

The production figures published to OPEC give the clearest measure of what this has cost.

Saudi crude output fell by 1.9 million barrels a day in August, to 6.238 million. That is the lowest monthly figure the Kingdom has reported since 1990, and it is below the previous wartime low set in April. Before the war, output ran above ten million barrels a day. Exports fell to roughly 3.2 million barrels a day, the lowest in more than a decade. Brent has traded above $100.

The most revealing number is not any of those. It is that Saudi Arabia told OPEC its crude "supply to market" in August was 7.122 million barrels a day — nearly 900,000 barrels a day more than it produced.

You cannot supply more than you produce for long. The gap is inventory: crude drawn from storage to honour commitments that current output cannot cover. That figure is a clock. It tells you that the Kingdom has been buying time with stocks, that the stocks are finite, and that the strategy of absorbing the disruption while waiting for the strait to reopen has a term limit measurable in months rather than years. It also tells you that the headline production collapse understates the pressure, because the export figure has been propped up by drawdown.

Confidence here is moderate rather than high. The figures are self-reported to OPEC, as all member production data is, and the "supply to market" line is exactly the sort of number that carries definitional room. But it points in one direction, and the direction is consistent with the physical picture.

Five weeks before the pipeline shut, Saudi Arabia signed the most significant security agreement in its modern history.

The Mecca Joint Defence Agreement was concluded in Mecca on 7 August 2026 between Crown Prince Mohammed bin Salman, President Erdoğan and Prime Minister Shehbaz Sharif, and given the name Mecca Defence Alliance at a follow-up meeting in Istanbul on 31 August. Its central clause is the one that matters in any alliance: an attack on one is an attack on all. A secretariat has been established in Riyadh, with the first secretary-general to be drawn from Pakistan. On paper the combination is formidable — some 1.4 million active personnel, 3,400 aircraft, 6,000 tanks, more than 340 naval assets, and combined annual defence budgets of around $124.4 billion, of which Saudi Arabia's $63.9 billion is the largest share.

Between the signature and this week, Saudi Arabia has had a refinery struck, its oil production driven to a thirty-six-year low, a maritime embargo declared against its shipping by a non-state actor holding a strait, and its only alternative export artery shut by drones from a neighbouring state. The alliance has produced statements of condemnation.

It would be unfair to call this a failure of the agreement, and the fairer reading is more uncomfortable. The pact is doing precisely what it was designed to do, which is to deter a conventional attack on Saudi territory by a state adversary. No such attack has occurred. What has occurred is a campaign conducted below that threshold by actors who are not states, launched from territory whose governments disclaim responsibility, against infrastructure rather than population.

The attack-on-one clause has no obvious purchase on any of it. Who, precisely, has attacked Saudi Arabia in a way that triggers collective defence? Ansar Allah is not a state. The Maysan drones were launched from Iraq, but Baghdad dismissed the responsible commander, which is the behaviour of a government disclaiming the act rather than committing it. Tehran's fingerprints are visible throughout and legally nowhere.

This is the oldest problem in alliance theory, and the Mecca pact has run into it within five weeks. Collective defence guarantees are written against invasions. The coercion actually being applied to Saudi Arabia is calibrated, deniable and aimed at infrastructure — designed, one has to assume deliberately, to stay under every threshold that would activate anyone's treaty obligations. Deterrence by alliance requires an adversary who attacks in the manner the alliance anticipates. Nobody is obliged to oblige.

Three implications run outward from this, and none of them is confined to Saudi Arabia.

The first concerns how energy security is planned everywhere. The Saudi case is the most expensive demonstration yet that route diversification is not the same thing as risk diversification. Two routes that terminate in different oceans still failed together, because the analysis behind them counted geography and did not count adversaries. Any state or company currently reassuring itself that it holds alternative supply corridors should ask a harder question than whether the corridors are physically separate. The question is whether a single actor can reach all of them, and whether the corridors share an origin point, a chokepoint, or a threat axis. Europe should ask this about its own post-2022 energy architecture with some urgency.

The second concerns the corridor debate that has dominated Eurasian connectivity planning for three years. The argument for overland routes between Asia and Europe has rested substantially on the unreliability of maritime chokepoints — Suez, Hormuz, Bab al-Mandab, Malacca. That argument has just been strengthened and complicated in the same week. Strengthened, because two of the four chokepoints named are now compromised simultaneously and the case for land alternatives has never looked better. Complicated, because the Petroline is a land route, and land routes turn out to be attackable too, by the same cheap aerial weapons, across the same borders, with the same deniability. A corridor is not safe because it is dry. It is safe if somebody can defend its entire length, and very few can.

The third concerns Riyadh's strategic position, which is now genuinely difficult. The Kingdom has lost its eastern export route to a state adversary, its southern maritime exit to a non-state one, and its overland artery to a third party operating from a fourth country. It has drawn down inventories to maintain supply. Its new alliance cannot be triggered by any of the methods being used against it. Its American security relationship has not restored the strait in six months of war. There is no version of this in which Saudi Arabia does not have to negotiate, and the only actor who can deliver all three of the files — Hormuz, the Houthis, and the militias in Maysan — is the one applying the pressure.

Five indicators will tell us which way this goes over the next quarter, and they are worth stating in advance so the assessment can be tested rather than merely asserted.

Whether the Petroline restarts, and at what throughput. A return to five million barrels a day suggests localised damage and effective repair. Sustained operation below that, or repeated precautionary shutdowns, indicates that the attacker has established a standing veto over the route, which is a far more serious condition than any single strike.

Whether the "supply to market" figure converges back towards production. If exports fall towards the 6.2 million barrel output level, the inventory buffer is exhausted and the Kingdom's room for patience has run out.

Whether Sanaa's Saudi-specific exemption holds. If Houthi enforcement stays narrow, the movement is running a coercive strategy aimed at a negotiation. If it broadens to allied or Western shipping, the calculation has changed and so has the risk to everyone else.

Whether any Mecca Alliance activity moves beyond declaratory. Joint air defence deployment to Saudi territory, a Pakistani or Turkish contribution to interior air defence, or a formal invocation of the attack-on-one clause would each mark a real shift. Continued condemnation without deployment confirms the reading above.

And whether Gulf–Iran talks on reopening Hormuz produce anything. Reporting suggests they are being weighed. If they begin and the pipeline attacks stop at the same time, the linkage between the pressure campaign and the negotiation will have been demonstrated rather than inferred.

There is an argument, which will be made, that none of this was foreseeable — that no planner could have anticipated a simultaneous closure of the Gulf, the seizure of the Red Sea's southern gate and a drone campaign from Iraq against a pipeline in the Saudi interior.

That argument is wrong, and it matters that it is seen to be wrong. Each of those three events was individually anticipated in the open literature. Hormuz closure has been modelled for forty years. Houthi capability against Red Sea shipping has been demonstrated repeatedly since 2015 and at scale since 2023. Iraqi militia drone strikes on Saudi territory are not novel. What was not anticipated was the combination — and the combination was not anticipated because redundancy was assessed against the failure of individual links rather than against an adversary with the reach to pull several at once.

The lesson is not that Saudi Arabia planned badly. By the standards of the discipline, it planned unusually well and spent accordingly. The lesson is that the discipline itself has been assessing the wrong thing. Infrastructure resilience has been treated as an engineering problem, measured in redundant capacity and hardened facilities. It is a strategic problem, and the right measure is not how many routes a state holds but how many of them one adversary can reach.

By that measure, Saudi Arabia held one route the entire time. It simply had two ends.

Saudi Arabia built two routes out of a war zone and lost both within eleven days. This assessment traces how a thirty-year hedge against the Strait of Hormuz collapsed into a single, indefensible point of failure.

The Saudi Kingdom Has No Way Out

On the morning of 10 September 2026 the Houthi forces, who call themselves Ansar Allah, entered the port city of Mocha on Yemen’s Red Sea coast and took the port and the airport without a serious fight, because the Saudi backed government forces holding the city had already withdrawn during the night. Three Yemeni government sources confirmed the loss to Reuters, and the government itself said nothing for most of the day, which in this war is usually the clearest confirmation of all.

Mocha was not an ordinary town. It had been in the hands of the anti Houthi coalition since early 2017, it was the main military and logistical base of Tariq Saleh’s National Resistance Forces on the western coast, and before that it had served as a base for the United Arab Emirates. Losing it removes the last significant government anchor on the Red Sea, and it sits between sixty and seventy five kilometres north of the Bab al-Mandab strait.

The fall of Mocha was the end of a campaign that had run for a week rather than an isolated event. Since 3 September the Houthis have attacked along several connected fronts at once in western Taiz and southern Hodeidah, and within roughly twenty four hours at the end of that week they took a coastal belt of about 2,600 square kilometres, which is around seven times the size of the Gaza Strip. The towns that fell in sequence were Hays, Al-Khokha, Al-Mafraq, Jabal al-Nar, Mawza, Al-Wazi’iyah and finally Mocha itself, together with the Khalid ibn al-Walid military camp, and by the end of the day the coastal town of Dhubab had gone as well, which is the last settlement on the Yemeni shore before the strait.

Then the campaign went to sea. Houthi units carried out missile and rocket strikes followed by boat borne landings and took Jabal Zuqar island, and reports through the day placed further landings on the Hanish islands, both Greater and Lesser, and on Mayyun, which is better known as Perim and which sits inside the strait itself. Iranian state linked outlets claimed outright that Ansar Allah now controls Bab al-Mandab. That claim runs ahead of what has been confirmed, and the situation on the islands was still unsettled at the time of writing, but the direction is not in doubt.

At the same time the Houthis have been striking Saudi Arabia directly. Satellite imagery from Sentinel-2 confirms that four oil storage tanks were destroyed at Aramco’s Jazan refinery, with three more at the Jazan bulk plant and five at the Abha bulk plant, and at least fourteen tanks of various sizes have been destroyed at the Jazan refinery alone since July. Missile and drone attacks have reached Abha, Jazan, Khamis Mushait and Najran, wounding dozens of people and forcing the temporary shutdown of several energy facilities.

2. The battlefield, and how it moved so quickly

The campaign has a clear geographic logic and it was fought in a sequence rather than as a scattering of separate attacks.

The first move was inland, against the mountain belt west of Taiz, where Houthi forces attacked Maqbanah, Al-Kadaha and Jabal Habashi with heavy preparatory fire and drone support, and captured Jabal Sheikh Saeed and the remaining Qahban positions. The purpose of taking those heights was not the ground itself but the road, because whoever holds that belt controls the route between Taiz and the sea. The second move ran down the coast from the north, where Hays and Jabal Dubas were fought over for about forty eight hours before the defence broke, and once Hays fell the coastal road to Al-Khokha and then to Mocha lay open. The third move joined the two, running from Mawza and Al-Wazi’iyah in the interior down to the coast, which cut Mocha off from the inland positions that were supposed to support it.

What turned steady pressure into a collapse was not firepower alone. Government units withdrew from Al-Khokha, from Al-Mafraq and from Jabal al-Nar before the fighting was decided, and the same happened at Mocha in the early hours of 10 September, with footage showing government forces leaving before Houthi fighters arrived. In several places the line did not break under assault, it simply moved.

Two further points deserve recording because they explain the speed. The first is that the anti Houthi camp was already divided before the offensive began, since the Emirati policy of backing the Southern Transitional Council against Saudi aligned forces has left the coalition politically fragmented and operationally uncoordinated, so there was no single commander able to hold a line that ran across several different forces. The second is information warfare. During the offensive an audio recording appeared that was made to sound like Lieutenant General Tariq Saleh announcing a withdrawal, and it was amplified to an audience of around half a million people before the National Resistance denied it and said the recording had been fabricated. Whether or not that single recording changed any decision, its existence tells us that the Houthis are now fighting for the morale of the opposing side as deliberately as they fight for ground.

The government response has been real but late. The Giants Brigades and the Nation’s Shield forces have been sent into Taiz and southern Hodeidah, Major General Hamdi Shukri of the Fourth Military Region and Major General Abdulrahman al-Lahji of the Nation’s Shield have been moved to Mocha to run coastal operations, and government aircraft have struck Houthi targets from Hays down to Al-Kadaha. Saudi aircraft flew about forty strikes on Taiz, Hodeidah, Al-Jawf and Marib on 10 September according to Houthi sources, and the government has tried to relieve the pressure by opening new fronts of its own in Al-Jawf, Marib and Al-Bayda. None of this has yet stopped the movement on the coast. At least 917 families have been displaced from Maqbanah and Jabal Habashi alone.

3. How the Houthis fight, and what this success actually means

The Houthis do not fight the war their opponents keep preparing for. They combine four things, and the combination rather than any single element is what makes them difficult.

They use tempo. The offensive did not pause between objectives, and each captured position was used immediately as the starting line for the next attack, which never gave the defence the day or two it needed to organise a new line. They use layered fire, meaning artillery, rockets, drones and missiles used together against the same position, so that a defender who can handle one of these has to handle all of them at once. They use small boats, and the island landings show that a force with no navy at all can take islands from a coalition that has warships, provided it moves at night, moves quickly and picks targets that nobody has bothered to garrison properly. And they use politics as a weapon, through fabricated recordings, through announcements timed to arrive before the fighting is settled, and through offers such as the statement from Sanaa that navigation in the Red Sea is safe for everyone except Saudi shipping.

That last statement is worth reading closely, because it is the whole strategy in a single sentence. It tells the shipping industry that the Houthis are a reasonable authority that can be dealt with. It tells Riyadh that it alone will pay. And it quietly assumes something that no government has yet conceded, which is that the Houthis are now the party that decides who may pass.

The wider meaning of the past week is therefore not really about Mocha. It is that a movement which began as a highland insurgency in the far north of Yemen has become a coastal power with a hand on one of the world’s most important shipping routes, and it has done this while the state that has been fighting it for eleven years watches from the air. The Houthis have also shown that they can raise support well outside their own Zaydi Shia base, because the coastal population of Tihama is largely Sunni and the advance did not stall in hostile country. Some of that is coercion and some of it is the ordinary tribal arithmetic of a place where alliances follow whoever is winning, but the effect is the same.

The response so far has three parts, and each of them has a visible weakness.

The first part is air power, and here the striking fact of the past week is how little of it there has been. Saudi Arabia holds one of the largest and most modern combat air fleets in the region, and yet government forces on the ground appear to have been fighting without reliable Saudi air cover and without American air cover, which does more damage to morale than to the order of battle. The wider problem is that air power did not defeat the Houthis between 2015 and 2022 either, and the reason was never the quality of the aircraft. What decides an air campaign is the joining together of intelligence, target selection, strike, damage assessment and coordination with the troops on the ground, and that has been the weak point of the Saudi effort from the beginning. The Houthis of 2026 are also a harder target than the Houthis of 2015, because they have far more combat experience and considerably better air defences, so sustained air operations against them now carry a real cost in aircraft and crews.

The second part is reinforcement, and the Giants Brigades and Nation’s Shield are genuine formations with real fighting records. The difficulty is that the anti Houthi camp is a coalition of forces that answer to different masters, since the National Resistance answers to Tariq Saleh, the southern forces answer to the Southern Transitional Council and through it to Abu Dhabi, and the Presidential Leadership Council has never truly commanded any of them. Reinforcement only produces recovery if these units can be made to operate as one force in one theatre, and that has not been demonstrated.

The third part is alliance building, and this is the newest element. Saudi Arabia, Pakistan and Türkiye finalised a mutual defence pact in August 2026, known as the Mecca Defence Alliance, and Pakistan has now warned Iran that continued Houthi attacks on Saudi territory could bring that pact into operation. This is a serious development on paper. In practice it is hard to see Islamabad or Ankara sending ground forces into Yemen, and the one power that did successfully mount an amphibious operation in this war, the United Arab Emirates when it retook Aden in 2015, is now estranged from Riyadh and is very unlikely to repeat it.

There is also a timing problem that has no good answer. Saudi Arabia will be tempted to launch a ground offensive to reverse the losses, but rebuilding Yemeni government forces into a force capable of that would take months, and during those months the Houthis in the south and the Iran aligned armed groups in Iraq to the north could do very serious damage to Saudi energy infrastructure. Saudi crude production last month fell to 6.238 million barrels a day, which is the lowest figure since 1990 and about 1.9 million barrels a day below recent levels. A country that is already producing less than it has at any point in thirty six years is not in a strong position to invite more strikes on its refineries.

5. What this means for the regional balance

The honest way to describe the past week is that the map of pressure in the Middle East has changed more than the map of Yemen.

Since the closure of the Strait of Hormuz to normal traffic earlier this year, Saudi Arabia has moved a large part of its crude exports to the Red Sea through the East-West pipeline to Yanbu, precisely in order to avoid the Gulf. That decision made commercial and military sense at the time. It also meant that the Red Sea route stopped being an alternative and became the main road, and the Houthi campaign of the past week has been aimed at exactly that road. If Bab al-Mandab comes under reliable Houthi control, then both of the exits from the Gulf region are subject to the same coalition, and the alternative that was created to escape the first chokepoint runs through the second.

The geography makes this worse than it sounds. The Strait of Hormuz is about fifty kilometres wide at its narrowest point, while Bab al-Mandab is about twenty kilometres wide and is divided into two channels by Perim island. A force holding the Yemeni shore and the islands does not need anti ship missiles to interfere with traffic, because ordinary artillery reaches across that distance. That is a fundamentally different kind of threat from missile and drone attacks, which can be intercepted, deterred and insured against. Artillery on a headland is a permanent condition rather than an incident.

For Iran the value of this is obvious and does not depend on any single ship being hit. Tehran is under severe economic pressure and is fighting a war it did not choose the timing of, and its ability to raise the cost of that war for others is now its main currency. Holding leverage over two chokepoints at once gives it something to trade in any negotiation with Washington, and it gives it that leverage without needing to use it. As Yemen’s ambassador to Qatar put it this week, a grip over these two points would be worth more to Iran than any nuclear arsenal, and while that is a partisan statement it is not a foolish one.

It would be wrong, however, to treat the Houthis as simply an Iranian instrument. They receive weapons, training and targeting help from the Islamic Revolutionary Guard Corps, and their leaders speak openly about the relationship, but they are also fighting their own war for control of Yemen and they have their own reasons for wanting the coast. The useful way to think about it is that Iranian and Houthi interests overlap almost completely at the moment, which makes the question of who is directing whom less important than it looks. What should worry planners is a further point, which is that the Houthis are a working model of what an armed movement can achieve without a state behind it, so even a serious weakening of Iran would not necessarily remove the pressure on the Red Sea.

Their immediate aim has been achieved, which was to take the whole Yemeni Red Sea coast and the approaches to the strait, and to cut Taiz off from the sea. Taiz is now supplied by a single remaining route.

The larger aim is recognition. Everything in the Houthi public position points in the same direction, from the foreign ministry statement in Sanaa saying that Yemen poses no threat and seeks relations based on mutual respect, to the offer to join a regional security arrangement and help protect maritime navigation alongside the other coastal states, to the notice that shipping is safe for everyone except Saudi vessels. These are the statements of a movement that wants to be treated as the government of Yemen and is using the shipping lane as the argument. Control of Bab al-Mandab is the strongest bargaining position any Yemeni actor has ever held, and it is not being collected in order to be given away.

The immediate question is whether they will now push further south towards Aden, which is the temporary seat of the internationally recognised government. Their own messaging has suggested they will stop at the old border between North and South Yemen and negotiate with the southern forces from a position of strength, and most experienced observers expect that, since taking Aden would unite the southern factions against them and turn a fast campaign into a long occupation. That expectation may hold. It is worth remembering, though, that a week ago the same reasoning said that Mocha would be defended.

7. Energy, shipping and the wider consequences

Between ten and twelve per cent of global trade passes through Bab al-Mandab in normal times, and the strait is the only sea route between the Mediterranean and the Indian Ocean that does not require the long passage around the Cape of Good Hope. The consequences of losing confidence in it fall into several layers.

The first is price. Brent was trading around 101 dollars a barrel on 10 September, having risen above 100 for the first time since May, and the market is already pricing in disruption at both chokepoints rather than one. Analysts at Goldman Sachs have identified more intense attacks in Hormuz and the Red Sea as the most likely cause of further increases and expect the disruption to run into 2027. The important point for European governments is that this is not a spike caused by a single incident, it is a structural change in the cost of moving energy, and structural changes do not reverse when the news cycle moves on.

The second is insurance and routing. War risk premiums for the southern Red Sea will rise again, and some owners will simply stop using the route, which adds roughly ten to fourteen days and a great deal of fuel to every voyage between Asia and Europe. That cost lands on European consumers through goods prices rather than through the oil price, and it lands on Egypt directly, since Suez Canal revenues have already been badly damaged by the attacks that began in 2023 and a further collapse would hit an economy that can afford it least.

The third is the exposure of Saudi Arabia itself. The whole point of the East-West pipeline and the Yanbu terminal was to give the kingdom an exit that did not depend on the Gulf. That exit now runs down a sea whose southern gate is held by the movement currently bombing its refineries, and the refineries themselves are being destroyed tank by tank. This is the reason Saudi production has fallen to a thirty six year low, and it is why Riyadh has no good options: acting decisively invites more strikes on energy infrastructure it cannot fully protect, while not acting concedes the coast.

The fourth is the effect on the countries that are usually forgotten in this discussion. Yemen imports the overwhelming majority of its food and fuel, and a war fought along its ports will produce a humanitarian emergency well beyond the 917 families already displaced this week. The Horn of Africa depends on the same lane. Djibouti, Eritrea and Somalia sit on the western side of a strait that is about to become a military frontier.

For the South Caucasus and for the wider region that this centre studies, there is a further consequence that is easy to miss. Every month that both maritime routes out of the Gulf remain unreliable strengthens the argument for overland corridors between Asia and Europe, which is the argument on which the Middle Corridor through Azerbaijan, Georgia and Türkiye rests. Trade does not move to a new route because the new route is cheaper, it moves because the old route stops being dependable, and that is precisely what is happening now.

The clearest indicator will be Perim, or Mayyun, because holding the island is what turns influence over the strait into control of it. Watch also for the appearance of radar sets, surveillance equipment and missile or drone launchers on the newly captured coast and on the islands, since that is the step that converts territory into the ability to strike shipping.

On the government side, the question is whether the Giants Brigades, the Nation’s Shield, the National Resistance and the Taiz Axis can be made to work as a single force under one command, and whether Saudi air power will be used properly or held back. Newly captured ground is at its weakest before it is entrenched and mined, so there is a window for a counterattack, and that window is measured in days rather than weeks.

Politically, the question is whether the Mecca Defence Alliance turns out to mean anything, and whether Washington chooses to act. And the question that matters most for the rest of the world is a simple one, which is whether the first ship is fired on from the shore rather than from the air, because that would be the moment when the southern Red Sea stops being a dangerous route and becomes a closed one.

Half measures have failed in Yemen for eleven years. The Houthis have understood that better than anyone else in this war, and last week they acted on it.

On the morning of 10 September 2026 the Houthi forces, who call themselves Ansar Allah, entered the port city of Mocha on Yemen’s Red Sea coast and took the port and the airport without a serious fight, because the Saudi backed government forces holding the city had already withdrawn during the night. Three Yemeni government sources confirmed the loss to Reuters, and the government itself said nothing for most of the day, which in this war is usually the clearest confirmation of all. Mocha was not an ordinary town. It had been in the hands of the anti Houthi coalition since early 2017, it was the main military and logistical base of Tariq Saleh’s National Resistance Forces on the western coast, and before that it had served as a base for the United Arab Emirates. Losing it removes the last significant government anchor on the Red Sea, and it sits between sixty and seventy five kilometres north of the Bab al-Mandab strait.

The Houthis Close the Gate

On the evening of 5 September 2026 the American envoys Steve Witkoff and Jared Kushner sat down with Vladimir Putin in the Kremlin, and the meeting ran for more than three hours. Before it began, the Russian president made a point of telling his guests that Kyiv had not been struck since midnight. It was offered as a gesture of good faith, and much of the reporting that followed described it as a ceasefire. It was not a ceasefire. It was a decision to stop hitting one city for three days, taken for the duration of a visit, and everything that lay outside that narrow line continued exactly as before.

The distinction matters more than it might appear, because the shape of the pause tells us a great deal about how Russia understands the war, what it is willing to trade, and what it is not.

The sequence is worth setting out carefully, because the different accounts have been used to support very different conclusions.

On 4 September, ahead of the American visit, President Volodymyr Zelensky proposed a truce in the air. His offer was reciprocal and specific. Ukraine would refrain from striking Moscow from that moment until the end of Monday, and it expected Russia to do the same with regard to Kyiv. Russia did not accept those terms. During the night of 5 September Russian forces struck across Ukraine with ballistic missiles and jet-powered drones, killing at least five people, and they hit the airports at Kyiv and Boryspil. Zelensky read those strikes as a deliberate signal about the possibility of the American delegation flying into Ukraine, which would have been the first arrival of a foreign official by air since the invasion began.

Only on the afternoon of 5 September did the Kremlin spokesman Dmitry Peskov confirm the pause, and his wording was precise. Putin had instructed that no strikes be carried out against Kyiv for three days from midnight, and this was connected to the preparation and conduct of the American meetings in the capital. Putin himself later told Witkoff that Ukraine had published its call for a wider ceasefire but that Russia had never discussed it with the Ukrainian side, and that Russia had not agreed to any pause longer than three days.

The Ukrainian side described the position in almost the same terms from the opposite direction. The presidential communications adviser Dmytro Lytvyn said that the proposal had been passed to Russia but that no actual ceasefire had taken place. The General Staff, speaking through the government’s Centre for Strategic Communications, said that a ceasefire cannot be unilateral or asymmetrical, and that the tasks of the defensive operation were being carried out in full.

There was also a period of genuine confusion on the Ukrainian side, which is worth recording honestly. An influential Telegram channel reported a Ukrainian order for a seventy two hour ceasefire from midnight on 5 September, yet no other outlet confirmed it, individual units reported receiving no such order, and Russian strikes continued through the night. It remains unclear whether an order existed. What is clear is that nothing observable stopped except strikes on Kyiv.

So the true content of the arrangement was this. One capital was granted three quiet days because foreign guests were coming to it. Kamianske in Dnipropetrovsk region was hit with eight ballistic missiles, and large fires broke out at the Kametstal metallurgical plant. The Kyiv region outside the city continued to be attacked, and two people were killed there on the same day. Drone strikes went on elsewhere, including against a parked aircraft at Zhulyany and against the office of the acting head of the Ukrainian security service. A pause that protects the venue of a negotiation while the rest of the country remains under fire is not a step toward peace. It is a demonstration of control over the tap.

The three days should be read against a decision that Putin announced only four days earlier, and which has not been withdrawn.

On 1 September, speaking after the Shanghai Cooperation Organisation summit in Bishkek, Putin said that the Russian armed forces had been instructed to prepare and carry out massive strikes against Ukraine’s energy facilities, and that he had given the order himself. He presented it as a response to Ukrainian strikes on Russian refineries and civilian infrastructure. On the same trip he claimed that Russian troops were twelve kilometres from the Sumy ring road and that Sviatohirsk had been captured. Ukraine’s Centre for Countering Disinformation pointed out that in June he had given the same distance as ten and a half kilometres, that Sviatohirsk had not fallen, and that Sumy has no ring road. A similar pattern appeared around Kozacha Lopan in Kharkiv region, which Russian sources declared captured and which the Ukrainian 14th Army Corps says remains under Ukrainian control amid heavy fighting.

The order to strike the energy system therefore sits in the background of the Kremlin meeting, unrevoked. Independent monitoring of Russian bomber movements suggests that the package was being assembled while the envoys were in the air. By 5 September there were reported to be three Tu-160M and six Tu-95MS aircraft armed with cruise missiles at Ukrainka airbase in the Far East, which is the number normally associated with a large combined attack. The expectation among observers was that the attack had not been cancelled but postponed by a few days.

This is the pattern that has defined the last month. Escalation is decided first, and then paused briefly and reversibly when there is something to be gained from the pause.

The reasoning behind that choice appears to be domestic rather than strategic in any conventional sense.

Russia’s army has been effectively stalemated for years, and the war has settled into a contest of drones and missiles that damages both countries. Yet the Kremlin is not looking for an exit. Reporting from people close to the Russian leadership indicates that Putin spent several weeks over the summer weighing his options and decided to escalate. There are signs that he read the unexpected arrival of the CIA director John Ratcliffe in Moscow on 25 August as evidence of American weakness rather than as a warning. Russian accounts say that Ratcliffe came to argue that Russia’s military and economic position was worsening and that a deal now would serve Moscow better than waiting, that he met the heads of the foreign intelligence service and the security service, and that Putin then declined to see him at all once the purpose of the visit was understood. A proposal for a three way meeting between Trump, Zelensky and Putin was dismissed by Peskov as an information hoax.

The pressures behind the decision are visible enough. The silent public consent that supported the war has been eroding. Deep Ukrainian drone strikes and internet restrictions have made the war part of ordinary Russian life, the Levada Centre finds that fifty seven per cent of Russians describe the situation as tense, and Putin’s approval rating is falling. The mood among the elite is reported to be worse, with the dominant complaint being that nearly five years have been consumed for no result. Putin is said to have told associates at one point that failure to secure the whole of Donbas would end with him being hanged. In a system that tolerates violence but punishes visible failure, simply continuing is not an attractive option, because inertia consumes money and political capital without producing anything to show.

What escalation means in practice is mostly more of what is already happening. Within Ukraine it means heavier strikes on energy and logistics. On the ground, Russian sources say Putin is committed to feeding a further three hundred to four hundred thousand men into the fighting, with recruitment handled by regional governments so that any resulting discontent stays local, and the new system is already being tested in Penza. Outside Ukraine it means more hybrid pressure on the countries that supply Kyiv, and Western agencies are concerned about European factories that make weapons for Ukraine. The same agencies judge that Putin does not want a direct confrontation with NATO. Repression at home is being tightened in parallel, which is best understood as a display of strength rather than a response to any real threat.

The most consequential part of the escalation is the coming attack on Ukraine’s power system, and the balance there has shifted against Kyiv over the past year even though Ukraine has prepared far better than before.

Ukraine’s preparations are real. A programme to harden the grid began in April and is running ahead of schedule, with about half the work finished. The country expects to enter winter with surplus generation capacity, fourteen billion cubic metres of gas in storage, physical protection at key nodes and new private air defence companies supporting the air force. The State Restoration Agency has shielded roughly two dozen of Kyiv’s 110 kilovolt transformers with netting and concrete, having protected none of them before this year, and its director believes the new defences will hold against anything short of a direct missile hit. By December, according to the energy expert Oleksandr Kharchenko, three times more will have been done than in any previous year.

The threat has moved faster. Russian production of ballistic missiles has risen to roughly one hundred a month, against an estimated three a month in 2022, while Ukraine’s stock of interceptors has fallen. Ukraine is seeking an additional three hundred Patriot missiles from partners this winter with no guarantee of receiving them, and domestic interceptor projects are years away. Maxim Timchenko, the chief executive of DTEK, notes that monthly strike records have been set four months running and accepts that a concentrated ballistic attack can destroy any one of the company’s power plants, which are too large to protect fully.

The newer problem is the jet-powered drone. Ukraine had learned to shoot down more than nine in ten of the propeller driven Shahed type, but the jet models fly at up to six hundred kilometres an hour rather than two hundred, and in recent attacks a combination of fighter jets, shoulder launched missiles and automated guns brought down only about thirty per cent of them. They have reached high value targets that ballistic missiles had missed for four years. Russia appears to have solved the production problem by obtaining 3D printing equipment for turbines despite sanctions, and the unit cost has fallen by half or more to somewhere between fifty and seventy thousand dollars. The tactic has changed with the technology. Instead of assembling a large swarm every few nights, Russia now sends one or two jet drones at a time without pause, which keeps air raid alerts sounding almost continuously in Kyiv and exhausts the population without any single spectacular attack.

Kyiv is the weak point. Some substations around the capital have no backups, nor do its connections to the nuclear plants that supply most of its electricity, and the power supply to the water system is vulnerable, which means that a sustained outage would take down sewage and central heating together. Two and a half thousand of the city’s fifteen thousand buildings have no back up source of heat. One senior official expects the capital to be tormented once the frosts arrive.

Alongside the energy campaign, a second and less reported campaign has been aimed at the movement and storage of goods.

The data assembled by the analyst Clément Molin shows one hundred and forty recorded strikes on Ukrainian warehouses in August, of which more than ninety five were against civilian logistics facilities. The monthly average had been about thirty for sixteen months, then roughly sixty after Ukraine began mid range strikes on Russian logistics, and then one hundred and forty in August. The geography moved too, with Kharkiv region the most targeted between February and June and Kyiv city and region the most targeted in August.

Molin argues that the surge is connected to Ukraine’s own campaign against Russian logistics, including strikes on Wildberries facilities, and that Russia used it as justification for a much larger effort against Ukrainian civilian storage. This is a contested reading. Others point out that Russia has struck warehouses throughout the war and that many of the targets, particularly in Kyiv, have both military and civilian uses. Both things can be true at once. The pattern through this war has been that each side adopts the other’s methods, from fuel truck strikes answered by attacks on filling stations, to Ukrainian drone use answered by mass Russian drone deployment, to Russian fortification lines copied two years later by Ukraine, to Russian strikes on the grid answered by Ukrainian strikes on power supply in occupied areas. What is not in dispute is the effect. Ukraine’s agrarian policy minister Taras Vysotskyi says that around ninety per cent of the country’s modern storage and warehouse capacity has been destroyed, and while he sees no risk of famine, he warns that continued attacks will affect the availability and variety of goods in shops.

There are also signs that the campaign is moving west, with Russian reconnaissance flights reported in the direction of Rivne and Ternopil in early September. Regions that have been comparatively safe for four years should not assume they will remain so.

While attention is on the air war, the region the whole war is nominally about is quietly ceasing to be a place where people live.

Of the Donetsk region still under Ukrainian control, about two thousand five hundred and twenty square kilometres out of roughly four thousand one hundred and thirty are now zones of mandatory evacuation for families with children, which is around sixty one per cent. Combining occupation with mandatory evacuation, about ninety four per cent of the region is either in Russian hands or effectively empty. Measured across Donbas as a whole, roughly ninety seven per cent is occupied or dead, leaving something close to three per cent where ordinary civilian life continues. That remaining sliver contains most of Slovyansk and Kramatorsk, the towns of Novodonetske, Cherkaske and Oleksandrivka, and a number of villages. About thirty per cent of Slovyansk and ten per cent of Kramatorsk already fall inside evacuation zones.

The population figures show how quickly this is happening. There are 92,932 civilians left in Ukrainian controlled Donetsk region and none at all in the sliver of Luhansk region still held. In January 2026 the figure was 192,600, so the civilian population has more than halved in eight months and has fallen by nearly forty thousand in the last two. Novodonetske, one of the last two significant liveable towns in the western part of the region, is now under intensified glide bomb and drone attack with the front line still around seventeen kilometres away.

On present trends the remaining areas will be placed under mandatory evacuation in the coming months, and Donetsk region will join Luhansk and Crimea as territory that is either occupied or empty. After nearly thirteen years of war that began in 2014, the end of 2026 may mark the end of Donbas as a functioning part of Ukrainian society. This deserves more attention than it receives, because a settlement that hands Russia the whole of Donbas would be transferring a region that has already lost the population the argument was supposedly about.

None of this means Russia is comfortable, and the financial picture explains part of the urgency.

Oil and gas provide a fifth of Russian budget revenue, and in the first half of 2026 that revenue came to 3.7 trillion roubles, twenty three per cent lower than a year earlier, even though the average oil price used for tax purposes was higher than the budget assumed because of disruption around the Strait of Hormuz. The deficit for the first half alone reached 5.7 trillion roubles, about two and a half per cent of annual output, against a projected 3.8 trillion for the whole year.

A strong currency made this worse, because exporters earn dollars and the budget is written in roubles. The rouble averaged seventy seven to the dollar in the first half against a budget assumption of ninety two, so the government moved on 5 August to increase daily purchases of foreign currency and gold by about a fifth, and the rouble duly weakened to around eighty five by mid month. The finance ministry has also been given the right to borrow above the statutory limit without further approval, and Russian businessmen have been asked to make voluntary contributions to the budget, which have produced 384 billion roubles since the start of the year. Inflation has risen to six per cent, helped by the petrol shortage caused by Ukrainian strikes on refineries, and a weaker rouble will push it higher.

This is a government managing a squeeze rather than one facing collapse. Ukrainian long range strikes are imposing real costs, but the Russian ability to keep fighting is not about to run out, and Western assessments that assume otherwise have been wrong before.

The most significant development of the last month may not be on the battlefield at all, but in how European governments have started to plan.

Sweden is the clearest example, because it has moved further than most. Two years after joining NATO and abandoning two centuries of neutrality, support for Ukraine is so uncontested there that the war barely featured in the campaign for the election of 13 September. Prime Minister Ulf Kristersson says he sees no sign that Russia has changed the basic logic behind the war, that Russia wants to conquer Ukraine, and that nobody knows what its aims would be after that. His government has tied Sweden to Ukraine as a long term industrial partner, including Gripen fighter aircraft, and has expanded conscription while deepening ties with its Nordic and Baltic neighbours and with Britain, France, Germany, the Netherlands and Poland. On the centre left, which polls suggest will win the election, senior figures describe Russia as a long term existential threat and talk about a fifty year commitment to national defence.

The wider picture supports that reading. Finland has asked for Swedish aircraft and ships to help monitor its border for incursions. Germany has publicly blamed Russia for a drone attack at Leipzig airport on 4 August. Romania has destroyed a Russian naval drone that targeted a drilling platform in its waters, and Slovakia has foiled an arson attack on a drone factory. These are small events individually, but together they describe a campaign that is already under way against countries that are not at war.

What worries Stockholm is not only the war but its ending. There is a real fear that a ceasefire would produce a frozen conflict rather than a durable peace, and that European unity, which has held up better than expected, would come under strain as countries further from Russia press for normal relations. There is quieter anxiety about what would happen if Washington sought peace at any price, by restricting the transfer of American weapons and parts, cutting off intelligence, or threatening to withdraw troops from Poland and the Baltic states.

Three days of quiet over one city, granted for the convenience of a meeting and framed by its author as a favour, is the most accurate available summary of where this war stands.

The war is intensifying, because Russia has decided to escalate and has the missiles, the drones and the manpower plan to do it. It is expanding, because the pressure is now being applied to Ukraine’s suppliers and neighbours as well as to Ukraine, and because the strike campaign is moving toward regions that were previously spared. And it is stuck, because neither side can impose a result and both consider defeat worse than continuing.

The diplomacy is real, and the arrival of American envoys in Moscow and Kyiv in the same weekend is not nothing. But the terms on which the pause was granted show what Russia is prepared to concede at this stage, which is a temporary and geographically limited quiet that costs it nothing and can be withdrawn at any moment. The bombers were being loaded while the talks were going on. Until a pause covers the whole country, applies to both sides, and survives the departure of the visitors, it is a piece of stagecraft rather than a step toward the end of the war.

On the evening of 5 September 2026 the American envoys Steve Witkoff and Jared Kushner sat down with Vladimir Putin in the Kremlin, and the meeting ran for more than three hours. Before it began, the Russian president made a point of telling his guests that Kyiv had not been struck since midnight. It was offered as a gesture of good faith, and much of the reporting that followed described it as a ceasefire. It was not a ceasefire. It was a decision to stop hitting one city for three days, taken for the duration of a visit, and everything that lay outside that narrow line continued exactly as before. The distinction matters more than it might appear, because the shape of the pause tells us a great deal about how Russia understands the war, what it is willing to trade, and what it is not.

Putin's Three-Day Favour

The war has changed shape. Neither side can move the front line in any way that decides the outcome, so both have shifted their main effort to destroying the other’s economy from the air and at sea. Russia is closing Ukraine’s sea exports by attacking merchant ships rather than ports. Ukraine is burning Russian refineries and airbases at ranges that now reach the Gulf of Finland.

Two developments in the last forty-eight hours matter more than anything else in this report. Ukraine struck the KINEF refinery at Kirishi in Leningrad Oblast, which is one of the largest refineries in Russia, and it did so as part of a mass attack reaching towards St Petersburg. At the same time Ukraine disclosed intelligence that the Kremlin has ordered its military to prepare ground offensives against Kyiv and Chernihiv.

The two facts point in opposite directions, and holding both is the analytical problem of the moment. Ukraine is inflicting deeper damage on Russia than at any previous point in the war, while simultaneously warning that Russia is preparing to widen it.

A note on sources appears at the end, because the material for this period comes in three very different tiers and they should not be read alike.

Overnight on 29 and 30 August Ukrainian long-range drones attacked the KINEF refinery at Kirishi in Leningrad Oblast and started a large fire. The plant processes around 20 million tonnes of crude a year, which is roughly 6 per cent of Russia’s total refining capacity, making it the most significant single refining target struck so far. Hundreds of drones were reported moving towards Leningrad Oblast and St Petersburg in the same wave, with Russian air defence engaging both there and in the regions the drones crossed.

Two airbases were hit in the same period. Satellite fire-detection data indicated a large fire at Millerovo in Rostov Oblast, which hosts Su-30, MiG-29 and Su-25 aircraft and serves as a launch point for Geran-2 and Gerbera drones. At Yeisk in Krasnodar Krai, home to Su-30SM and Su-34 aircraft, the strike appears to have hit a storage facility holding KAB glide bombs, with secondary detonations continuing for over an hour and multiple fires still growing.

Striking glide-bomb stocks is significant beyond the immediate damage, because KABs are the weapon doing most of the work against Ukrainian positions in Donetsk oblast, and they are cheap, plentiful and difficult to counter in flight.

This campaign has produced cumulative results. Strikes on refineries have taken out upwards of 30 per cent of Russia’s operating refining capacity. Petrol rationing at its peak affected 80 regions and continues in 18. Reuters has reported that Russian petrol output fell to about 70 per cent of domestic demand in late August after attacks forced the Perm, Nizhny Novgorod and Yaroslavl refineries offline. President Zelensky has said that Russia is importing petrol for the first time in its history while spending budget funds to support its oil companies.

These figures need care. Ukraine’s general staff claims 13.5 billion dollars of losses to Russian refining since August 2025, while an analyst at the Carnegie Endowment assesses the real immediate business losses at perhaps a tenth of that and says state revenue has barely been touched. Both sides inflate.

Figures for the twelve hours to 30 August illustrate a change that has happened quietly. The Ukrainian Air Force reported 130 Russian drones of the Geran and Gerbera families. The Russian Ministry of Defence claimed Ukraine had launched at least 494 mid to long-range drones.

Each figure comes from the opposing side and both should be treated as claims, but the direction is consistent with other reporting. Ukraine is now launching more drones than Russia on some nights. What Russia retains is the ballistic and hypersonic missile force, which is the category Ukraine can no longer intercept.

The Ukrainian Air Force has also stopped publishing its interception rates, which is consistent with the collapse described below.

The most consequential military fact of the summer is that Ukraine has almost exhausted its Patriot interceptors, the only weapons it holds that can engage Russian ballistic and hypersonic missiles.

The intercept rate against ballistic missiles was around 70 per cent in March, fell to 20 per cent in July, and on the night of 5 August there were no interceptions at all. Russia has responded by increasing missile production. In July it fired a record 198 ballistic and hypersonic missiles, most at the Kyiv region, and on 27 August it conducted an hours-long attack on Kyiv and other cities using ballistic missiles together with 258 drones, aimed at ports, industrial sites and the distribution centres of retail chains.

Ukraine has asked for at least 300 Patriot interceptors. Its allies do not have them either. The Pentagon fired close to two thirds of its pre-war Patriot stock during this year’s war with Iran, and rebuilding to pre-war levels is estimated to take until 2029. A PAC-3 MSE interceptor costs four to five million dollars while the drones it is used against can cost as little as a used car, and defenders normally fire two interceptors at each incoming missile.

President Zelensky acknowledged in a video address on Friday 28 August that Ukraine needs to significantly accelerate its own production of interceptors capable of downing Shahed-type drones. That is an admission that the Patriot supply problem will not be solved externally in the timeframe that matters.

The turning point came in July, when Russia stopped attacking port infrastructure and began attacking ships. On 19 July cruise missiles sank a Turkish-owned cargo vessel leaving Chornomorsk and killed ten people. Since then every vessel entering or leaving Odesa has been targeted, and no vessels entered Odesa at all during the first weeks of August.

Russia has extended the campaign to the Danube. On 17 August it struck the Izmail district and damaged a Togo-flagged civilian ship, injuring four people. The Danube was Ukraine’s fallback route in 2022, so closing it removes the alternative as well as the main channel.

Ukraine has done the same in reverse. Its drones have effectively closed the Sea of Azov to Russian shipping, and in mid-August all three grain terminals at Novorossiysk, the largest wheat-exporting port in the world, suspended operations. More than 90 per cent of Russia’s grain export capacity in the basin is offline in the busiest weeks of the season.

Taken together, roughly 97 per cent of the combined Russian and Ukrainian grain export capacity in the Azov and Black Sea basin is now out of action.

The scale is global. In August 2025 the two countries exported 6.3 million tonnes of wheat between them, out of about 16 million tonnes traded worldwide that month. This August their combined exports may reach only about 2.5 million tonnes. Ukrainian grain exports fell around 75 per cent in the first two weeks of August year on year, and Kyiv has said port closures may halve its farm exports for the year, with 50 million tonnes of harvest still to move. Estimates in the trade press put the potential global supply gap at 30 to 40 million tonnes, though these are projections rather than observed outcomes.

The damage runs both ways. Fourth-class Russian wheat has fallen to about 12,000 roubles a tonne from 15,000 a year earlier, precisely because Russia cannot export it while world prices rise.

The Russian summer offensive is running on three axes, which are Sloviansk, Kostyantynivka and Dobropillya.

 is the most advanced. Russian forces hold significant positions throughout all parts of the town. Ukrainian forces retain significant positions in only about a quarter of it, concentrated around Lomonosova Street and the adjacent residential blocks in the west, where the two sides overlap. The accurate description is that neither side controls the town, since most of it is a large contested zone, but the trend is against Ukraine. Fresh geolocations at the end of August led one detailed mapping analyst to adjust the western grey zone slightly in Ukraine’s favour, so the position there is not yet settled.

 is being approached from the east and south-east after Russian attempts to take it from the north-east through Lyman failed earlier in the offensive. No confirmed advances were recorded in the last week of August.

 is under increased artillery and glide-bomb attack, with large fires burning around Dobropillya and Bilozerske in the last twenty-four hours. Druzhkivka, in the same belt, is described by observers as reduced to a ghost town.

Elsewhere, Russian forces continued offensive operations in the Pokrovsk, Novopavlivka and Oleksandrivka directions in the last week of August without advancing. Ukrainian forces retook the small town of Novoselivka in the Lyman direction after a battle lasting about six months.

The wider picture is of Russia creeping forward at heavy cost while Ukraine gains ground locally, with neither achieving anything decisive. That is why both have moved their main effort into the air and the sea.

On 29 August Pavlo Palisa, deputy head of the President’s Office, told reporters that Russian military chiefs have been tasked by the Kremlin’s political leadership with planning and preparing attacks on Kyiv and Chernihiv, based on intelligence Ukraine says it has received and verified.

The caveat he attached matters as much as the claim. Palisa said there is not yet evidence that Russian troops are gathering to carry out such attacks, and that Ukraine’s priority is to prevent the conditions in which any offensive becomes possible. This is therefore a report about planning, not about observed force concentration.

It should also be read against Russia’s manpower position. Western assessments reported in late August indicate that Russia is losing around 6,000 more troops each month than it recruits. The Wall Street Journal has reported that Russia is testing its capacity to mobilise further troops, and there is a growing exodus of Russians who expect a new mobilisation to be announced. Ukrainian and foreign intelligence have reported a Russian General Staff plan for an additional 600,000 mobilised across 2026 and 2027.

An offensive against Kyiv would require forces Russia does not currently have available, which is consistent with Palisa’s own caveat. The planning is credible; the capability is not yet demonstrated.

Russia has been striking civilian logistics deliberately. A warehouse supplying a Kyiv food retailer was hit on 19 July, two distribution hubs of a larger chain were struck two weeks later with six workers killed, and the private postal service much of the country depends on has been hit repeatedly. One businessman estimated three weeks of strikes cost Ukrainian companies around 500 million dollars.

The pattern continues. Recent Geran-4 jet-drone strikes in Mykolaiv oblast targeted the Trykhaty 330 kilovolt electrical substation, a marine transhipment terminal and a logistics warehouse, coordinated with Geran-2 drones, glide bombs, Banderol cruise missiles and Iskander-M ballistic missiles. Kyiv was struck again on 30 August, including a target in the Holosiivskyi district that caused a large fire.

Two capability points follow. The Geran-4 jet-drone has demonstrated a flight of approximately 980 kilometres to Rivne oblast, which is a new record and places western Ukraine within reach. Analysts expect attacks on Lviv, Chernivtsi, Rivne and Ivano-Frankivsk oblasts in the coming days, with logistics warehouses, large retail centres and electrical substations as likely targets. Western Ukraine has until now been the country’s rear area.

The economic exposure is concentrated in agriculture. A Ukrainian investment firm estimates the sea blockade will cost about 0.6 per cent of GDP this year. If the harvest cannot be sold, farmers are likely to skip winter sowing, and the head of the Ukrainian Agribusiness Club puts the resulting losses at 10 to 12 billion dollars, or around 5 per cent of GDP. President Zelensky disclosed on 22 August a 27 billion dollar shortfall in the defence budget. Ukraine says it holds sufficient gas reserves for the winter.

The military position is complicated by a political one. On 18 August Mykhailo Fedorov, dismissed as defence minister in July after a clash with senior generals, released a video challenging Zelensky to hold elections. Wartime elections are prohibited under martial law and Zelensky has described the idea as a tsunami for the country, with polling showing Ukrainians strongly opposed.

The following morning parliament confirmed his successor, General Yevhenii Khmara, by 312 votes to two. On the same day the anti-corruption agencies released material from two linked investigations into a scheme to launder bail money for a former justice minister facing charges. Iryna Mudra, a deputy head of the President’s Office, was charged and dismissed. An unpublished poll taken before these events put Zelensky at 21 per cent, former commander-in-chief Valery Zaluzhny at 18, Fedorov at 16 and Kyrylo Budanov at 8.

A debate has opened about the anti-corruption bodies themselves, with supporters describing them as a necessary check on a centralised presidency and critics, including some inside government, calling them a parallel government.

, now head of the President’s Office, said on 28 August that Russia categorically does not want to stop the war along the line of contact, that talks are limited mainly to prisoner exchanges, and that another large exchange will follow shortly. Asked whether Ukraine would still be fighting in winter, he said he was afraid so, adding that the war cannot be stopped now not because Ukraine does not want to but because it is unrealistic. He also said Ukraine must mobilise at least 30,000 people a month simply to hold what it has, that the new minister will not change that figure, and that the number of dissatisfied people will therefore not fall either.

 said on 29 August that Russia is bombarding Kyiv with virtually nonstop attacks from jet-powered drones.

, the Polish prime minister, said on 29 August after a briefing from the NATO Secretary General that escalation from Russia’s side is progressing, that Ukraine needs more air support, and that the next six months will be decisive and Poland and NATO must be prepared for various scenarios.

 told Axios that he is not concerned at all about Russia attacking NATO members, and denied that the recent Ratcliffe visit to Moscow carried any warning, describing the meeting as routine.

 said on 27 August that the United Kingdom and France are playing with fire by supplying Ukraine with technology capable of striking deep inside Russia, and threatened British military facilities both in Ukraine and abroad. 

, a deputy chairman of the State Duma, argued on state television that Russia should bomb Ukrainian cities openly, after twenty-four hours notice, until the population is demoralised enough to surrender.

The diplomatic track, and a contradiction worth noting

John Ratcliffe, the CIA director, travelled to Moscow around 25 August and met the heads of the FSB and the foreign intelligence service. Three accounts of that meeting are now in circulation and they do not agree.

The New York Times, citing American officials, reported that he gave a pessimistic assessment of Russia’s position and urged genuine negotiations, but issued no threats about the consequences of continuing the war. Politico reported that he warned Russian officials not to escalate the conflict with America’s NATO allies. President Trump denied that any message was carried at all and called the visit routine.

These cannot all be correct, and the divergence is itself informative about how the administration is managing the file. Subsequent reporting indicates Ratcliffe raised the idea of a trilateral summit between Trump, Putin and Zelensky, and judged the FSB director a potentially constructive participant in restarting talks. Zelensky has been briefed and is reported willing. Putin has rejected the format before.

This channel runs in parallel with those maintained by the special envoys Steve Witkoff and Jared Kushner, who may travel to Kyiv and Moscow. Ukrainian officials remain sceptical and point to the intelligence on northern offensive planning as evidence that Moscow is not preparing to negotiate.

Separately, the European Union is expected to resume preparations from 1 September for opening negotiating clusters with Ukraine and Moldova.

Four judgements follow from the material above.

The sea is the decisive theatre. Russia can close Ukraine’s exports and Ukraine cannot fully reciprocate, because Russia retains Baltic and Caspian alternatives while Ukraine has only the Danube, which is constrained by low water and is now itself under attack. Russia is therefore likely to continue regardless of what happens in diplomacy.

The interceptor shortage converts every other Ukrainian problem into a worse one. Without air defence, strikes on logistics, energy and ports cannot be prevented, only absorbed. The extension of Geran-4 range to nearly a thousand kilometres removes the western rear area as a place of relative safety, which affects both civilian morale and the storage of incoming Western supplies.

Ukraine’s deep-strike campaign is working, but on a slower clock than Russia’s. Burning 6 per cent of Russian refining capacity in a night is a real achievement and the airbase strikes reduce the glide-bomb and drone effort at source. The question is whether that accumulates faster than the damage to Ukraine’s exports and grid, and on the present evidence it does not.

The warning about Kyiv and Chernihiv should be treated seriously but read precisely. Ukraine reports Kremlin-directed planning, not force concentration, and Russia is currently losing more men each month than it recruits. The most likely reading is that Moscow is preparing options rather than executing a decision, and that disclosing the intelligence serves Kyiv’s argument for more air defence, which Tusk made explicitly the same weekend.

The scenario discussed quietly in Kyiv is that strikes on energy, water and sewage could force mass emigration. One businessman described a company meeting originally titled Autumn Actions being renamed Survive Till April.

The war has changed shape. Neither side can move the front line in any way that decides the outcome, so both have shifted their main effort to destroying the other’s economy from the air and at sea. Russia is closing Ukraine’s sea exports by attacking merchant ships rather than ports. Ukraine is burning Russian refineries and airbases at ranges that now reach the Gulf of Finland. Two developments in the last forty-eight hours matter more than anything else in this report. Ukraine struck the KINEF refinery at Kirishi in Leningrad Oblast, which is one of the largest refineries in Russia, and it did so as part of a mass attack reaching towards St Petersburg. At the same time Ukraine disclosed intelligence that the Kremlin has ordered its military to prepare ground offensives against Kyiv and Chernihiv. The two facts point in opposite directions, and holding both is the analytical problem of the moment. Ukraine is inflicting deeper damage on Russia than at any previous point in the war, while simultaneously warning that Russia is preparing to widen it. A note on sources appears at the end, because the material for this period comes in three very different tiers and they should not be read alike.

Russia-Ukraine: The War Turns to Strangulation

Two and a half years into Russia's full-scale invasion, and roughly ten weeks into its 2026 spring-summer offensive, the front line barely moved in the first half of August. Ukraine's DeepState monitoring group recorded Russian net gains of about 36–88 km² in July (estimates vary because of reporting lag), a rate of advance of roughly 1.2 km² per day — glacial by any measure, and a pace the Institute for the Study of War (ISW) says has produced no operationally significant breakthrough despite Russia's stated ambition of enveloping Ukraine's "Fortress Belt" cities in Donetsk Oblast from the southwest via the Pokrovsk and Dobropillya axes.

Two and a half years into Russia's full-scale invasion, and roughly ten weeks into its 2026 spring-summer offensive, the front line barely moved in the first half of August. Ukraine's DeepState monitoring group recorded Russian net gains of about 36–88 km² in July (estimates vary because of reporting lag), a rate of advance of roughly 1.2 km² per day — glacial by any measure, and a pace the Institute for the Study of War (ISW) says has produced no operationally significant breakthrough despite Russia's stated ambition of enveloping Ukraine's "Fortress Belt" cities in Donetsk Oblast from the southwest via the Pokrovsk and Dobropillya axes.  That slow crawl came at a steep price. Ukraine's General Staff put Russian losses in July at roughly 42,900 killed and wounded, the highest monthly figure of 2026 and among the costliest months of the war. These are Kyiv's own figures and cannot be independently verified — a caveat worth stressing, since neither side publishes reliable casualty data and both have incentives to shade the numbers. Independent OSINT trackers, including the pro-Ukrainian account AMK Mapping, cited similar figures (around 43,000) while noting it was the second-worst month for Russian losses since January 2025. Russian-aligned commentators, such as the podcast Russians With Attitude, have pushed back on the reflexive trust some place in Ukrainian General Staff numbers — a fair point of methodological skepticism that applies to battlefield claims from both sides.

Research

The Saudi Kingdom Has No Way Out

A Week of Deepening US–Iran Strikes as Oil Tops $100

Commentary