A Dependency Investigation (#06) – May, 2026
About This Series
The Magician’s Hands is a series of dependency investigations. Each report examines a single case in which a structural dependency, between a state and an infrastructure owner, a farmer and a seed company, a continent and an energy supplier, was created, normalised, leveraged, and converted into power. The cases span domains and decades. The grammar beneath them does not change.
The series takes its name from a simple observation: the most consequential things happening in the world are rarely the things that take centre stage. While we watch the visible hands, something else is being built in the structural layer underneath. These reports are an attempt to make that layer legible.
The founding article, “”, sets out the full grammar. Each investigation that follows applies it to a specific case.
Scene-setter
On 24 February 2022, Russia launched its full-scale invasion of Ukraine. Within days, Ukrainian communications infrastructure was under severe pressure: ground-based networks damaged, military coordination systems strained, command structures disrupted. What followed was not a procurement process. It was an emergency adoption. Elon Musk announced Starlink service activation for Ukraine, and SpaceX began shipping terminals at a pace no state or intergovernmental body could have matched. By mid-2023, more than 42,000 terminals were in use across Ukraine by the military, hospitals, businesses, and humanitarian organisations, making Ukraine the largest single military user of the Starlink network.
The visible story of Starlink in Ukraine is one of technological heroism: a private company stepping in where governments moved too slowly, providing a lifeline to a nation under existential threat. Ukrainian commanders credited the network openly and repeatedly. Western commentators celebrated it as a demonstration of private sector agility. The narrative was clean, compelling, and genuinely partly true.
What this investigation looks at is the structural layer beneath that narrative. The speed of adoption, the depth of integration, and the documented episode in which a private individual declined to extend coverage during an active military operation, curtailing a Ukrainian offensive capability, tell a different story: not of a gift freely given, but of a dependency created under conditions of maximum vulnerability, normalised through military necessity, and already leveraged before most observers had noticed it was there. What the five moves will show is how a sovereign state, in the course of defending its existence, became operationally subordinate to a private actor accountable to no democratic body, bound by no treaty, and guided by his own geopolitical judgment.
Move 1: Creating the Dependency
The entry point here is creation, but creation under conditions that made the dependency feel like rescue. That distinction matters, because it shapes everything that follows.
Ukraine in February 2022 did not lack communications infrastructure in the ordinary sense. It had networks. What it lacked, within days of the invasion, was reliable, decentralised, jam-resistant connectivity that could function at the frontline under active electronic warfare conditions. Russia’s military doctrine includes systematic degradation of communications infrastructure as a first-phase priority. Ukrainian military planners knew this. What they did not have, and could not rapidly build, was an alternative.
Starlink was available. It was low-latency, broadband, and could be deployed terminal by terminal without dependence on ground infrastructure that could be targeted. The hook was not merely utility: it was the only thing that worked at the scale and speed required. Individual Ukrainian commanders adopted it not because they failed to see the long-term implications, but because the short-term operational need was overwhelming and the alternative was communicating less effectively while people died. This is how rational individual choices, made one at a time under conditions of acute stress, produce structural consequences that no individual choice-maker intended or could foresee.
The dependency became costly to reverse almost immediately. Starlink was not layered onto existing Ukrainian military communications doctrine: it was embedded into it. Drone targeting protocols, combined arms coordination, frontline command relay, all were rebuilt around the assumption of Starlink availability. The integration was not gradual in the way that software integrations usually are. It was rapid and total, because the conditions of war do not permit phased adoption strategies. Ukrainian officials declared as early as 2023 that Starlink had become the core of their communications infrastructure. By the time Ukraine had enough operational breathing room to assess its position, the assessment would have revealed that exiting Starlink would require reconstructing the communications architecture of an active military, under fire, with no equivalent alternative in existence.
The dependency creator in this case is SpaceX and, inseparably, Elon Musk. The two cannot be usefully distinguished. SpaceX is a private company with no public shareholders and no board empowered to override its controlling owner. The decisions about Starlink’s coverage, pricing, and operational parameters are Musk’s decisions. What Ukraine adopted was not a utility with distributed governance. It was a privately held infrastructure asset controlled by a single individual.
Move 2: Normalising the Dependency
Normalisation in this case operated with unusual speed, because the conditions of war compressed the timeline that ordinarily makes dependencies invisible. In most cases, a decade or more separates the moment of adoption from the moment of consequence. In Ukraine, the sequence ran in months.
The narrative normalisation was supplied primarily by Musk himself, by SpaceX’s communications, and by the Western press that covered the story. Starlink was framed as a humanitarian technology deployment, a lifeline, a demonstration of what private innovation could do where governmental bureaucracy could not. Musk was cast in some quarters as a wartime benefactor. The framing was not dishonest, exactly: the terminals were real, the connectivity was real, and the operational advantage was real. But the framing successfully kept a structural question out of view: what does it mean to build your military communications doctrine on infrastructure you do not own, cannot replicate, and have no contractual right to demand?
The counter-narrative existed. Ukrainian officials privately expressed concern about the dependency. Analysts in defence and technology policy raised the governance question. But the counter-narrative failed to land publicly for a predictable reason: raising it felt, in the context of active war, like complaining about the specifications of a life raft while drowning. The structural question was correct. The political moment made it unaskable.
Operational normalisation was faster still. Starlink stopped being a communications tool and became the communications infrastructure: the assumption underneath the doctrine rather than a component within it. Ukrainian military planners began designing operations around Starlink availability. Drone units built their targeting workflows on it. This is the point at which a dependency stops being a relationship and starts being simply how things work, and it arrived in Ukraine within the first year of the full-scale invasion. By the time the Crimea episode made the leverage dimension publicly visible, the operational normalisation was already complete. Starlink was not something Ukraine used. It was something Ukraine ran on.
Move 3: Leveraging the Dependency
The leverage in this case became publicly visible in the autumn of 2022, in a manner so direct and so documented that it functions as a kind of structural proof of concept.
Ukrainian forces planned an underwater drone attack on Russian warships anchored near Crimea and required Starlink connectivity extended to the operational area. Starlink service had never been enabled within approximately 100 kilometres of the Crimean coast. When Ukrainian officials, including Deputy Prime Minister Mykhailo Fedorov, contacted Musk directly requesting that coverage be extended to support the operation, Musk refused. His stated reasoning was fear of nuclear escalation, driven in part by conversations with senior Russian officials. He later posted publicly that activating coverage toward Sevastopol would have made SpaceX “explicitly complicit in a major act of war.” The attack on the Russian Black Sea Fleet did not proceed as planned. A sovereign military operation was curtailed not by enemy action, not by equipment failure, not by command decision, but by the refusal of the infrastructure owner to extend service. The episode was confirmed by Musk himself and documented in detail by biographer Walter Isaacson, whose initial account required subsequent clarification on the precise mechanics but not on the core fact: Musk made a unilateral decision that directly affected the operational capacity of a sovereign nation’s military during wartime.
The structural leverage layer operates differently and is less visible. Musk’s geopolitical positions on the conflict, expressed publicly, have been extensive and consistent in their direction. In October 2022, he proposed a peace plan on X that included formal Russian sovereignty over Crimea and Ukrainian neutrality, a proposal the Kremlin described as “very positive” and that drew a blunt rebuke from Ukrainian officials including President Zelensky. In early 2024, he publicly predicted Russia would gain further territory and recommended Ukraine seek a negotiated settlement. In March 2025, following the breakdown of a White House meeting between Zelensky and Trump, he criticised Zelensky directly for refusing ceasefire negotiations, calling the continuation of the war a choice to let “men die in trenches.” These are not incidental statements. They represent a documented, sustained disposition toward positions broadly aligned with Russian negotiating interests, expressed by the individual who controls Ukraine’s most critical military communications infrastructure.
Option-space leverage is the subtlest layer, and in this case it operates at the strategic level of the war itself. Ukrainian military planning cannot be indifferent to Starlink continuity. This means that any operation, any escalation decision, any strategic choice that risks provoking a coverage restriction must be evaluated against that risk. The dependency does not merely affect what Ukraine can do; it affects what Ukraine will consider doing. This is the deepest form of leverage: not the exercise of power, but the prior restructuring of the dependent party’s sense of what is possible.
Move 4: Obscuring the Mechanism
The Starlink dependency in Ukraine is unusual in the dependency grammar because it is not obscured in the way most structural dependencies are. The Crimea episode made the leverage dimension visible almost immediately. The dependency itself has been publicly acknowledged, analytically documented, and discussed in policy circles. This is not a hidden dependency in the conventional sense.
What is obscured, and what this move addresses, is the mechanism that made the dependency unavoidable and the mechanism that prevents its resolution.
The temporal structure of this dependency is compressed but still operates as a displacement device. The decision to adopt Starlink was made in the first days and weeks of the invasion, under conditions in which strategic assessment of long-term structural consequences was impossible. The politicians, military commanders, and institutional actors who made that decision were operating in an acute emergency. The consequences, including the Crimea episode and the ongoing option-space constraints, were not theirs to foresee or prevent. This is temporal displacement in its accelerated form: not decades separating cause from consequence, but weeks, during which the conditions of war made structural analysis a luxury that could not be afforded.
The narrative that obscures the mechanism is not a conspiracy. It is the honest story of genuine utility, framed in a way that made the structural question feel inappropriate. Starlink helped Ukraine survive. That is true. The structural question, what kind of relationship was created in the process of surviving, and who controls its terms, is not made less urgent by the fact that the dependency was genuinely helpful. But the narrative of heroic technology deployment made that question feel churlish to raise, and continues to make it difficult to discuss without appearing to argue against Ukrainian survival.
The structural opacity in this case is primarily jurisdictional. SpaceX is a private US company. Its decisions about Starlink coverage are not subject to Ukrainian law, NATO treaty obligations, EU regulatory frameworks, or any international telecommunications governance mechanism that currently exists. The information about coverage decisions may be technically available to some parties. The institutional architecture to contest those decisions collectively does not exist. Ukraine has no legal avenue for appeal. NATO has no mechanism for override. The gap between the visibility of the leverage and the institutional capacity to address it is the precise measure of the dependency’s depth.
Move 5: Converting the Value
What has been converted here, and who has paid?
The most immediate conversion is strategic. Musk demonstrated, in the Crimea episode, that a private actor controlling critical military communications infrastructure can exercise sovereign-equivalent influence over a wartime operation without any of the accountability structures that normally constrain sovereign actors. That demonstration has value beyond the specific episode. It establishes a precedent: private infrastructure can be withheld from a client state to shape the outcome of a military operation, and there is no existing framework to prevent or contest it. The value converted is not financial. It is the normative establishment of a new category of power: private actor with sovereign reach and private accountability.
The commercial conversion is ongoing. SpaceX has used Ukraine as the world’s largest operational demonstration of Starlink’s military utility, under the most demanding real-world conditions any satellite communications system has faced. The result is a product validation no controlled test could produce, a fact not lost on other governments evaluating military satellite communications procurement. The dependency Ukraine cannot exit has simultaneously become the most persuasive sales case in Starlink’s commercial history.
The geopolitical conversion concerns the United States. The Pentagon contracts that partially formalised the dependency, announced in June 2023, routed financial accountability through Washington but did not resolve the sovereignty question for Ukraine. The Pentagon declined to disclose the contract’s value, scope, or timeline. What those contracts did establish is that US government funding flows to a private company controlled by a single individual, whose geopolitical judgments may or may not align with US foreign policy at any given moment. This creates a second-order dependency: Ukraine’s access to Starlink is now contingent not only on Musk’s decisions but on the continuity of US government willingness to fund and maintain the arrangement. In February 2025, Reuters reported that the US had threatened to cut off Starlink service if Ukraine did not sign a deal on critical natural resources, a claim Musk denied but that illustrated precisely how entangled the two dependencies had become in the perception of all parties.
Who pays? Ukraine pays in option space: the strategic choices foreclosed by the knowledge that operational capacity depends on the continuing goodwill of a private actor. Ukrainian soldiers pay when coverage decisions affect operational outcomes. Ukrainian sovereignty pays in the simple fact that a private individual has exercised, and may again exercise, decision-making authority over its military operations. Those costs are borne not by the commanders who adopted Starlink in February 2022, because they had no real alternative, but by the state and its people who inherited the structural consequence of that necessary choice.
Analytical Notes – The Collision Dynamics
This case presents the collision dynamic explicitly, and it is more structurally complex than a single dependency would be.
Ukraine depends on Starlink for operational military communications. That dependency is held by a private actor with no public accountability obligations, governed by no treaty or regulatory framework, and already documented as willing to exercise unilateral coverage decisions based on personal geopolitical judgment.
Ukraine simultaneously depends on Western government support, primarily US military and financial aid, for its capacity to sustain the war. That dependency is held by democratic governments subject to electoral cycles, political leadership changes, and public opinion dynamics that do not necessarily align with Ukrainian interests at any given moment.
The collision point is the Pentagon contract layer. The formalisation of Starlink funding through US government contracts appeared to resolve the dependency by routing it through an accountable institution. What it actually did was entangle the two dependencies. Musk’s commercial relationship with the US administration now intersects with his control of Ukrainian military communications infrastructure. If US political conditions shift against Ukraine, both dependencies are affected simultaneously, through different mechanisms but with compounding effect. The Reuters report of February 2025, whether or not the threat was precisely as described, captures the logic of the collision: the leverage exercisable through Starlink and the leverage exercisable through aid decisions had become, in the perception of all parties, a single instrument. The partial formalisation did not reduce Ukraine’s vulnerability. It added a second layer of contingency to an existing one.
The dependent party caught between these two dependency architectures has no institutional mechanism for navigating the collision. Ukraine is actively pursuing alternatives: the EU’s GovSatCom system reached initial operational capability in January 2026 and has been identified as an interim measure, while the IRIS2 constellation, a 290-satellite multi-orbit network contracted to the SpaceRise consortium in October 2024, is expected to reach full operational capability between 2029 and 2030. Ukraine is in discussions to join IRIS2 as a non-EU participant. These are meaningful developments. They are also a decade behind the dependency they are designed to address, and none currently operates at a scale or latency comparable to Starlink.
Closing
What this case reveals about the grammar is something the founding article anticipates but cannot demonstrate from within its own frame: the grammar operates across compressed timelines when conditions of acute stress replace the gradual adoption dynamics of peacetime. Most dependencies take years or decades to reach the normalisation phase. The Starlink dependency in Ukraine reached it in months, because war does not permit the deliberative processes that might otherwise slow adoption or surface structural questions before the switching cost becomes prohibitive. The grammar is not merely patient. It is opportunistic. It installs itself most completely precisely when the dependent party is least positioned to assess what is being installed.
The case also reveals something about the relationship between visibility and accountability. The Crimea episode made the leverage dimension unusually visible. The dependency is publicly documented. The structural problem is analytically understood, and senior EU officials have named it in precisely those terms: “Elon Musk is in fact the guardian of Ukraine’s connectivity on the battlefield. And that’s a strategic vulnerability.” None of this has produced a remedy at the speed the situation requires. The gap between legibility and institutional response is itself a finding: the architecture to act on what we can now see does not yet exist at the necessary scale or speed. International telecommunications governance does not reach private satellite constellations used for military operations. No NATO framework governs the coverage decisions of a private contractor. No treaty exists that would require SpaceX to provide continuous service to a sovereign military customer. The dependency is visible and structurally unaddressed, which means the grammar has produced a form of its usual obscurity that is more durable precisely because it persists not through concealment but through the absence of tools to respond.
The question to carry forward is not whether Ukraine should have adopted Starlink. It had no serious alternative. The question is what it means for international order that the most consequential military communications infrastructure of the largest European land war since 1945 is controlled by a private individual, and that this fact is broadly known, widely discussed, and structurally unaddressed. The grammar does not require secrecy. It only requires that the moment for action arrives after the moment for action has passed.
Comments are closed