12 Million People Can't Turn It Off: How Starlink Became Elon Musk's Geopolitical Weapon

12 Million People Can't Turn It Off: How Starlink Became Elon Musk's Geopolitical Weapon

Millions of People and Dozens of Countries Rely on Starlink, Giving Elon Musk Unprecedented Leverage

There's a scene in the Ukraine war that gets cited a lot in policy circles. It's not a battle. It's not a treaty. It's a decision made by one man, in a room somewhere, about whether to keep the internet on.

You don't need to be a genius to understand the power dynamics of that. You just need to read the news. A commercial satellite internet service, designed for rural broadband and maritime connectivity, became the nervous system of a modern military. Then the guy who owns it, one guy, no shareholders to answer to, no board to consult, decided to restrict it. Temporarily. For a specific operation. Then he turned it back on.

That was 2025. By 2026, the world had 12 million Starlink subscribers in more than 160 countries. The Pentagon called the service "indispensable." The Iranian government had 6,000 terminals smuggled into its territory during civil unrest. Ukraine's drone operations ran on it. The United States Navy had naval drones that drifted for an hour off the California coast when Starlink hiccupped.

This is the situation. Not hypothetical. Not speculative. Twelve million customers. One owner. And a lever that a lot of powerful people wish they had never handed over.

The Number That Matters

Twelve million. That's the figure Starlink announced in mid-2026. It's a big number, but the growth curve tells a more interesting story. In 2023, the service had 2.3 million subscribers. By 2024, it doubled to 4.4 million. By 2025, it had nearly doubled again, 8.9 million. The acceleration isn't linear. It's gravitational. Once a critical mass of users, enterprises, and governments adopt a service, switching becomes harder than staying.

The geographic spread matters as much as the raw count. Starlink operates in 160 countries and territories. In Africa alone, it serves 26 markets, with Nigeria, Zimbabwe, and Zambia accounting for roughly 64% of the continent's subscribers. The top five markets by Speedtest sample counts are the United States, Mexico, Indonesia, Brazil, and Canada. That's not a niche service for remote cabins anymore. That's a global utility.

And the market share is staggering. Starlink accounted for 97.1% of all global satellite internet Speedtest samples in the third quarter of 2025. Viasat, the distant second place, had 1.7%. HughesNet had 1.0%. This isn't competition. It's a monopoly with decorative rivals.

The latency numbers explain why. Starlink's highest recorded latency, in the Marshall Islands, of all places, was 282 milliseconds. That's less than half the best latency any geostationary satellite provider could manage. For someone in a rural clinic or a military command post, the difference between 26 milliseconds and 600 milliseconds is the difference between a video call and a frozen screen.

Once you build the infrastructure, you own the pipe. And the pipe is getting harder to replace.

How Starlink Became the World's Backup Plan

Starlink didn't become critical infrastructure by accident. It became critical because terrestrial infrastructure kept failing, and Starlink kept working.

The pitch was simple. Ground-based internet requires cables, towers, fiber, permits, and functioning governments. Satellite internet requires a dish and a clear view of the sky. When a hurricane takes out a cell tower, Starlink keeps broadcasting. When a government shuts down the internet during protests, Starlink connects through the blackout. When a war destroys a country's telecom grid, Starlink fills the void.

The design is elegant in its simplicity. Thousands of small satellites orbit at roughly 550 kilometers, low enough for low latency, high enough to cover enormous swaths of the planet. SpaceX launches them by the dozen, using its own reusable rockets. As of early June 2026, there were over 10,400 Starlink satellites in orbit. That's not a constellation. That's a blanket.

The business model follows the same logic. Starlink charges $55 a month for a 100 Mbps residential plan, $85 for 200 Mbps, and $130 for the top-tier Residential Max. The hardware costs $349 up front, or can be rented for $10 a month. For a rural household with no fiber option, that's not expensive. It's the only option.

For governments and militaries, the calculus is different. The price isn't the monthly fee. The price is dependence. Once you integrate Starlink into your communications architecture, you've outsourced a piece of your sovereignty. And sovereignty is not something you get back easily.

Ukraine, The First Test Case

Ukraine is where the world learned what Starlink leverage looks like.

When Russia invaded in 2022, Ukraine's terrestrial communications infrastructure took a beating. Starlink filled the gap. Soldiers used terminals to coordinate artillery, guide drones, and share real-time intelligence. The service became the backbone of Ukraine's tactical communications. Not a convenience. A necessity.

Then Musk started making decisions. In 2025, reports emerged that he had ordered a temporary shutdown of service in a coastal area during a Ukrainian operation against the Russian fleet. Ukrainian officials said the outage directly hampered their offensive. Musk's defenders said he was trying to prevent escalation. The details are contested. The power dynamic is not.

The situation got messier. Russian forces started using Starlink terminals on their own drones. Ukraine appealed to Musk to restrict their access. He did. Russian units faced communication disruptions, and their attempts to switch to alternatives proved less effective. So Musk, a private citizen, not an elected official, not a general, was effectively making battlefield access decisions for both sides of a war.

That's the leverage. Not the kind you vote for. Not the kind you can subpoena. The kind that comes from owning the switch.

The Pentagon's Uncomfortable Marriage

The U.S. military has a Starlink problem. It needs the service, and it doesn't trust the service. Both things are true at the same time.

In the summer of 2025, a global Starlink outage hit during naval drone tests off the coast of California. Two dozen U.S. Navy drones lost control and drifted for an hour. The Pentagon had been developing those drones specifically to counter China. The outage was a technical failure, not a deliberate action. But it exposed a structural vulnerability: if Starlink goes down, American military capabilities go down with it.

Reuters reported that the Pentagon is "critically dependent" on Starlink satellite communications, and that this dependency poses significant risks in a potential conflict with China. The report cited two specific examples: SpaceX's refusal to provide connectivity over occupied Ukrainian territories, and its likely refusal to serve U.S. military personnel in Taiwan to avoid triggering Chinese sanctions.

SpaceX denied the Taiwan claim. The denial doesn't change the underlying reality. The company has made decisions about where its service is available and where it isn't. Those decisions are made by SpaceX. Not by the Pentagon. Not by Congress. Not by any democratic process.

The workaround is Starshield, a separate SpaceX product designed for government and military use, with encrypted communications and battlefield awareness capabilities. But Starshield runs on the same infrastructure, launched by the same rockets, managed by the same company. The dependence doesn't disappear. It gets a new logo.

Bryan Clark, an autonomous warfare expert at the Hudson Institute, put it plainly: "You accept those vulnerabilities because of the benefits you get from the ubiquity it provides". That's the deal. Everyone knows the terms. Nobody has a better option.

Iran, DRC, and the Quiet Interventions

Ukraine gets the headlines. The quieter interventions are where Starlink's geopolitical role becomes stranger.

In January 2026, during a period of civil unrest in Iran, the U.S. government secretly shipped approximately 6,000 Starlink terminals into the country to help protesters maintain communications. The Iranian government had been shutting down internet access to suppress the unrest. Starlink terminals bypassed the blackout. The operation wasn't public. It was a covert intervention conducted through a commercial service owned by a single individual.

The Democratic Republic of Congo tells a different story. Starlink was banned in the country in March 2024 over military and security concerns. The ban was lifted in May 2025. During the ban, armed groups in eastern Congo were reportedly using Starlink terminals to coordinate operations. The service that was supposed to connect the unconnected was connecting militias in a conflict zone.

The pattern is consistent. Starlink shows up where terrestrial infrastructure fails. It provides connectivity where governments can't or won't. And in doing so, it becomes a variable in conflicts and crises that have nothing to do with internet access. The technology is neutral. The consequences are not.

The Countries Saying No

For every country that embraces Starlink, there's another that blocks it. Those rejections are the most revealing data points.

China banned Starlink outright. Beijing views the service as a national security threat and has never issued SpaceX an operating license. Users of illegal terminals face penalties. The reasoning is straightforward: a foreign satellite network that bypasses state control over information is incompatible with China's internet governance model.

India is more complicated. Starlink has agreements with Indian telecom companies Airtel and Jio. But final regulatory approval remains frozen. Security agencies have raised concerns about data routing and localization. They've studied Starlink's track record in Ukraine and the Middle East, where SpaceX remotely geofenced coverage and altered routing during conflicts. India's question is simple: if SpaceX can flip the switch on a battlefield, why should India trust it with civilian infrastructure?

Namibia rejected Starlink's license application in March 2026. The Communications Regulatory Authority of Namibia cited ownership rules, Starlink is wholly foreign-owned, and concerns about data sovereignty, national security, and the government's ability to supervise the service. A local expert put it bluntly: Starlink's encrypted messaging systems allow communications to "bypass Namibia's cybersecurity system".

South Africa has similar ownership requirements that have blocked Starlink's entry. So has Angola. In Europe, Italy's government hit pause on a €1.5 billion contract with SpaceX after domestic opposition questioned whether handing national security communications to a company owned by a man with volatile political views was wise. Canada's Ontario province tore up its Starlink contract in protest of U.S. tariffs.

These rejections aren't about technology. Starlink works. They're about control. The countries saying no are asking a question the adopters haven't answered: what happens when the person who owns the switch decides he doesn't like you?

The Competition Question

There are competitors. On paper, at least.

Amazon's Project Kuiper, now rebranded as Amazon Leo, has deployed 375 satellites after 14 orbital launches. OneWeb, owned by Eutelsat, has 648 satellites. Starlink has over 10,400. The gap isn't a gap. It's a canyon.

Amazon Leo targets up to 400 Mbps download speeds, slightly higher than Starlink's typical range. But Kuiper's deployment is years behind, and its service availability is limited. OneWeb focuses on enterprise and government connectivity, with higher latency and lower density than Starlink. Neither is a consumer replacement for Starlink at scale.

The deeper problem is switching costs. A household that buys a Starlink dish, mounts it on the roof, and integrates it into their daily life isn't going to switch to a competitor that offers similar speeds for a similar price. The friction of change is too high. And for governments and militaries that have built Starlink into their operational architecture, the cost of switching is measured in years of planning and billions in reconfiguration.

Starlink's first-mover advantage isn't just about satellites. It's about habit. It's about the fact that when someone needs internet in a crisis, Starlink is the first name they think of. That mental availability is as valuable as the orbital infrastructure.

What Leverage Actually Looks Like

Leverage isn't always visible. Sometimes it's a tariff negotiation. Sometimes it's a regulatory approval that gets fast-tracked. Sometimes it's a diplomatic cable that mentions a company's interests in a trade deal.

In 2025, congressional oversight Democrats demanded an investigation into whether the Trump administration was using tariff negotiations to support Starlink's expansion. A State Department cable specifically mentioned Starlink in the context of Lesotho's trade negotiations with the United States. India's concessions on Starlink were described as "especially noteworthy".

The pattern is hard to prove and harder to ignore. When a company's owner has direct access to the president, and that company is seeking regulatory approval in dozens of countries, the line between business and foreign policy blurs. Musk's political engagement, his public feuds with European leaders, his commentary on Ukraine, his relationship with Trump, has become a liability for Starlink's expansion in some markets and an asset in others.

Deutsche Bank analyst Edison Yu noted that geopolitical factors are causing a "structural shift" in satellite communications. "Starlink is clearly still number one," he said, "but not the only one". That's the tension. Starlink's dominance is real. So is the resistance to it.

The Fragility of Dependence

Dependence cuts both ways. The more countries rely on Starlink, the more leverage Musk has. But the more leverage he exercises, the more incentive those countries have to find alternatives.

The Ukraine case is instructive. After Musk restricted service during the coastal operation, Ukrainian officials didn't just complain. They started looking for backups. European politicians who had been funding Ukraine's Starlink service began exploring alternatives. The trust was damaged. The dependence remained, but the assumption of reliability didn't.

The Pentagon faces the same dilemma. It needs Starlink now. It's investing in alternatives, Starshield, other commercial providers, government-owned satellite networks. But those alternatives take years to build and billions to deploy. In the meantime, the dependence persists.

The fragility isn't just technical. It's political. A single decision by a single person can change the availability of a service that millions of people rely on. That's not a bug in the system. It's the system.

Starlink's trajectory 

It points in one direction. More satellites. More subscribers. More countries. The company plans to expand its constellation to 12,000 satellites and beyond. It's targeting 1 billion potential new users in markets like India, Turkey, Morocco, and Bangladesh. The growth curve hasn't flattened. If anything, it's steepening.

The leverage will grow with it. Every new country that adopts Starlink adds another node of dependence. Every military that integrates the service adds another layer of vulnerability. Every government that approves a license accepts a trade-off between connectivity and control.

Musk has shown that he will use the leverage. He restricted service in Ukraine. He shipped terminals to Iran. He makes decisions about where Starlink operates based on his own judgment, political, strategic, personal. The question isn't whether he has leverage. The question is what happens when he uses it in a way that a government can't tolerate.

The answer, for now, is nothing. The dependence is too deep. The alternatives are too slow. The switch stays in his hand. And the world keeps connecting.

Twelve million people. One hundred sixty countries. One owner. The math is simple, and the implications aren't.

Starlink solved a real problem. It connected places that cable and fiber couldn't reach. It kept Ukraine online during an invasion. It gave protesters in Iran a way around a government blackout. Those are not small things.

But the same infrastructure that enables those outcomes creates a concentration of power that no private citizen has held before. Musk can decide which battlefields have internet and which don't. He can decide which countries get service and which get blocked. He can decide, in a crisis, who stays connected and who goes dark.

That's not a conspiracy theory. It's a description of the system as it exists. The system was built by a company. The company is owned by a man. The man makes the calls.

The world relies on Starlink. And Starlink relies on Elon Musk. That's the leverage. It's not going anywhere.