The sovereignty of Starlink, as of October 2026, sits inside one operational fact: SpaceX controls approximately 75 percent of all active maneuverable satellites in Low Earth Orbit, serves 13.6 million subscribers across sovereign territories, and holds the switch that determines what those subscribers can and cannot do.
One Hundred and Sixty Kilometres per Hour
The British imperial cable network, at its peak in 1902, ran through roughly forty chokepoints that London either owned outright or leased from a compliant sovereign. The comparison to Starlink is tempting and wrong in one exact way: the Victorians never claimed the cables were someone else's property while retaining the hand on the key.
In mid-October 2026, SpaceX imposed a 160 km/h velocity limit on Starshield terminals in Ukraine. Drones travelling faster than that threshold lost their satellite uplink. A speed cap — not a weapon, not a sanction, not a declared policy — quietly reached into an active theatre of war and changed what a country could do with hardware it believed it controlled.
This is the thesis detail, and it is sufficient. A country may own the terminal, but someone else holds the switch. That is not a new anxiety — small nations have lived inside it since the first undersea cable landed on their shore.
What is new is the scale: 11,150 satellites under private orbital control, seventy-five percent of all active maneuverable objects in Earth orbit, one corporate signature on the service agreement. The question from the small-nation floor is not technical. It is the oldest question in the map room: who is the borderland here, and who holds the chokepoint?
Eleven Thousand Satellites, One Balance Sheet
As of October 2026, SpaceX operates approximately 11,150 satellites in Low Earth Orbit. That figure is not a boast from a corporate press release. It is a structural condition: SpaceX now controls roughly 75 percent of all active maneuverable objects in Earth orbit.
No state, no alliance, no multilateral body comes close.
The financial weight behind that number is similarly unambiguous. Starlink reported 2025 revenue of 11.4 billion dollars and operating income of 4.4 billion dollars. A company generating that margin does not negotiate with regulators from a posture of supplication.
It negotiates as a counterparty with leverage, because the regulator's citizens are already subscribers, and the regulator knows it.
The human footprint of that dependency is 13.6 million consumer subscribers worldwide as of 30 September 2026. Nigeria alone recorded 98,642 active subscriptions by the second quarter of the same year, a number spread across commercial users, healthcare facilities, and rural communities that have no credible terrestrial alternative. Those are not customers who can absorb a service disruption and switch providers by Tuesday.
The scale argument is the structural argument. A network carrying 13 million subscribers across sovereign territories is not an amenity in the sense that a hotel chain is an amenity. It is infrastructure, and infrastructure has always been political, from the railway concessions that carved up Qing-dynasty China to the IMF conditionality packages that arrived in manila folders and rearranged governments.
The question of who controls the switch is not a regulatory technicality. It is the question. Ask the small question first: who is the borderland here, and who is the empire?
The Switch and the Silence Around It
Two statements sit side by side in the public record, and they cannot both be true. Elon Musk has said that Starshield — the government-grade tier of his network — is owned and controlled by the United States government. U.S. officials, asked about the October 2026 speed cap, denied ordering it.
The contractual void between those two positions is where Ukraine's drone operators discovered the limit of their battlefield connectivity.
Marco Rubio supplied the cleaner formulation. Operational decisions regarding Starlink rest with Musk and SpaceX. Read that sentence against the official denial and the load-bearing wall becomes visible: Washington may hold the deed on paper, but SpaceX holds the switch in practice.
The moulding — the language of government ownership, oversight, accountability — is decorative. The structure underneath it belongs to one private company.
A country may own the terminal, but someone else holds the switch.
The mechanism was already visible before October. In February 2026, SpaceX introduced a whitelist policy designed to prevent Russian forces from accessing terminals. The policy also disrupted Ukrainian military units.
That outcome was presumably unintended. What it demonstrated, unambiguously, was that a single corporate chokepoint can misfire against the party it is meant to protect — with no obvious recourse, no appeal, no override from the government nominally in charge.
The unknown that matters most is legal, not technical. No public instrument has been produced showing Washington's right to override SpaceX on Starshield operations — no treaty clause, no contract provision, no statute. If such an instrument exists and was not invoked in October, the question is why.
If it does not exist, then the language of government ownership describes an aspiration rather than a fact. The difference between those two readings is not a fine point of procurement law. It is the difference between a sovereign's tool and a subscription.
Twenty-Nine Points, and Other Assertions of Starlink Sovereignty
India's Ministry of Communications declined to issue Starlink an operating license in October 2026. The reason given was a 29-point security compliance framework — data localisation, lawful interception capability, real-time government access to subscriber locations. Musk's public response was to ask, in paraphrase, whether Mukesh Ambani was the real boss of India.
The question was rhetorical. It was also an answer.
Ambani's Reliance Jio holds the terrestrial infrastructure that Starlink would displace. The Indian state's interest and its national corporate champion's interest happened to coincide, and the 29 points were the instrument where they met. Whether that confluence was coordination or accident is a question the compliance framework is not designed to answer.
Brazil moved differently. In 2024, Supreme Court Justice Alexandre de Moraes suspended the X platform and threatened Starlink's operations directly, using content moderation orders as the lever. The logic was straightforward: if Musk would not comply through one service, the court would reach the other.
A judge, a platform ban, a satellite network — the chain of escalation is worth reading slowly.
Uganda and Vietnam chose a quieter instrument. Terminal import restrictions. Neither country could mount a full licensing standoff; both could control what crossed the border.
The result was the same — Starlink operating inside a regulatory perimeter it had not chosen — achieved at a fraction of the diplomatic cost.
The pattern across all four is legible. States reach for whatever leverage their market size and institutional capacity will support. India can demand 29 points.
Uganda restricts the hardware. The outcomes are proportional, and nobody mistakes that for fairness. The borderless network turns out to have borders wherever a government decides to draw one.
The Treaty That Arrived Too Early
The 1967 Outer Space Treaty was written for state actors with launch programs, not for a private company orbiting 11,150 objects. Article VI holds the United States internationally responsible for SpaceX's activities in space — a clause drafted when a satellite was a government program and its operator was a civil servant.
It remains the one legal hook on the entire structure. Everything else is moulding.
The ITU has spectrum authority and no enforcement mechanism over a going concern already in orbit. Rwanda and France grasped this quickly. Rwanda filed for 337,320 satellites and France for 116,640 through ITU processes — neither country possessing the industrial base to launch a fraction of that number, but both acquiring a legal claim to orbital slots and spectrum bands that Starlink would otherwise consolidate.
The treaty mechanism, repurposed as a chess piece. The comparison to the 1970s Law of the Sea negotiations is tempting — small states filing claims against large ones — and it breaks in exactly the same place: the filing is not the fleet.
The 1967 treaty comparison is similarly lazy in one direction and precise in another. It was not designed for private infrastructure at scale, which is accurate. But the state-responsibility clause was written broadly enough that it survives the era it did not anticipate.
The load-bearing wall is Article VI: Washington can be held to account for what Musk does in orbit, whether Washington ordered it or not. That is the structural fact. The ITU filings and the spectrum disputes are real, but they are the decorative layer above it.
What the Signal Carries
The constellation is not as closed as its operator assumes. Academic researchers have already recovered control-field and quasi-cyclic code representations from Starlink downlink signals, documented in arXiv:2610.29387v1, which suggests that the architecture SpaceX treats as proprietary is more legible to outside analysis than the company's posture implies.
Seventy-five percent of all active maneuverable satellites in Earth orbit belong to one firm. That concentration of technical opacity and orbital mass in a single private hand is the structural problem with orbital sovereignty at scale, not a symptom of it.
The next layer of dependency is being built before the first is resolved. The SatCraft framework, published in arXiv:2610.29588v1, deploys large language models to optimise joint scheduling across satellite-cellular edge services — a further integration of Starlink infrastructure into civilian and commercial telecommunications that will deepen the lock-in each year it operates unregulated.
Watch India's Ministry of Communications. If New Delhi converts its 29-point compliance framework into a binding license instrument before the end of the year, it will be the first mid-sized democracy to resolve the Starlink sovereignty question on its own terms. If the 29 points remain a permanent stall, the record will show something simpler: that the switch exists, that one man holds it, and that the rest of us are subscribers.
