International Affairs

Islamabad called it a military target. International humanitarian law calls it a protected facility. The evidence supports one of those positions. At approximately 9 p.m. on 16 March 2026, an airstrike hit the Omid Addiction Treatment Hospital in Kabul, a 2,000-bed drug rehabilitation facility near the city’s international airport, destroying large sections of the building.  Afghanistan’s Interior Ministry confirmed 408 dead and 265 injured. Rescue crews were still recovering bodies from the rubble the following morning. The patients were civilians in medical treatment for addiction. Pakistan’s Information Minister said the air force had carried out precise, deliberate, and professional strikes on military installations and terrorist support infrastructure, that secondary detonations clearly indicated the presence of large ammunition depots, and that no hospital, no drug rehabilitation centre, and no civilian facility had been targeted.  The factual dispute between Islamabad and Kabul has not been independently resolved. The legal analysis does not require it to be, because under international humanitarian law the evidentiary burden does not rest on the victim. It rests on the state that fired. International humanitarian law does not prohibit civilian deaths in armed conflict as such. It prohibits specific categories of conduct, and attacking medical facilities sits near the top of that list. Article 12 of Additional Protocol I requires that medical units be respected and protected at all times. Article 18 of the Fourth Geneva Convention states explicitly that civilian hospitals may in no circumstances be the object of attack. The Rome Statute, in Article 8(2)(b)(ix), classifies intentionally directing attacks against buildings dedicated to medical purposes as a war crime, provided those buildings do not constitute military objectives. Residents and a Reuters journalist present at the site confirmed it was the hospital that was struck, and that the Omid hospital and Camp Phoenix, the former NATO base Pakistan claims to have targeted, were not the same location.  The facility held protected status under four separate instruments of international humanitarian law. Its location beside a former NATO base that had been repurposed by Afghan authorities after 2021 does not extinguish that protection. The central legal question is whether Pakistan can demonstrate that the facility’s protected status had been lawfully forfeited before the strike was ordered. Under IHL the threshold for forfeiture is narrow and procedurally demanding. A medical facility loses its protection only when it is actively used to commit acts harmful to the enemy, not when a state suspects proximity to militants, not when it occupies ground adjacent to a former military installation, but when the facility itself is engaged in hostile military conduct. Even then, a warning must be issued, a reasonable deadline set, and that warning must go unheeded before an attack becomes lawful. Pakistan issued no warning. Its claim that secondary detonations indicated ammunition storage was made after the strike, not before it. Post-hoc assertion is not pre-strike evidence, and the burden of proof rests entirely on the attacking party. Article 50 of Additional Protocol I is explicit: in case of doubt, civilian status is presumed. That presumption applied to the Omid centre. Pakistan made no demonstrated effort to rebut it before firing, which means the strike was unlawful from the moment the order was given. The proportionality and precaution analysis is an independent and equally serious exposure. Article 57 of Additional Protocol I requires commanders to do everything feasible to verify that targets are military objectives, to select means and methods that minimise civilian harm, and to refrain from attacks where civilian losses would be excessive relative to the anticipated military advantage. These are binding obligations, not operational guidelines. The strike occurred at 9 p.m. in a populated district of Kabul, against a 2,000-bed medical facility, with no warning issued to staff or patients.  Pakistan has not defined the military advantage it anticipated, has not quantified it, and has not demonstrated that any proportionality assessment was conducted before weapons were released. The precautionary duties of Article 57 exist precisely to prevent this scenario. They were not discharged. Pakistan’s stated defences do not survive legal scrutiny. The first is that it struck a legitimate military objective, which requires verified pre-strike evidence of hostile use and established forfeiture of protected status. Neither has been demonstrated. The second is that Afghanistan provides sanctuary to Tehrik-e-Taliban Pakistan fighters, giving Islamabad just cause. This is irrelevant to the targeting legality of a specific building on a specific night. The principle of distinction requires attacks directed at identified military objectives, not at territory as collective accountability for the conduct of armed groups operating there. The third, implicit in Pakistan’s public framing, is that Taliban cross-border attacks on Pakistani civilians provide reciprocal justification. That argument was explicitly and permanently rejected at the Nuremberg Tribunals in 1946. Reciprocity does not suspend the laws of war. An adversary’s violations do not authorise your own. On the mental element, the Rome Statute does not require proof that Pakistan intended to kill patients. It requires that the attack be intentionally directed at a protected site, and recklessness satisfies that threshold. A commander who orders munitions onto a compound at night, without verifying it is a lawful military objective, without issuing a warning, when a civilian medical population is foreseeably present, has met the intent standard through recklessness even absent specific malice. On the present public record, every element of the war crime of attacking a protected medical facility is satisfied. The site held protected status. No forfeiture was established. No warning was issued. The proportionality obligation was not discharged. The UN High Commissioner for Human Rights has called for an investigation and for those responsible to be held to account in line with international standards.  That call will almost certainly go unmet. Pakistan is not a party to the Rome Statute, and a Security Council referral would face veto from states with their own unresolved targeting exposure. The legal classification and the probability of accountability are two entirely separate questions. The strike constitutes a war crime. Whether that finding produces consequences depends on

Iran found the gap in missile defence. It’s called a cluster munition. The weapon is banned by international treaty. Neither Iran, Israel, nor the United States signed it. The ban is technically alive and operationally irrelevant. On the night of 5 March, Israelis watching the sky over greater Tel Aviv did not see a single inbound fireball. They saw dozens, organised and steep, descending in unison. What registered on the ground was not one explosion but a dispersal event across multiple city blocks simultaneously. That distinction is the substance of this conflict’s most consequential weapons development, and it is not a new problem. It is a very old one that the international community has spent fifteen years pretending to have solved. A cluster munition is a dispenser, not a warhead. The ballistic missile re-enters the atmosphere and at altitude the casing splits open. Submunitions scatter across a radius of several kilometres with no guidance and no correction after release. Each falls independently, arms on descent, and detonates on impact. Iran’s standard configuration releases between 24 and 80 submunitions per missile, each carrying approximately 2.5 kilograms of explosive. The heavier variants carry a thermal coating allowing survival through re-entry on a long-range trajectory, a design feature that distinguishes Iran’s cluster-configured missiles from shorter-range tactical systems and speaks directly to the intended application. The target area is not a point. It is a neighbourhood. The operational rationale is straightforward attrition management. With its launcher fleet degraded by sustained strikes, Iran faces a diminishing capacity to generate mass salvo effects. Cluster munitions restore per-round efficiency, converting one interceptable airframe into dozens of simultaneously falling submunitions and solving the salvo problem at the munition level rather than the force level. Beyond immediate lethality, the weapon imposes a compounding cost on the defender’s interceptor stockpile, forcing the expenditure of multiple intercepts against submunitions from a single parent missile. Under attrition conditions on both sides, that exchange ratio carries strategic weight. The damage to layered missile defence architecture is structural. Long-range intercept systems are optimised against the parent missile and have no role after cargo release at altitude. Terminal-phase systems can engage individual submunitions but were not engineered for simultaneous saturation across a dispersal field. In documented cases the parent missile has been successfully intercepted while submunitions already released continued to impact. The intercept, by conventional measure a success, produces no protective effect on the ground. This is not a gap that investment in existing intercept technology closes. It is a consequence of the weapon’s physics. Secondary hazard extends lethality beyond the strike itself. Many submunitions fail to detonate on impact and remain on the ground as live ordnance, functionally equivalent to scattered anti-personnel mines. Post-conflict clearance operations in Lebanon, Kosovo, Laos, and Yemen have recovered active submunitions decades after original delivery. The strike ends. The contamination does not, and civilian populations absorb that residual cost for generations. The legal framework governing the weapon is the Convention on Cluster Munitions, concluded in Dublin in 2008 and in force from 2010. It is categorical in its prohibition on use, production, transfer, and stockpiling, with no military necessity clause and no permitted reservations. 111 states have ratified it. Iran has not signed. Neither has Israel. Neither has the United States, which approved five separate cluster munition transfers to Ukraine between July 2023 and April 2024. Neither has Russia, which has deployed them extensively across Ukrainian urban terrain since 2022. Neither has China or Pakistan. The states absent from the convention are not peripheral actors. They are every military power with a serious land war in its operational planning, every state with a meaningful cluster munitions stockpile, and every belligerent currently employing the weapon in an active theatre. The convention’s membership and the world’s military reality occupy almost entirely separate categories. Customary international humanitarian law binds all states regardless of treaty membership and prohibits indiscriminate attacks against civilian areas. Multiple human rights bodies have assessed cluster munition use in populated areas as a war crime. That assessment applies equally to every non-signatory deploying the weapon in an active theatre. Past use in Afghanistan, Iraq, and Laos caused civilian harm that persists to this day. The law does not distinguish between belligerents. The accountability mechanisms do. The Convention on Cluster Munitions did not fail because it was poorly drafted. It failed at the point of design. A prohibition that the United States, Russia, China, and Israel declined to join was never a ban. It was an agreement among states that had already decided not to use the weapon. Iran’s current employment is the most visible expression of that void, not the cause of it. When the architects of the international security order exempt themselves from its humanitarian constraints, the system does not bend. It breaks.

Over Taiwan, silence. Over West Asia, everything. The PLA’s unusual air lull after the Iran war opened is not a puzzle, it is a window into where Chinese strategic attention was actually focused. For anyone who has tracked PLA activity around Taiwan with any regularity, the daily bulletins from the ROC Ministry of National Defense have a familiar rhythm. Since Lai Ching-te’s inauguration in May 2024, Beijing settled into a higher and more continuous pressure tempo than anything seen in prior years. Air incursions became less dependent on political flashpoints and more routine in character. The median line, once treated as a tacit boundary by both sides, was crossed with increasing regularity. Joint air-sea patrols normalised entry vectors across the southwestern, northern, and at times eastern ADIZ sectors. The pattern was deliberate: make previously exceptional behaviour unremarkable enough that no single incident could be cleanly identified as escalatory. By late 2025, analysts tracking monthly incursion data had stopped treating individual sorties as news. That is how normalised the pressure had become. Which is precisely why eleven consecutive days of zero PLA aircraft sorties, from 1 March through 11 March 2026, registers as a pronounced anomaly. This is not the kind of variance that appears in the historical record as routine operational fluctuation. It is a clean break from an established pattern, and it coincides exactly with the opening phase of the U.S.-Israeli military campaign against Iran, which began on 28 February. What the data actually shows The ROC MND bulletins are unambiguous. Between 28 February and 16 March, the PLA recorded 41 aircraft sorties around Taiwan. Twenty-six of those, 63 percent of the entire period’s total occurred on a single day, 15 March. For the preceding eleven days, the number was zero, broken only by a probe of two sorties on 7 March. The PLAN surface fleet, by contrast, never stood down. Six to eight vessels maintained their patrol pattern every single day throughout the same period, producing 105 ship-presence counts across the window. That air-maritime dissociation is the critical data point. A general stand-down would have reduced both. A logistical pause would show variance across both instruments. What the data shows instead is a deliberate choice to maintain Taiwan presence through lower-escalation naval assets while withdrawing the more operationally demanding and politically consequential air component. Ships hold position. Aircraft cross lines. Why West Asia absorbed the attention To understand where that attention went, it is necessary to understand what Beijing had at stake in the conflict it was watching. China is Iran’s largest trading partner and purchases an estimated 80 percent of Iranian oil exports. More than 55 percent of China’s total oil imports in 2025 came from the West Asia, approximately 13 percent from Iran itself, most of which must pass through the Strait of Hormuz. The 2021 Comprehensive Strategic Partnership between Beijing and Tehran, a 25-year agreement, positioned Iran as a critical node in the Belt and Road Initiative’s overland corridors connecting East Asia to Europe, corridors that cannot function effectively without stable access through Iranian territory. China is Iran’s largest trading partner, and the current conflict effectively closed the Strait of Hormuz, an energy artery where approximately 50 percent of Chinese energy imports transit. Beijing’s exposure was not merely economic. China became Iran’s largest trading partner, importing significant oil volumes and providing technological support, including radar systems and navigation tools. Reports emerging from the conflict’s opening weeks indicated that Chinese material contributions, including spare parts for missile systems were helping Tehran sustain defensive operations, stopping well short of combat involvement but material enough to attract U.S. attention. Analysts described these contributions as enabling Iran to sustain its defenses without escalating to a broader confrontation involving Beijing. Whether China was learning from the conflict or, in more limited ways, actively participating in it through a supporting role to Tehran, the strategic calculus was the same: West Asia in the opening weeks of March 2026 was where Chinese interests were most immediately at risk, and where the most consequential decisions about those interests were being made in real time. What a live war offers that no exercise can Separately from its direct stake in Iran’s survival, Beijing had an observational interest in the conflict that ran independently of the outcome. The U.S.-Israeli campaign represented the most significant live test of American air, naval, and command systems in a high-intensity environment since at least 2003. How carrier strike groups manage multi-axis saturation threats. How command networks absorb simultaneous kinetic, electronic, and information operations. How the decision chain from theatre commander to national command authority actually functions under real time pressure, not scripted exercise conditions. These are the precise questions that PLA modernisation; anti-access systems, carrier-killer missiles, electronic warfare architecture, the entire military competition with the United States in the Pacific has been built around for two decades. A live conflict involving U.S. forces against an adversary operating layered air defences, ballistic missiles, and drone swarms provides data that no simulation generates. A Chinese geospatial intelligence company, MizarVision, was already circulating annotated satellite imagery of American bases in Jordan, Qatar, and Saudi Arabia within days of the campaign opening carrier positions, runway activity, logistics signatures at forward operating sites, machine-sorted and distributed in near real time. The collection infrastructure was active from the first hours. Running high-profile air operations over Taiwan while that observational window is open is a poor allocation of military attention. The eleven-day air lull was not restraint for its own sake. It was concentration. What is usual and what is not Here it is worth being precise. Periodic dips in PLA air activity are not unprecedented. Operational cycles, weather, and political calendars have all historically produced short gaps. What is unusual is the duration, the timing, and the context together. An eleven-day complete absence, coinciding with the opening of a major conflict directly involving China’s most significant West Asian partner and a live test of its primary strategic adversary, falls outside any normal operational variance.

India’s Defence Forces Vision 2047 signals something rarer than modernisation, a change in how the military thinks about war itself. Something shifted when India released Defence Forces Vision 2047. Not the release itself. Long-range planning documents are neither rare nor automatically consequential. What shifted was the register. This is not a document about what India wants to buy. It is a document about what kind of military power India intends to become, and why that question can no longer be deferred. For decades, Indian defence planning operated within a particular institutional grammar. Threats acknowledged obliquely, ambitions framed modestly, modernisation treated as a procurement exercise rather than a strategic project. Vision 2047 breaks from that tradition with unusual directness. It places the armed forces at the centre of India’s emergence as a developed nation, argues that economic and military power are not parallel ambitions but co-dependent ones, and states plainly that a Viksit Bharat which cannot secure its trade routes, defend its borders, or resist coercion below the threshold of open war is not, in any meaningful sense, developed. That is not a bureaucratic formulation. It is a declaration of how India now understands the relationship between national power and national security. The document’s ambition on jointness alone represents a historic shift in institutional intent. Theatre commands, integrated logistics, tri-service doctrine, a joint operations coordination centre. These ideas have circulated in Indian defence circles for the better part of two decades. Seeing them anchored in a formal long-range vision, with new institutional bodies proposed to carry them forward, signals that the conversation has moved from aspiration to architecture. The distance between those two things is enormous, and crossing it begins with exactly this kind of formal commitment. What distinguishes Vision 2047 most sharply from its predecessors is that it thinks about the nature of war itself. It does not simply list formations to be restructured or platforms to be acquired. It grapples seriously with AI, autonomous systems, quantum technologies, hypersonics, and cognitive operations, and asks what kind of institution India must build in response. It recognises that future conflict will be multi-domain, that the line between peace and war has effectively dissolved, and that the adversary of 2047 will not be defeated by the organisational logic of earlier decades. The most ambitious claim in the paper is also its conceptual spine: that warfare is evolving from network-centric to data-centric and ultimately to intelligence-centric models, and that India intends to build its future force around that trajectory. The destination is right. The framing rewards closer examination to appreciate what it is actually reaching for. Network-centric warfare, as it was theorised in the late 1990s, was always about converting informational advantage into decision advantage. Data centricity was not a later stage of that idea. It was the original premise. What Vision 2047 is pointing at, more precisely, is the collapse of decision timelines. The compression of the entire sensor-to-shooter cycle to machine speed, across every domain simultaneously. That is the real rupture that AI, autonomous systems, and edge computing are now producing in military competition. Find, fix, decide, strike, before the adversary can move, disperse, or retaliate, at speeds that exceed human cognition. The document senses this clearly. Intelligence-centric warfare is the right direction of travel. It now needs operational definition, intelligence for what decisions, at what echelon, against which adversary, to drive the specific force structure choices that must follow from it. That work lies ahead, and Vision 2047 has created the mandate to do it. Equally significant is the document’s insistence on intellectual sovereignty. It calls for shedding colonial institutional practices and building a strategic culture rooted in Indian knowledge, Indian geography, and Indian threat realities. The argument is that a genuinely self-reliant military must also be self-reliant in thought. Borrowed frameworks produce borrowed outcomes, and Indian doctrine built on foreign templates will always fit imperfectly. This is a more radical proposition than any of the new commands or agencies the paper proposes. A Cyber Command can be stood up by notification. A genuinely native strategic culture takes a generation to build. Vision 2047 names that project and takes ownership of it. The three-phase roadmap, transition by 2030, consolidation by 2040, excellence by 2047, is sequential. Restructure first, integrate second, mature into a world-class force third. What matters is that the sequencing reflects a genuine understanding that transformation of this scale is not an event but a sustained institutional process, one that must survive budget cycles, government changes, and the friction of organisations that resist their own reinvention. Most defence establishments, when confronted with the pace of change in modern warfare, default to hardware. Platforms are concrete. Paradigms are not. India has chosen to lead with the paradigm, to ask what kind of war is coming before asking what to build for it. That choice, embedded formally in a long-range vision document, changes what is possible in every planning conversation that follows. Vision 2047 does not solve India’s defence challenges today. It does something arguably more important. It reframes them. Transformation of this kind begins not in the procurement cycle or the budget, but in the willingness to say clearly what you are building toward, and why. India has said it out loud. That is where it starts.

How Handala’s double-strike operation against Stryker combined mass data theft with deliberate destruction, and why that combination signals a dangerous new phase in state-aligned cyberwarfare.   Steal Everything, Then Burn It Down When Handala claimed responsibility for the cyberattack on Stryker Corporation last week, the group did not simply boast about wiping tens of thousands of devices. Buried in its statement was the detail that made security analysts sit up straight: before destroying anything, the group says it pulled out 50 terabytes of data. That sequence, extraction followed by annihilation, is not accidental. It is a deliberate two-phase architecture that tells us a great deal about how Iran-aligned threat actors have matured their operations. To understand why the order matters, it helps to think about what each phase accomplishes independently. Exfiltration is about intelligence and leverage. Wiping is about punishment and disruption. Done separately, each is effective in its own way. Done together, in that specific sequence, they become something considerably more dangerous than the sum of their parts. The exfiltration phase would have been the quieter, more technically demanding portion of the operation. Moving 50 terabytes of data out of a corporate network without triggering alerts requires patience, planning, and usually an extended period of undetected access known as dwell time. Attackers staging an operation of this scale typically establish a foothold through a phishing campaign, a compromised credential, or exploitation of an unpatched vulnerability. From there, they map the network, identify high-value file stores, and begin staging data in a location they can reach from the outside. Exfiltration itself is often done slowly, in chunks, timed to blend with normal business traffic. Fifty terabytes is an enormous volume. For context, that is roughly equivalent to fifty million documents or the entire text contents of a large university library. Moving that much data without detection is a significant technical and operational achievement. The choice of target within the Microsoft environment is also telling. Microsoft 365 environments typically contain email archives, SharePoint document libraries, Teams conversation histories, and OneDrive file stores. For a company like Stryker, which holds a 450 million dollar Department of Defense contract and operates across 79 countries, those repositories could contain procurement details, military supply chain documentation, internal communications about government relationships, and personnel records. A harvest of that breadth would be operationally valuable to Iranian intelligence independently of anything that came afterward. Once exfiltration was complete, the destruction phase began. Wiper malware, the tool believed to be responsible for the device erasure employees encountered, is fundamentally different from ransomware in its intent. Ransomware encrypts data and demands payment, meaning the attacker has a financial incentive to keep the data recoverable. A wiper has no such incentive. It is designed to overwrite data at a low level, often targeting the master boot record or file allocation tables to make recovery effectively impossible even with forensic tools. The fact that remote employee devices running Windows, including laptops and mobile phones connected to Stryker systems, were wiped suggests the malware propagated through the Microsoft environment rather than requiring physical access to each machine. The psychological and operational logic of wiping after exfiltration is layered. First, it destroys forensic evidence of how the attackers got in and what they accessed, complicating any attempt to understand the full scope of the breach. Second, it maximises disruption to the victim organisation, forcing a costly and time-consuming rebuild of infrastructure. Third, it sends a public signal, in this case a geopolitical one, amplified by Handala displaying its logo on affected login screens. The destruction is the message. The data theft is the prize collected quietly before the message is delivered. Security researchers have noted that this steal-then-wipe pattern has been appearing with greater frequency in state-aligned operations. It was visible in Russian operations against Ukrainian targets before and during the 2022 invasion, and Iran-linked groups have been refining similar techniques over the past several years. What distinguishes the Stryker incident, if Handala’s claims are even partially accurate, is the claimed scale of data removed before the wiper was triggered and the apparent reach into devices across dozens of countries simultaneously. The healthcare dimension of the attack also deserves technical scrutiny. Stryker’s Lifenet system, which transmits electrocardiogram data from ambulances to hospitals, was knocked offline in Maryland and potentially elsewhere. This is not a consequence of targeting medical devices directly. It is a consequence of the underlying Microsoft network infrastructure being disrupted, which in turn severed the connectivity that medical systems depend on. Modern hospitals and medical device networks are deeply entangled with corporate IT infrastructure, and that entanglement creates cascading failure pathways that attackers may not even need to deliberately target. The broader lesson from the technical architecture of this attack is that the most damaging cyberoperations are no longer defined by a single dramatic action. They are defined by sequencing. Getting in quietly, staying long enough to extract what is valuable, and then detonating on the way out. That model requires a higher level of operational discipline and technical capability than a simple smash-and-grab. The fact that a hacktivist group with documented ties to Tehran appears to be executing it at this scale suggests that the line between loosely organised hacktivism and structured intelligence operations has become very thin, or perhaps no longer exists at all.

The next soldier will not bleed, will not tire, and will not hesitate. It is already being built, and the race to send it to war is underway. In late January 2026, three Russian soldiers emerged from a destroyed building to surrender. There was no Ukrainian infantryman waiting for them. There was an armed ground robot, holding the position. The humans were already behind the line. That moment was not a military curiosity. It was a marker of where war is heading, and how fast it is getting there. When U.S. and Israeli forces struck Iran in February 2026, AI was embedded across the entire operation, from target identification to guiding autonomous drones through GPS-denied, signal-jammed environments. Nearly 900 strikes in the first 12 hours, a tempo no previous conflict had achieved.Two wars. Two continents. Same conclusion. The age of AI war is not arriving. It is already here. While Ukraine remains the world’s most consequential testing ground for autonomous war, its front line increasingly held not by soldiers but by machines and the skeleton crews that control them, Iran has shown what the next level looks like in combat. In the strikes on Iran, air defense networks, drone salvos, and electronic warfare operated simultaneously across multiple theatres at a speed and complexity that compressed years of strategic assumption into days.  In both wars the pattern is identical. The human body has become the most vulnerable object in modern war. The machine has become the primary fighter. The soldier has become support. Every serious military establishment on earth is watching, and accelerating. What they are accelerating toward is a new generation of bipedal robots designed to do what a soldier does. Carry weapons. Breach doors. Move through terrain. Hold a position. Resupply under fire. The most advanced can pick up and operate rifles, pistols, shotguns, and grenade launchers already in service across existing armies. The design logic is deliberate. Decades of weapons, vehicles, and military infrastructure have been built for human hands and human bodies. A robot engineered to fit that existing architecture requires no new logistics chain. It steps into one already built. Ukraine proved these systems endure. Iran proved they can decide. The most advanced humanoid built explicitly for war is the Phantom MK-1, developed by Foundation, a San Francisco startup with U.S. Army, Navy, and Air Force research contracts and approved military vendor status. At 5 feet 9 inches and 180 pounds, it is designed around one principle: operate with everything a soldier already carries. Two units are currently on reconnaissance trials in Ukraine. The Marine Corps is training them on breach entry, placing explosives on doors so troops stay back from the fatal funnel. Current per-unit cost sits at approximately $150,000, projected to fall below $100,000 by 2028 and below $20,000 at scale. Production targets for 2026 stand at 10,000 units, scaling to between 40,000 and 50,000 by end of 2027. At that price a robot battalion becomes economically competitive with a human one, without the casualties, the trauma, or the political cost of repatriated bodies. The United States is not alone in this. Anduril, founded by Palmer Luckey, builds autonomous drone interceptors, electromagnetic warfare systems capable of collapsing enemy drone swarms, and the Ghost Shark, a fully autonomous submarine already operational with the Australian Navy. Scout AI demonstrated in February 2026 a complete autonomous kill chain in which seven AI agents identified, located, and neutralized a target with no human involvement at any stage. Boston Dynamics, majority owned by Hyundai, has been testing its Atlas bipedal robot in environments with direct military adjacency since 2021. Figure AI is developing general purpose humanoids with clear dual-use potential. China’s People’s Liberation Army has been funding humanoid robotics research through state institutions including Beijing Institute of Technology and Zhejiang University since at least 2015. Russia is developing dual-use platforms under direct military sponsorship, with the Central Research Institute for Robotics and Technical Cybernetics in St. Petersburg among the primary state facilities. Iran unveiled Aria, a domestically built autonomous combat robot, in September 2025, built entirely under international sanctions. Goldman Sachs projected between 50,000 and 100,000 humanoid robots shipping globally in 2026 alone. Morgan Stanley forecasts the total humanoid market exceeding $5 trillion by 2050. The largest share of that growth is in defense. Every major power is building. None are waiting. While the race accelerates, the technology has real distance left to travel. A humanoid moves through roughly 20 individual motors, each a potential failure point under combat stress. The platforms are heavy, power dependent, and not yet proven against sustained rain, mud, extreme cold, and kinetic impact. A captured or compromised humanoid is not simply lost equipment. It carries intelligence, has potential software access points, and could in the wrong hands be turned. These are engineering problems, and engineering problems get solved. Expert consensus places initial combat deployment at two to three years for leading platforms, with broader fielding across multiple militaries by the early 2030s. The harder problem is judgment. International Humanitarian Law requires that any use of force distinguish between combatants and civilians, that it be proportionate, and that all feasible precautions be taken to avoid civilian harm. These obligations do not change because the trigger is pulled by a machine. But in both Ukraine and Iran that standard is already under pressure. In Ukraine, when communications are jammed, drones default to onboard AI targeting because the operational alternative is paralysis. In Iran, AI systems processed and prioritised over a thousand targets at a speed no human oversight structure was built to match. These are black box decisions, made by opaque models running on algorithms whose reasoning cannot be audited, reconstructed, or explained after the fact. The law says one thing. The war is doing another. That gap is where the most consequential argument of this era is playing out. The United Nations and the International Committee of the Red Cross have jointly called for a binding treaty prohibiting autonomous weapons that operate without

AI-enabled geospatial intelligence is reshaping military competition, but the revolution is more precise, and more dangerous, than its most dramatic advocates suggest.   In February 2026, as the United States accelerated its military build-up around Iran, a Chinese geospatial intelligence company called MizarVision began circulating satellite imagery that was striking not for its resolution but for its organisation. High-definition photographs of American bases in Jordan, Qatar, and Saudi Arabia were annotated, labelled, and formatted into something closer to an operational briefing than a news photograph. Carrier positions, runway activity at al-Udeid and Ovda, logistics signatures at forward operating sites: all machine-sorted and publicly distributed in near real time. The episode attracted commentary framing it as proof that a new era of warfare had arrived, one in which the first shot would be fired not from a runway or missile silo but by an algorithm that had already identified every target from orbit. That framing captures something real. But it also obscures the more precise and more actionable strategic truth. The Shift That Actually Happened Satellite imagery is not new. What is new is that machine vision can now process it at a scale and speed no human organisation can match. For decades, the constraint on imagery intelligence was not collection but interpretation. Trained analysts, compartmented bureaucracies, and slow dissemination pipelines throttled the operational value of what satellites could see. That bottleneck has moved. Computer vision systems can now detect, classify, and disseminate findings across a commercial imagery feed faster than any national intelligence centre. The significance of this shift is not merely technical. It is structural. What once required a state-level investment in imagery exploitation infrastructure can increasingly be approximated by software stacked on commercial cloud architecture and trained on open-source data. The private sector is no longer supplying pictures. It is supplying theatre awareness, and it is supplying it to anyone with the access and the model. This produces a political asymmetry that Western governments have been slow to absorb. During the Ukraine war, commercial satellite imagery was widely celebrated as a democratising force: a check on Russian operational deception, a tool for accountability, a structural advantage for the transparency-friendly side of the conflict. The implicit assumption was that open-source overhead surveillance naturally served liberal democratic interests. MizarVision and the Chinese geospatial intelligence infrastructure behind it dispose of that assumption cleanly. Structural capabilities do not carry ideological loyalty. The West helped normalise the instrument. It cannot now monopolise its use. Where the Argument Runs Ahead of the Evidence The leap from “AI can classify a carrier strike group” to “an algorithm fires the first shot” is too large, and the gap between them matters strategically. Commercial imagery at fifty-centimetre resolution can confirm the presence of an F-35 on a ramp. It cannot reliably determine its fuel state, weapons load, maintenance window, or sortie schedule. The difference between knowing an asset exists and knowing precisely when it is vulnerable is the difference between surveillance and targeting. Conflating them overstates the current operational readiness of these systems and, more importantly, misidentifies where the real danger lies. Authorisation also still lies with humans. In every major military doctrine, American, Chinese, and Russian alike, kinetic action requires a political decision. An algorithm that has identified every target from orbit still requires a commander, and ultimately a head of state, to act on that picture. The compression of the sensing-to-understanding timeline does not automatically compress the understanding-to-decision timeline. Institutional friction, escalation calculus, and signalling dynamics all intervene. Treating the kill chain as a purely technical problem mistakes software capability for strategic agency. Countermeasure adaptation is also systematic, not absent. Emissions control, signature management, deliberate deception, dispersal, and spoofing are active and evolving military disciplines, and they are evolving in direct response to persistent overhead surveillance. The question is not whether these countermeasures exist but whether they remain sufficient at scale. That is a genuine empirical question, and one that defence establishments are actively working through. Presenting militaries as passive recipients of surveillance pressure misreads both their doctrine and their investment priorities. The Risk That Is Actually There The genuine strategic danger is narrower than autonomous AI warfare, and in some ways more insidious for it. AI-enabled geospatial intelligence shortens the window between a commander’s decision and the availability of an actionable targeting picture. It reduces the cost of continuous monitoring for state and non-state actors alike. And it creates new pathways for sensitive operational intelligence to enter public or adversarial channels through commercial intermediaries who are not bound by the classification architectures that have traditionally governed such material. These are serious problems. They demand serious investment in signature management, commercial imagery governance, and a harder-edged understanding of escalation dynamics in a world where both sides can observe each other’s preparations in near real time. But they do not support the conclusion that interpretive speed alone determines who prevails in future conflict, or that the pre-kinetic phase has already superseded the kinetic one. The algorithm does not fire the first shot. It narrows the decision space of the human who does. In a world moving as fast as this one, that is consequential enough to deserve serious attention on its own terms, without inflation.

Iran always knew this day would come. For two decades it built a warfare architecture that could not be centred, could not be decapitated, could not be won from the air. On the morning of March 8, 2026, black rain fell on Tehran. The Iranian capital was engulfed in a cloud of toxic smoke that unleashed oil-tainted rainfall dozens of miles away after overnight Israeli strikes hit several fuel depots, causing fires to burn for hours. Four oil depots and a petroleum products transfer center in the Tehran and Alborz provinces were under Israeli fire and damaged, and four personnel, including two oil tanker drivers, were killed. By 10:30 in the morning, cars on Valiasr Street, Tehran’s main north-south artery, still needed their headlights on to navigate the darkness. It was a catastrophic image, and it was designed to be one. But here is what the architects of this air campaign may be miscalculating: Iran was not built to survive this war from the top. It was built to survive it from the bottom. This is the Mosaic Defence, and it is arguably the most consequential military framework to emerge from the Middle East in the past two decades. Its origins trace back to 2009, when then-IRGC Commander Mohammad Ali Jafari formally reorganised the Revolutionary Guards around a single, haunting lesson drawn from watching American military power eviscerate two neighbouring states. Afghanistan fell in weeks. Baghdad collapsed in three. In both cases, destruction of centralised command produced almost immediate systemic failure. Tehran incorporated those lessons: don’t fight the enemy’s preferred war. The US advantage is high-end airpower, precision strikes, and intelligence dominance. Mosaic Defence tries to make those strengths less decisive by ensuring there is no single headquarters, city, or leader whose loss collapses the fight. The architecture that emerged is methodical. Each of Iran’s 31 provincial IRGC commanders operates with his own weapons arsenal, logistics chains, intelligence services, and Basij militias, explicitly trained to make independent military decisions, plan attacks, and wage guerrilla warfare without consulting Tehran. The formal language inside IRGC operational culture refers to this as the “operational autonomy protocol,” triggered automatically when central command goes dark. Iranian Deputy Defence Minister Reza Talaeinik confirmed publicly that each figure in the command structure has named successors stretching three ranks down. You kill the general, his brigadier already has orders. You kill the brigadier, the colonel carries on. On March 1, after Israeli strikes killed Supreme Leader Ayatollah Ali Khamenei, Foreign Minister Abbas Araghchi posted on X in direct, unflinching terms: “Bombings in our capital have no impact on our ability to conduct war. Decentralised Mosaic Defense enables us to decide when, and how, the war will end.” It was not bravado alone. It was a precise articulation of a deeply embedded strategic posture. The Basij is the human tissue that holds this organism together. Established in 1979 by Ayatollah Khomeini as a people’s volunteer force and now operating as a subsidiary arm of the IRGC, its estimated one million members form the paramilitary backbone beneath the Revolutionary Guard’s 150,000 professional troops. In the coastal provinces, “Ashura” and “Imam Hussein” battalions are organised in towns to operate autonomously, defending designated geographic areas, leveraging proximity to logistics centers and coastal road networks to ensure flexible, rapid movement of combat assets between sectors. These are not conscript armies waiting for radio orders. They have pre-assigned mission packages. They know their terrain the way a farmer knows his field. The strategic calculation is brutally simple: to defeat Iran, you do not take Tehran. You take 31 separate, motivated, geographically embedded armies simultaneously. Operationally, this manifests in ways that have already unnerved American planners. In February’s “Smart Control” exercises in the Strait of Hormuz, IRGC fast-attack craft swarmed in coordinated patterns, electronic warfare systems blinded radars, and decentralised orders were executed without central authorisation. This is the rehearsal. The Strait, through which roughly a fifth of the world’s traded oil passes, is now overseen not by a single naval command in Tehran but by distributed coastal units that can independently initiate harassment, mining, or blockade operations. The UAE’s Ministry of Defence reported intercepting over 1,400 drones, eight cruise missiles, and 238 ballistic missiles from Iran in under a week of conflict. Some of that volume reflects this posture, not desperation: swarm the adversary’s interception capacity until something gets through. The darker edge of this framework is its unpredictability under pressure. While disciplined elite units will sustain coherent operations, less experienced units will fall victim to confusion and disorder, raising the risk of uncoordinated strikes and navigation errors that could trigger unintended escalation. The Iranian drone that reportedly struck Oman, a country actively mediating ceasefire talks, illustrated exactly this: autonomous units operating on pre-issued orders with no one in Tehran in a position to call them back in real time. The oil rain over Tehran, apocalyptic as it appeared, does not break this system. Iran’s oil distribution company confirmed that despite the strikes, sufficient gasoline reserves remained. Fuel disruption to a city of ten million is a genuine hardship and a psychological blow. But Mosaic Defence was never designed around keeping Tehran’s refineries lit. It was designed around the premise that even if Tehran burns, Khuzestan fights, Isfahan launches, and the IRGC navy at Bandar Abbas decides on its own when to close the Hormuz chokepoint. The question the US and Israel face is not whether they can win a battle. It is whether there is a battle to win. You cannot break a mosaic; you can only rearrange its pieces. And the pieces, right now, are fighting on their own.

Beijing did not send troops to West Asia. It sent a marketing clip. Source: http://eng.mod.gov.cn/2025xb/D/V/16446709.html On 6 March 2026, as Iranian Shaheds continued to breach air defences across six Gulf Cooperation Council (GCC) states simultaneously, China’s Ministry of National Defence posted a 35-second clip on its official English-language website. With few sentences about detecting “low-altitude, low-speed, and small aerial targets such as drones.” The timing was surgical. The product was not. This is what Chinese defence marketing looks like in 2026: exploit a live war, insert an unproven system into a panic-driven procurement conversation, and bank on customers too frightened, too indebted, or too technically unsophisticated to ask the right questions. War That Created the Window On 28 February 2026, Israel and the United States struck Iran’s military infrastructure under Operation Roaring Lion and Operation Epic Fury. Tehran answered within hours. Operation True Promise IV sent ballistic missiles and UAS simultaneously into Bahrain, Qatar, Kuwait, Saudi Arabia, Iraq and the UAE. Within 36 hours, all six GCC states had been struck by Iran. No drill. No simulation. Nightmare Gulf war planners had war gamed for twenty years arrived at once. Iran’s UAS campaign did not relent. By 5 March, UAE alone had tracked 1,072 inbound UAS and 196 ballistic missiles. On that single day, 131 aerial threats were engaged over Emirati airspace. Iran’s Shahed variants, types 136, 107, and 238, constituted the bulk of confirmed rounds. The economics were catastrophic for defenders. Gulf interceptors ran between three million and twelve million dollars a shot. A Shahed costs hundreds. Iran could sustain the arithmetic indefinitely. Gulf capitals could not. Defence ministries across Asia, Africa, and the West Asia drew the same conclusion simultaneously: counter-UAS capability was no longer optional. They needed a system. They needed to procure one publicly. They needed it now. Beijing had been waiting for precisely this moment. What the Release Actually Says The MND release is worth reading with forensic care. The Radar-Video Fusion Platform, it states, “combines radar and video means” and is “capable of guiding the video system to conduct real-time tracking once targets are detected by radar.” It identifies “moving ground targets within the designated area” and “low-altitude, low-speed, and small aerial targets such as drones” as its detection targets. Strip the language and what remains is this: a fixed post, a radar that cues a camera, operating within a bounded area. The system detects and tracks. It does not intercept. It does not jam. It does not kill. No engagement range. No reaction time. No kill mechanism of any kind. This is the front end of a kill chain presented without the kill chain. Against 131 inbound Shaheds in a single operational day, a border camera that hands off to a video tracker is not a counter-UAS solution. It is a perimeter sensor with a marketing budget. PLA Combat Record That Should End the Conversation The question of whether Chinese military technology performs under fire is no longer theoretical. It has been answered, repeatedly, in the field, by China’s own export customers. Operation Sindoor, May 2025. Pakistan deployed its Chinese-supplied air defence grid against Indian Air Force strikes. Chinese-made HQ-9 and HQ-16 surface-to-air missile systems failed to intercept a single incoming missile. The YLC-8E anti-stealth radar at Chunian Air Base was destroyed. Wing Loong-II UAS were shot down by Indian air defences. Indian Rafale jets using SCALP precision missiles bypassed the Chinese-supplied grid entirely. PL-15 air-to-air missiles fired by Pakistani J-10C jets either missed or malfunctioned, with some reportedly landing in Indian territory. Pakistan’s defeat was total. Its arsenal was 81 percent Chinese-supplied. The pattern did not begin in 2025. Myanmar grounded the majority of its Chinese-supplied jets due to radar defects and unresolved structural faults years after delivery. Nigeria returned seven of nine Chengdu F-7 fighters to China for urgent repairs after a series of crashes, then abandoned the fleet entirely and purchased Italian M-346 aircraft instead. Pakistan’s F-22P frigates reported radar degradation, engine overheating, faulty Gimbal Assembly motors, and compromised missile guidance. Chinese manufacturers acknowledged the defects and declined to repair them on any workable timeline. Saudi Arabia acquired China’s SkyShield laser counter-drone system. In desert operational conditions it experienced significant performance degradation. A laser counter-drone platform that fails in desert heat is not a serious military proposition. This is not a pattern of isolated incidents. It is a pattern of systemic failure across platforms, across countries, across years. A Camera on a Stick China’s approach to military exports relies on perception management over battlefield performance. Advanced-looking systems. Orchestrated reveals. English-language portal releases timed to maximum global anxiety. The 6 March video is the template made visible: a border post dressed as a solution, a sensor dressed as a kill chain, published at the precise moment that counter-UAS procurement panic was highest in recorded history. Radar-Video Fusion Platform may perform adequately on a quiet frontier against a lone surveillance UAS in permissive conditions. That is what it was built for. It was not built to operate inside a Shahed saturation campaign. It cannot engage. It cannot degrade. It cannot stop a single inbound round. Against 131 aerial threats in a single day it can watch and record them arriving. In the Gulf war of 2026, that is not a military capability. It is a camera on a stick. The release was not written for engineers. Any competent defence engineer notes the absence of an engagement mechanism, reads “within the designated area,” and closes the browser. It was written for procurement officials in anxious capitals under political pressure to show populations that something is being acquired. In that market, Beijing is not selling a solution. It is selling the appearance of one. Based on the record from Islamabad to Lagos to Naypyidaw, the customers are still buying. They just keep finding out what they actually paid for.

By striking the Gulf rather than concentrating on Israel or US targets, Tehran has taken the war to the one geography where economic contagion is immediate and globally felt. The war began at 02:14 Tehran time on 28 February. Operation Roaring Lion and Operation Epic Fury decapitated Iran’s leadership, gutted its air defence architecture, and announced to the world that the post-1979 regional order was finished. Washington and Tel Aviv pulled the trigger. What followed was not what either capital planned for. Iran did not fold. A regime collapse as anticipated never materialised. In six days, the Islamic Republic of Iran has fired over 400 ballistic missiles and close to 1,000 drones across the Persian Gulf. The targets tell the story: Dubai International Airport, Jebel Ali port, Ras Tanura refinery, Hamad International in Doha, the US Fifth Fleet headquarters in Bahrain. Israel, by comparison, has absorbed a fraction of that volume. Tehran has made its calculus explicit. This is not a war Iran intends to fight against Israel or US. It is a war Iran intends to fight through the Gulf. This is not improvisation. It is strategy. Military planners have long understood doctrines where targeting civilian infrastructure to maximise chaos, delegitimise adversaries, and accelerate international pressure for a ceasefire. Iran has now adopted the same logic. Airports, ports, oil facilities, and luxury hotels are not military objectives in the conventional sense. They are the load-bearing pillars of Gulf economic identity. Iran is not trying to destroy them. It is making them ungovernable under fire, forcing every airline, insurer, and sovereign wealth fund to recalculate their exposure. The Strait of Hormuz closure clinches the argument. Twenty per cent of global oil and gas passes through that chokepoint. It is now shut. Oil markets have spiked. Bond yields are moving. Iran does not need to win a single air engagement to prosecute this strategy. It needs only to sustain pressure long enough for the global economic cost to exceed Washington’s political appetite for the campaign. The GCC’s Trilemma The Gulf states spent three years doing what may seem “everything right”. Qatar mediated. Oman back-channelled. Riyadh quietly denied offensive basing rights to US and Israeli aircraft. These were calculated risks taken at real political cost. None of it mattered. Iran has now struck every GCC member state within a single operational sequence. This is historically unprecedented. The Gulf’s core brand proposition, stability as a geopolitical product, has been punctured. The GCC faces a trilemma. They cannot retaliate alongside Israel without catastrophic domestic legitimacy costs; the Palestinian cause remains the organising moral framework for Arab public opinion, and any ruler seen fighting Tehran on Tel Aviv’s behalf risks street-level pressure that authoritarian stability cannot easily absorb. They cannot remain passive indefinitely as their cities burn. And they cannot negotiate from visible weakness without undermining the deterrent architecture they have spent hundreds of billions constructing. Interception is holding for now. But the economics of missile defence are brutally asymmetric. Iran’s Shahed-class drones and short-range ballistic missiles cost a fraction of the interceptors being expended against them. Analysts assess Iran can sustain the current rate of fire for approximately one month. The stockpile problem is real and worsening daily. The Decentralisation Problem Iran’s military planners anticipated regime decapitation. The IRGC has shifted to a mosaic defence posture, dispersed launch cells, mobile platforms, decentralised command. B-2 strikes on fixed facilities cannot suppress a doctrine built on mobility and redundancy. CENTCOM claims 17 Iranian naval vessels destroyed and Iran’s conventional air force eliminated. What it has not suppressed is the IRGC’s distributed short-range strike capacity, because that capacity was designed precisely to survive this campaign. Killing Khamenei removed the supreme decision-maker. It did not remove institutional will. Larijani’s Interim Leadership Council, announced on 1 March, confirmed what analysts feared: the succession mechanism is functioning and signalling continuity, not collapse. Real Target Was Always the Global Economy Strip away the operational detail and Iran’s strategy reduces to a single proposition. It cannot defeat Israel or the United States militarily. What it can do is make the cost of this war prohibitive for everyone else. By striking the Gulf rather than concentrating on Israel or US targets, Tehran has taken the war to the one geography where economic contagion is immediate and globally felt. Dubai going dark sends a signal that no missile hitting Tel Aviv can replicate. Gulf states are not Iran’s enemy. They are Iran’s instrument of leverage. Tehran is squeezing them to make their pain loud enough to force Washington’s hand. Trump’s contradictory signals on 1 March, simultaneously announcing Iranian negotiation outreach and claiming Iran has no navy remaining, suggest the economic pressure is already registering at the White House. GCC understands the game Iran is playing. The question is whether they can hold their nerve, sustain their defences, and resist being drawn into a military response that would finally give Iran the regional war framing it needs. That answer will determine not just this war’s outcome, but the shape of West Asia for a generation.