When Starlink activated service in Zambia in late 2023, a clinic in Serenje District got reliable internet for the first time. Patient records that once traveled by motorbike could move instantly. That single connection is a small data point, but multiply it across dozens of countries and the pattern becomes harder to dismiss: SpaceX’s low-Earth-orbit satellite network is reaching places that traditional providers never prioritized.
By mid-2025, SpaceX founder Elon Musk stated on X that Starlink had crossed 5.5 million active subscribers in 118 countries. The company has not filed audited subscriber figures with any regulator, so those numbers cannot be independently confirmed in the way a publicly traded telecom’s quarterly earnings can. Still, multiple external signals point in the same direction, and the growth trajectory has caught even skeptics off guard.
What independent data actually shows
The most rigorous outside check on Starlink’s footprint comes from network researchers, not from SpaceX press materials. A 2024 preprint study titled “A Large-Scale IPv6-Based Measurement of the Starlink Network” used IPv6 scanning to estimate active Starlink routers across countries and regions. Because the method counts real hardware responding on the public internet, it sidesteps the question of whether SpaceX’s own definitions of “active user” match what regulators or analysts would accept. The study, which has not yet completed formal peer review, confirmed a broad and growing global presence consistent with the company’s general claims, even if exact totals differ.
Inside the United States, the Federal Communications Commission maintains the National Broadband Map, which tracks fixed broadband availability down to individual addresses. The FCC publishes these datasets in CSV and GIS formats, letting researchers and journalists check coverage claims location by location rather than relying on national averages. The agency’s Broadband Data Collection program requires providers, including Starlink, to report where service is available and how many subscribers they have. Consumers and local governments can challenge inaccurate filings, creating a feedback loop that sharpens the map over time.
Together, the IPv6 research, the FCC’s broadband datasets, and the BDC reporting framework form the strongest available foundation for evaluating Starlink’s impact. None depends on SpaceX’s own accounting.
The measurement gap cuts both ways
SpaceX says rural dead zones are shrinking faster than it originally projected, but the company has never published that original timeline. Without a public baseline, the claim functions as a talking point rather than a testable statement.
Federal data introduces its own distortion. Satellite services can expand coverage far more rapidly than the FCC’s reporting cycle captures. A household that activated Starlink last month may not appear as “served” on the federal map for months. That lag can make official maps overstate the number of remaining dead zones at any given moment, understating satellite-driven gains. On the other hand, provider-reported availability data sometimes overstates coverage by listing areas where service is technically offered but where equipment costs, capacity limits, or waitlists prevent actual adoption.
The result: company claims may run ahead of what regulators can verify, while government maps may trail behind what consumers already experience. Neither picture is complete on its own.
Speed, cost, and the congestion question
Even where Starlink is available, availability does not automatically equal meaningful access. The Starlink kit, as of early 2026, costs $499 for standard hardware, with monthly service starting at $120 in the United States. For households in Appalachia, on tribal lands, or in sub-Saharan Africa, those prices can be prohibitive without subsidies.
Federal programs could help, but the picture is complicated. In 2022, the FCC revoked an $886 million Rural Digital Opportunity Fund award to SpaceX, citing concerns that Starlink could not meet the program’s speed and latency benchmarks. The Biden administration’s Broadband Equity, Access, and Deployment (BEAD) program has since directed billions toward state-level broadband buildouts, though most of that funding targets fiber and fixed wireless rather than satellite. Whether future subsidy rounds will treat LEO satellite as an eligible technology remains an open policy question heading into June 2026.
Performance is another variable. Ookla’s Speedtest Intelligence data has shown that median Starlink download speeds in the U.S. declined as the subscriber base grew, a predictable consequence of more users sharing finite bandwidth in each coverage cell. A community may look well served on a coverage map while residents experience slowdowns during peak evening hours. Neither the FCC maps nor the IPv6 study directly quantify how close any given area is to those capacity limits.
Competitors are circling the same market. Amazon’s Project Kuiper plans to begin commercial service with its own LEO constellation, and Eutelsat OneWeb already offers connectivity in high-latitude regions. More competition could drive prices down and push capacity higher, but none of these alternatives has matched Starlink’s current scale.
What this means for people still waiting for broadband
For rural Americans, the most actionable step is also the simplest: check the FCC’s National Broadband Map for your address. If it inaccurately shows your home as served, file a challenge. Those corrections feed directly into funding decisions that determine which areas qualify for federal broadband investment. Outside the United States, where comparable public mapping systems are rare, independent studies like the IPv6 research offer the best available window into where Starlink is actually operating.
The best-supported conclusion as of May 2026 is that Starlink is materially expanding high-speed options in hard-to-serve locations around the world. The IPv6 data confirms a broad and growing footprint. The FCC’s systems are gradually adapting to capture satellite-driven change. And millions of people who had no realistic broadband option a few years ago now have one.
But the precise scale of that expansion, whether it is 5.5 million users or somewhat more or fewer, and the speed at which dead zones are disappearing remain genuinely uncertain. Until SpaceX submits to independent auditing or regulators close the reporting lag, the boldest numbers should be treated as directional, not definitive. The story Starlink is telling about itself is probably in the right neighborhood. Pinning down the exact address will take better data and more time.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
