Hormuz remains selectively open as Ansar Allah widens Saudi pressure and Russia-Ukraine strikes reciprocate

Bottom Line

US Central Command and the Iranian Revolutionary Guard Corps Navy are very likely to remain in an active maritime confrontation through 5 October 2026, while commercial traffic through the Strait of Hormuz stays severely reduced but not demonstrably halted. Ansar Allah, also known as the Houthis, is likely to retain the ability to threaten Saudi targets and shipping from positions at or near Mokha port and Mayyun Island, also known as Perim, but the reviewed reporting does not establish continuous Ansar Allah control of the Bab al-Mandab traffic separation scheme.

Russia and Ukraine are very likely to sustain reciprocal deep strikes through 5 October. No independently corroborated evidence reviewed as of 21 September establishes a physical attack on NATO territory, although a reported explosive-carrying drone entered Lithuanian airspace. Humanitarian conditions in Yemen, Sudan and Gaza are very likely to worsen if fighting continues. Overall confidence is moderate because the strongest judgements use maritime monitors, commercial tracking, UN agencies and major media, while attribution and territorial-control judgements rely partly on belligerent statements, local media and single-source reporting.

Key Developments

Hormuz remained severely disrupted, not demonstrably closed

Reported fact: The Iranian Revolutionary Guard Corps Navy claimed that it struck the Togo-flagged tanker Trend on 17 September. UK Maritime Trade Operations reported further projectile incidents. The International Maritime Organization reported 80 attacks on international shipping since 28 February and at least 22 seafarer deaths. Kpler recorded four AIS-visible commodity-vessel transits on 17 September. Separate reporting identified at least two LNG movements during the week, while US Central Command claimed that more than 1 billion barrels had moved under protection over a longer period.

Assessment: Commercial traffic through Hormuz is very likely to remain severely constrained through 5 October, but a complete and continuous closure is unlikely. Confidence is moderate for the direction of disruption, high for the occurrence of reported maritime incidents, and low to moderate for Iranian attribution of incidents not independently assigned by UK Maritime Trade Operations.

Source quality and independence: The IMO and UK Maritime Trade Operations provide stronger institutional and operational reporting on attacks and navigational warnings. Kpler provides commercial tracking, but AIS visibility is incomplete. The Trend account is an Iranian military claim and is not equivalent to independent attribution. US Central Command's escorted-vessel and oil-volume figures measure a different population and period from Kpler's AIS-visible commodity-vessel count.

Ansar Allah widened the threat to Saudi Arabia

Reported fact: Ansar Allah claimed missile and drone attacks involving Riyadh and Yanbu. Saudi authorities reported interceptions or thwarted attacks involving Riyadh, Bisha, Taif, Farasan and Yanbu. Reporting places Ansar Allah at or near Mokha port and Mayyun Island. Kpler and other shipping reports record lower Bab al-Mandab traffic, but the available counts use different baselines.

Assessment: Ansar Allah is likely to continue attacks or attempted attacks against Saudi targets through 5 October and is likely to retain the capability to threaten vessels near the Bab al-Mandab approaches. Ansar Allah likely holds or contests Mokha port and Mayyun Island, but has not demonstrated continuous control of the international traffic separation scheme, the ability to inspect all vessels or uncontested control of the wider strait. Confidence is moderate.

Source quality and independence: Saudi and Ansar Allah statements confirm that both parties are presenting the confrontation as active, but their success claims are opposed. Major-media reporting and commercial vessel data provide some independent evidence of disruption. Reports repeating an Ansar Allah territorial claim are not independent corroboration. Independent imagery, vessel tracks, port records and geolocated video would be required to establish continuous control of Mokha or Mayyun.

Russia and Ukraine entered a more reciprocal strike cycle

Reported fact: Russian forces struck Kyiv, Odesa, rail and logistics infrastructure between 18 and 20 September. Ukrainian forces conducted a large drone operation against Moscow and surrounding regions, with damage near the Kapotnya refinery reported but not independently reconciled. Ukraine's 3rd Army Corps reported gains around Lyman, while Russian forces retained offensive activity in Donetsk and other sectors.

Assessment: Russia and Ukraine are very likely to sustain high-tempo long-range strikes through 5 October. Confidence is high for continued strike activity, moderate for the direction of territorial change and low to moderate for specific target effects because many claims originate with the belligerents.

Source quality and independence: Ukrainian and Russian official reporting is useful for activity and force claims but is not independent on damage or casualty effects. Major media and open-source geolocation support the existence of a reciprocal strike cycle. The Lyman territorial gain is unresolved because reported estimates range from approximately 85 square kilometres to more than 200 square kilometres and are not supported by a single agreed geospatial baseline.

Russian-linked hybrid activity remained below open NATO-Russia conflict

Reported fact: NATO aircraft reportedly downed an explosive-carrying drone that entered Lithuanian airspace from Belarus. Poland conducted preventive air operations during Russian strikes on Ukraine without reporting a violation. Germany, Denmark, France and Poland continued to report sabotage, cyber activity, espionage or recruitment linked to Russia.

Assessment: Russia is very likely to sustain a hybrid campaign against European states supporting Ukraine. A further hazardous airspace or sabotage incident is likely over the next three months, while open Russia-NATO kinetic conflict remains very unlikely through 21 December. The Lithuanian incident demonstrates an airspace violation, not a confirmed physical attack on NATO territory. Confidence is high for the campaign-level activity and moderate for attribution of individual incidents.

JNIM and ISWAP sustained pressure against state forces

Reported fact: JNIM's 10 September assault at Dioura was followed by reports of more than 100 Malian soldiers killed, five Russia Africa Corps personnel killed, prisoners taken and the camp abandoned. ISWAP attacked FOB Mairari and FOB Gajiram in Borno State on 17 September. Nigerian forces reported repelling both attacks with intelligence, surveillance and reconnaissance and close air support, while the outcome at Gajiram remains disputed.

Assessment: JNIM is likely to sustain attacks on FAMa and Africa Corps positions, and ISWAP is likely to continue attacks on Nigerian bases through 21 December. The reporting does not establish a region-wide surge. Confidence is moderate for continued activity and low to moderate for attack effects and casualty totals.

Humanitarian deterioration accelerated

Reported fact: UN and IOM reporting places Yemen's recent internal displacement between more than 112,000 and more than 125,000 people, depending on the reporting window. UN-linked reporting records between more than 2,000 and roughly 3,000 arrivals in Djibouti. WFP and IOM warn of food and relief reductions in Sudan, where about 20 million people face acute hunger. OCHA reports that nearly 84 per cent of Gaza's population lives in overcrowded displacement sites or damaged shelters.

Assessment: Humanitarian conditions are very likely to worsen through 21 December if fighting continues. Confidence is high for the direction of deterioration because UN agencies independently report displacement, food and shelter pressures. Confidence is lower for aggregate totals because agency definitions and reporting periods differ. The Yemen figures should not be added together.

China-Philippines coercion intensified

Reported fact: China Coast Guard vessel CCG-21585 and Philippine BFAR vessel BRP Datu Magat Salamat collided near Sabina Shoal on 18 September. The Philippine vessel suffered reported starboard-side damage. Beijing and Manila assign responsibility to each other.

Assessment: China is very likely to sustain China Coast Guard pressure around Sabina Shoal, Scarborough Shoal and Second Thomas Shoal over the next three months. Another dangerous encounter is likely, while a sustained armed clash remains very unlikely. Confidence is moderate.

Source quality and independence: The collision and damage are supported by opposing official accounts, Philippine Coast Guard material and major-media reporting. Responsibility remains unresolved because the two governments describe the decisive manoeuvre differently. The reviewed reporting does not independently establish firearms or other military weapons use.

Analysis

Change since the 14 September briefing

Prior judgement	Current judgement	Direction of change	Evidence driving the change	Confidence change
US-Iran maritime confrontation would continue through 28 September and Hormuz would remain severely constrained but not fully closed.	US Central Command and the IRGC Navy are very likely to remain in an active confrontation through 5 October. Hormuz traffic is very likely to remain severely constrained and a complete continuous closure is unlikely.	Continuity on the core forecast, with stronger evidence of selective disruption.	The 17 September Trend claim, further UK Maritime Trade Operations warnings, the IMO tally, Kpler's four-vessel count and continued LNG movement.	Remains moderate. Confidence in severe disruption is higher than confidence in attribution or exact traffic volume.
Ansar Allah would retain a presence and partial tactical control at Mokha, Dhubab and Mayyun, without established uncontested Bab al-Mandab control.	Ansar Allah likely holds or contests Mokha port and Mayyun Island and is likely to continue threatening Saudi targets and nearby shipping, but full control of Bab al-Mandab remains unconfirmed.	Threat to Saudi territory increased; territorial-control judgement remains broadly unchanged.	Claims of attacks on Riyadh and Yanbu, Saudi interception reports and continued reporting on Mokha and Mayyun.	Remains moderate because the territorial evidence is partly party-derived.
Russia and Ukraine would sustain high-tempo strikes through 28 September, with an intentional Russian attack on NATO territory very unlikely.	Russia and Ukraine are very likely to sustain reciprocal deep strikes through 5 October. A further hazardous NATO airspace incident is likely over three months, but a physical attack on NATO territory remains very unlikely.	Strike cycle became more reciprocal; no categorical probability change is claimed for accidental spillover because the prior briefing did not provide a comparable probability.	Ukrainian strikes against Moscow and Russian rear areas, Russian strikes against Ukrainian logistics, and the reported Lithuanian airspace incident.	High for continued activity, moderate for effects and spillover.
China Coast Guard patrols east of Taiwan and coercive encounters with the Philippines would continue; a Taiwan blockade was very unlikely by 28 September.	China Coast Guard pressure on Philippine vessels is very likely to continue and another dangerous encounter is likely over three months. No new large-scale PLA exercise is independently established in the current window.	Immediate South China Sea encounter risk increased after the Sabina Shoal collision; Taiwan assessment remains continuity.	CCG-21585 collision with BRP Datu Magat Salamat on 18 September and continued reporting of Taiwan pressure.	Moderate for the South China Sea pattern, low to moderate for specific Taiwan activity.
SAF-RSF combat would continue and Sudan's humanitarian situation would worsen; direct Sudan-Ethiopia war was unlikely.	SAF-RSF combat is likely to continue through 21 October, humanitarian conditions are very likely to worsen, and direct Sudan-Ethiopia war remains unlikely.	Continuity, with stronger warning of an end-September aid shock.	WFP ration concerns, IOM stock warnings, fighting around El Obeid and continued allegations of Ethiopian support to RSF drones.	High for humanitarian direction, moderate for battlefield and cross-border assessments.
Niger's junta would retain Niamey while military cohesion remained vulnerable.	No new independently corroborated mutiny or change of control is established in the current window. The prior judgement is retained.	No assessed change.	Current reporting focuses on Mali, Burkina Faso and Nigeria; absence of new evidence is not evidence that Nigerien cohesion has improved.	Remains low to moderate because the original mutiny reporting was thin.
Israeli military activity in Gaza and southern Lebanon would continue, with humanitarian deterioration.	Israeli operations in Gaza are very likely to continue rendering the October 2025 ceasefire operationally ineffective through 5 October; Israeli-Hezbollah exchanges are likely to continue.	Continuity, with further evidence of persistent Gaza violence and southern Lebanon incidents.	OCHA, UNICEF, UNIFIL, IDF and Lebanese reporting through 20 September.	High for Gaza activity and humanitarian deterioration, moderate for individual Lebanon incidents.
The prior briefing did not provide a comparable probability for accidental Russia-NATO spillover.	A further hazardous airspace incident is unlikely but consequential; no numerical probability change is claimed.	Assessment refined, not re-baselined.	Lithuanian airspace reporting and Polish preventive air operations.	Moderate for incident risk, very low for open conflict.

The primary forecast horizon is 5 October because it covers the 14 days following the end of this reporting window. Secondary judgements use 21 October for Sudan and 21 December for three-month military and political trends. This replaces the prior briefing's 28 September and 14 October dates rather than implying a direct probability change.

Stakeholder implications

Commercial carriers and war-risk insurers face a route-eligibility problem rather than a simple closure decision. Kpler AIS-visible counts, US Central Command escorted movements and LNG transfers are not interchangeable. A second tanker strike, compulsory Iranian inspection or a three-day absence of loaded tanker and LNG movements would increase route and insurance risk. Sustained same-method traffic recovery would reduce it.

Saudi Aramco, Yanbu operators and Gulf refiners face different risks from Hormuz and Bab al-Mandab. A verified outage at the East-West Pipeline, an impact at Yanbu or another successful Ansar Allah strike would affect Saudi export routing. Asian refiners should distinguish delayed cargoes from permanent supply loss because ship-to-ship transfers and Oman-route movements continue in limited form.

US Central Command and Gulf governments face a higher attribution burden after the Trend claim and the Saudi attack reports. A named vessel boarding, independent imagery or a crew statement would carry more weight than another party claim. Qatar, Oman and Pakistan face a narrower but still viable mediation window. A usable maritime protocol would require specific navigation, escort and verification terms, not only political contact.

NATO governments, particularly Lithuania, Poland, Latvia and Romania, face increased air-defence and attribution pressure. A further airspace entry would confirm recurring hazard, but only a physical impact or casualties on NATO territory would establish a different escalation category. The distinction matters for public attribution and crisis management.

UN agencies, WFP, IOM, OCHA, UNICEF and UNHCR face separate pressure points. Yemen's displacement and health-facility damage threaten access in the west; Sudan's ration and stock warnings create an end-September supply risk; Gaza's shelter, power, fuel and debris-removal gaps create a winter exposure. The agency figures should not be aggregated without reconciling reporting windows.

The Philippine Coast Guard and China Coast Guard face a recurring collision risk around Sabina Shoal, Scarborough Shoal and Second Thomas Shoal. A further collision or injury would materially raise the chance of a treaty-related crisis, while another close encounter without physical contact would support continuation of coercive law-enforcement behaviour below the armed-conflict threshold.

Indicators & Warnings

Hormuz traffic, worsening: Use Kpler and satellite-derived AIS-visible counts for the same category of loaded commodity vessels. The current reference point is four commodity-vessel transits on 17 September. Zero loaded tanker or LNG transits for 72 consecutive hours, combined with a UK Maritime Trade Operations warning or a port departure gap at Qatar, the UAE or Oman, would indicate further deterioration within 14 days.
Hormuz traffic, easing: Use the same Kpler and satellite methodology, not US Central Command's separate escorted-vessel count. More than 10 AIS-visible loaded commodity transits per day for seven consecutive days, with at least three independent port or vessel records and no new UK Maritime Trade Operations strike warning, would weaken the severe-disruption forecast.
Iranian enforcement: An IRGC Navy boarding, inspection, detention, new compulsory permit notice or named vessel refusal would confirm an enforceable transit regime. A second named vessel incident within 14 days would strengthen the Trend assessment. A joint US-Oman or Qatar-Iran navigation notice would indicate deconfliction.
Bab al-Mandab traffic: Use Kpler counts against its own mid-July reference of 47 vessels on comparable days. Fewer than 20 vessels per day for three comparable days, combined with independently verified Ansar Allah positions at Mokha or Mayyun, would confirm further pressure. A verified Yemeni government presence at either location for seven days would weaken the territorial-control assessment.
Saudi target effects: Use Saudi Ministry of Defence releases, airport NOTAMs, flight-tracking data, Aramco operational notices, satellite imagery and independently geolocated video. A verified impact crater, fire or outage at Riyadh, Yanbu, Bisha, Taif, Farasan or a named Aramco facility would confirm increased strike effectiveness within 14 days.
US-Iran pause: A joint US-Iran, Qatar-Oman or Pakistan-Iran statement specifying a cessation of strikes, routing rules, verification arrangements and an implementation date would be a leading indicator of de-escalation. A 72-hour period without a reported strike would provide confirmation. A new US strike followed by an Iranian attack on a US facility would confirm continuation.
Russia-Ukraine strike tempo: Track Ukrainian Air Force reports, Russian Ministry of Defence statements, satellite imagery and municipal casualty releases separately. Russian launches exceeding 150 drones and missiles in one 24-hour period, or Ukrainian attacks damaging a Russian refinery, airfield or logistics hub in two consecutive reporting periods, would confirm continued reciprocity. A 72-hour reduction in rear-area strikes by both sides would challenge the forecast.
NATO spillover: Track NATO, Polish, Lithuanian, Latvian and Romanian defence statements, NOTAMs and radar-based incident reporting. A second confirmed airspace entry within 30 days would raise the frequency assessment. A physical impact, civilian casualty or confirmed weapon effect on NATO territory would materially change the escalation outlook.
Sahel jihadist pressure: Use ACLED event data, FAMa, Africa Corps and Nigerian military releases, local reporting, geolocated imagery and convoy-tracking accounts. Another successful attack on an isolated FAMa, Africa Corps or Nigerian base, or two fuel-convoy attacks on routes into Bamako within 30 days, would confirm sustained operational pressure. Continuous FAMa or Africa Corps reoccupation of Dioura for seven days would weaken the assessment.
Sudan humanitarian deterioration: Monitor WFP ration announcements, IOM stock reports, OCHA access updates and verified fighting around El Obeid and Ad-Damazin. A WFP ration reduction or IOM stock depletion before 30 September would confirm an aid-supply shock. A SAF-RSF humanitarian pause with external monitoring and restored deliveries for seven days would challenge the forecast.
Gaza and southern Lebanon: Use OCHA casualty and access updates, IDF statements, Hezbollah claims, UNIFIL incident reports and independently geolocated imagery. Israeli strikes in Gaza on three consecutive days, or Hezbollah drone or rocket activity near Ali al-Taher or Mount Dov followed by Israeli fire, would confirm continued erosion. A verified 72-hour reduction in fire with expanded heavy-machinery, fuel and shelter access would indicate a material change.
China-Philippines: Track Philippine Coast Guard, China Coast Guard, AIS data, vessel imagery and crew statements around Sabina Shoal, Scarborough Shoal and Second Thomas Shoal. Another collision, water-cannon incident, ramming attempt or injury within three months would confirm the recurring-dangerous-encounter forecast. Firearms use or PLA Navy weapons employment would break the current low-intensity assessment.
Taiwan mobilisation: Monitor Chinese implementing regulations after 1 October, public requisitioning of civilian roll-on-roll-off ships or aircraft, PLA Eastern Theater Command notices and a Taiwan-specific mobilisation order. Requisitioning tied explicitly to Taiwan, combined with a named exercise and logistics movement, would indicate movement beyond a legal preparedness framework. The absence of these indicators would not prove that no classified preparation exists.
Bolivia mobilisation: Monitor the Central Obrera Boliviana, transport unions, road authorities and local media for road closures, strike notices and police-military deployments. A nationwide blockade affecting three or more major corridors before or after 10 October would confirm renewed mobilisation. Continued road access with only local protests would weaken the national-escalation forecast.

Alternatives

The scenarios below are not mutually exclusive. A maritime protocol and a further strike can occur in the same period. Probabilities refer to the individual scenario occurring by its stated horizon, not to a single exhaustive distribution.

Limited US-Iran maritime protocol: Roughly even chance. Qatar, Pakistan and Oman retain mediation channels, and both sides retain incentives to preserve selected energy flows. A protocol covering escorts, permits or designated lanes could raise traffic without resolving the wider war. A joint statement with verification arrangements and three days of compliant vessel movements would distinguish this scenario from continued coercion.
Further tanker or Saudi energy strike: Likely. The operating environment includes the Trend claim, further UK Maritime Trade Operations warnings, Ansar Allah attacks on Saudi targets and disputed Saudi infrastructure damage. A confirmed crew death, major fire or prolonged loss of a Saudi export route would increase the likelihood of retaliatory decisions by the United States or Saudi Arabia.
Ansar Allah retains threat capacity without controlling the whole chokepoint: Likely. Ansar Allah can threaten shipping through missiles, drones, mines and shore-based observation without holding every reported position. Continued attacks alongside no independently verified occupation or logistics chain at Mayyun would support this alternative.
Local Russian or Ukrainian operational gain without a wider breakthrough: Roughly even chance. Conflicting territorial claims around Lyman and Donetsk support the possibility of a local gain by either side while reciprocal deep strikes continue. Geolocated movement of more than 10 square kilometres over seven days, linked to a disrupted supply route or abandoned position, would distinguish a local operational gain from attrition.
Contained Russia-NATO incident: Very likely if a further airspace violation occurs, but the trigger itself is unlikely. NATO and affected states have so far used defensive and diplomatic responses, while Baltic intelligence services assess that Russia lacks the forces for a successful incursion. A physical impact causing NATO civilian casualties would invalidate the contained-incident assessment.
JNIM and ISWAP maintain high activity without a region-wide surge: Likely. Dioura, Mairari and Gajiram demonstrate continued pressure, but the reviewed reporting does not establish a common operational plan or an independently verified simultaneous offensive across Mali, Burkina Faso, Niger and Nigeria. A verified ECOWAS operation or simultaneous multi-front offensive would challenge this alternative.
Temporary SAF-RSF humanitarian pause: Roughly even chance. US, UN, AU and regional diplomacy remains active, while both parties face pressure from food and displacement conditions. A geographically limited pause around El Obeid, Tawila or aid corridors is more plausible than a nationwide ceasefire. Verification by OCHA, WFP and IOM would be required.
Israeli operations continue in Gaza while southern Lebanon exchanges temporarily decline: Roughly even chance. Gaza violence is supported by more extensive reporting than the southern Lebanon incidents, whose casualty and attribution details remain contested. A 72-hour absence of Hezbollah drone or rocket activity near Ali al-Taher and Mount Dov, confirmed by UNIFIL, would show localised de-escalation without changing the Gaza assessment.
China-Philippines encounters remain coercive but avoid another collision: Likely. Both governments continue to frame activity as coast guard or law-enforcement operations, and weapons use is not independently corroborated. Repeated close approaches without physical contact would support this lower-intensity alternative. A collision, water-cannon incident or injury would confirm the higher-risk forecast.
Bolivia experiences localised protests rather than nationwide blockades: Roughly even chance. Diesel prices increased sharply and the Central Obrera Boliviana opposes the IMF programme, but current nationwide road closures are not established. Local demonstrations and isolated corridor closures without three or more major national routes blocked would support this alternative.