September 17, 2022
Elon Musk personally ordered Starlink switched off over Crimea.
Not a government. Not a court. Not a UN resolution. One person determined the outcome of a military conflict. The Ukrainian military had planned a drone strike on the Russian Black Sea Fleet using Starlink as the navigation backbone. Musk decided it shouldn't happen. He withdrew the satellite coverage. The drones went dark.
This was not a failure of Starlink. It was Starlink working exactly as designed — a single private infrastructure that a single private individual controls.
Starlink & Digital Sovereignty — CME Ethics Tool · 2026
Starlink works. That is what makes it hard. Communities across sub-Saharan Africa, Southeast Asia, and Latin America that had never had reliable internet access became connected in weeks. Rural hospitals reached specialists in capitals. Schools accessed curricula they could never afford in print. Displaced families in conflict zones reconnected with relatives.
By 2025, Starlink operates in over 100 countries, with more than 7,000 satellites — the largest constellation in human history. For an estimated 600 million people in sub-Saharan Africa alone who still lack reliable connectivity, satellite internet is not a luxury. It is access to healthcare, education, economic participation, and safety.
For many communities, Starlink is not a choice between providers. It is the only option that exists. This matters for how you think about the dilemma. This is not a free market. The absence of alternatives was not an accident.
The technology works. The terms are the problem.
When a country subscribes to Starlink, its national data passes through infrastructure owned by a private American company, stored under US law, accessible to US federal authorities under the CLOUD Act without local court approval. Service can be suspended at SpaceX's "operational discretion" with thirty days notice. There is no arbitration clause. There is no exit mechanism that a small nation can realistically exercise. When rates increase — and they have — governments pay or go dark.
"Infrastructure is politics." — Laura DeNardis, The Global War for Internet Governance (2014)
These terms are not unusual for a private tech company. They are standard practice. That is the point. The infrastructure connecting the Global South to the internet was not built under terms negotiated by the people it serves. It was built under terms set by those who built it.
Colonialism was not only about conquest. It was about who built the infrastructure, who it served, and where the resources flowed.
The railways that crossed colonial Africa did not connect African communities to each other. They connected mines to ports — extraction infrastructure dressed as development. The human benefit to local populations was real and entirely incidental. Kwame Nkrumah identified this pattern in 1965: formal independence had arrived, but economic and infrastructural dependency kept the colonial relationship intact under a different name. Neo-colonialism — control without conquest.
Satellite internet follows a different trajectory. The human benefit is genuine and intended, not incidental. But the structural questions are identical: Who owns the infrastructure? Under whose law does it operate? Who can revoke it? When the infrastructure that connects a community is owned entirely by external actors, that community's ability to govern its own data, communications, and connectivity — its digital sovereignty — is compromised regardless of intent.
Couldry and Mejias (2019) call this "data colonialism": the same logic of appropriation and extraction as historical colonialism, now applied to human data and attention. The colony pays. The metropole accumulates. The relationship is maintained not by force but by dependency — and dependency, once deep enough, looks indistinguishable from choice.
Kwanda is a fictional country. The dilemma it faces is not.
The tool places you in the role of a decision-maker who has no good options. Adopt Starlink and your people are connected; your sovereignty is mortgaged to a private American corporation. Reject it and your sovereignty is intact; your people remain excluded from the global economy, the healthcare network, and the information commons that the rest of the world takes for granted.
Neither path is clean. The experience is designed to make that impossible to ignore. There is no correct ending. Every path ends with a cost that the people who made the decision will not personally pay. That asymmetry — who decides, who bears the consequence — is what post-colonial ethics is designed to examine.
The goal is not to tell you what to think about Starlink. The goal is to make the weight of the question feel impossible to ignore.
The tool is a single-page web simulation built in vanilla HTML, CSS, and JavaScript with no framework or build step. It runs entirely in the browser from three files — index.html, style.css, and script.js — and requires no installation or server. An interactive map is provided by Leaflet.js.
The design is structured in five sequential phases, each serving a specific function in the ethical argument.
Before any choice is presented, the user passes through six scenes that establish why no good option exists. These are not background reading — they are the argument. Scene 1 introduces Amara Diallo, a nurse at a rural clinic whose patient died for lack of internet access. Scenes 2–4 show the three alternatives that have already failed: a World Bank rejection, a regional provider whose prices are triple the average monthly income, and an African satellite consortium nine years from completion. Scene 5 summarises the stakes (312 clinics offline, 847 schools without connectivity, 2.4 million people excluded). Scene 6 presents the SpaceX offer.
The purpose of this sequence is to make the choice structurally coerced before the user encounters it. Testing showed that users who arrived directly at the choice without this context treated it as a straightforward policy decision. The onboarding is designed to close that exit.
The main experience opens on an interactive map showing six communities across Africa, Southeast Asia, and Latin America as grey dots. The user must visit at least three communities before the choice buttons activate — this ensures they encounter the human cost of exclusion before deciding.
Choosing to adopt Starlink triggers a contract overlay — a realistic legal document with four clauses drawn from standard technology service agreements. The clauses cover data jurisdiction (US law, CLOUD Act), service termination rights (SpaceX discretion, no arbitration), pricing modification rights, and data sovereignty requests (no binding commitment). The user must sign — they cannot skip the contract.
After signing, a data-flow primer overlay shows the extraction architecture before the joy phase begins: Kwanda → Starlink → US Servers. This is the only moment in the tool where the structure is made explicitly visible rather than experienced through consequence.
The joy phase simulates three years of connectivity in 20 real seconds. A day counter runs from Day 0 to Day 1,095 using a requestAnimationFrame loop. As time passes: map markers shift from grey to glowing green, community stat labels fade in one at a time (staggered to avoid information overload), toast notifications appear sequentially — one at a time, each fully visible before the next — reporting real connectivity benefits (school pass rates, clinic stockouts, language archives). A phone notification from the user's daughter reads: "Papa, we have internet at school!!" A second message, from the US Ambassador, congratulates Kwanda on its "forward-thinking leadership."
At Day 1,095, rather than auto-advancing, the tool shows a prompt: "Something has shifted. The data is flowing — but where?" The user clicks to continue. This moment of user-controlled pacing was introduced after testing showed that auto-advancing felt disorienting. The pause gives the user a beat of reflection before the Crimea flash.
The Crimea flash is a full-screen overlay: "September 2022 — Elon Musk personally orders Starlink switched off over Crimea. A private individual just determined the outcome of a military conflict. No vote. No court. No appeal." The screen is clickable to dismiss. After dismissal, the map redraws: markers shift to orange, data flow lines appear between communities and a server node, a money counter begins ticking upward ($2.4M per month leaving Kwanda), and anger-phase toasts report surveillance, pricing increases, and subpoenaed patient records. The same connectivity that caused joy is revealed as the mechanism of extraction.
The reject path runs in parallel: markers turn dark, a days-without-access counter begins climbing, and three years of consequences unfold through grief-phase toasts — clinic deaths, school closures, missed development opportunities. Neither path is designed to feel comfortable.
After the consequence phase, the user faces a second decision. On the adopt path: negotiate better terms with SpaceX (partial concessions, continued dependency), accept the dependency and build domestic capacity in parallel, or fund the African Union satellite consortium. On the reject path: capitulate and adopt Starlink anyway, join a regional fibre consortium, or accept Chinese infrastructure on equivalent terms. Each of the six options represents a real-world policy response that governments in the Global South have actually taken or considered.
Each of the six paths ends on a dedicated screen. Every ending names what actually happened — what the real-world equivalent of this choice produced — and identifies the post-colonial theorist whose framework explains why it was structurally inevitable. Endings are not graded as good or bad. Each ends with a question the user is left to sit with. An "About & Sources" link leads to this page.
No correct answer. The tool does not signal which choice is right. Both paths produce harm. The impact bars (Human Welfare, Economic Growth, Sovereignty, Infrastructure) shift in both directions on every path. The endings name failure modes, not victories.
Information sequencing over information density. Early versions fired all stimuli simultaneously — map labels, toasts, sidebar statistics, and top-panel text all appeared at once. Testing showed this produced confusion, not reflection. The current version staggers every information layer: the sidebar delays 3 seconds after the choice, map labels appear one community at a time 900ms apart, and toasts queue so only one is visible at a time. The argument is the same; the pacing allows the user to actually receive it.
Specificity over abstraction. Every statistic in the tool is drawn from real data — drug stockout rates in East African clinics, UNESCO literacy statistics, actual Starlink contract terms, the 2022 Crimea incident. Kwanda is fictional; its conditions are not. The decision to use a fictional nation was deliberate: it removes the cognitive escape hatch of "but I know how this story ends" that real country names provide.
Click any entry to read the annotation and see how it connects to the tool's design.
The primary theoretical framework for this tool. Couldry and Mejias argue that "data colonialism" is a structural continuation of historical colonialism — the same logic of appropriation and extraction, now applied to human data and digital connectivity. Their concept of "data relations" as a mode of dispossession directly informs the dependency phase: the moment the viewer realises that connectivity and extraction are inseparable in the Starlink model.
Nkrumah argued that formal decolonisation created a new mode of colonial control — dependency maintained through economic and infrastructural means rather than direct rule. His observation that neo-colonial infrastructure presents itself as development while serving the interests of the controlling power maps precisely onto Starlink's expansion model. The colonial railway metaphor in this tool draws directly from his analysis of infrastructure as extraction.
Fanon's analysis of colonial dependency — particularly how dependency structures the agency and self-determination of colonised peoples — provides the ethical ground for why substantive sovereignty matters beyond formal independence. His critique of post-colonial elites who reproduce colonial structures informs the tool's refusal to frame rejection as a clean victory: rejecting Starlink in the name of sovereignty is also a choice made by elites who already have connectivity.
DeNardis demonstrates that control over internet infrastructure constitutes political governance power. Her central argument — "infrastructure is politics" — grounds the claim that Starlink's private ownership is not merely a commercial question but a question of who governs global connectivity. Her framework explains why a developing nation's dependency on SpaceX represents a governance deficit, not simply a market dependency.
Kwet argues that US technology corporations are enacting a new form of imperialism in the Global South through control over software, hardware, and network infrastructure. His analysis of how technical standards, licensing regimes, and platform dependencies reproduce colonial economic relationships maps directly onto SpaceX's position in the satellite internet market.
This volume examines how communication infrastructures — cables, satellites, towers — embed social and political power relations that are typically rendered invisible to their users. Several chapters address satellite infrastructure specifically, arguing that the apparent universality of satellite coverage conceals deeply situated political and economic interests. This provides the bridge between post-colonial theory and the material specificity of Starlink as a physical infrastructure.