By mid-2026, SpaceX says its Starlink constellation serves 5.5 million active subscribers across 118 countries, and that rural connectivity gaps are closing faster than the company’s own internal projections anticipated. If accurate, the numbers mark a steep climb from the roughly 4 million subscribers SpaceX disclosed in September 2024. For farming families in Montana valleys, emergency crews on wildfire perimeters, and fishing villages in Indonesia that have never had a reliable broadband option, the trajectory matters more than the marketing. But verifying SpaceX’s claims requires data the company has not yet released, and independent researchers are only now assembling the tools to fill that gap.
What independent measurements actually show
Two peer-independent studies, both posted as preprints on arXiv and neither funded by SpaceX, offer the most transparent look at how Starlink performs under stress.
The first produced a longitudinal dataset called WetLinks, pairing months of continuous Starlink throughput, latency, and availability readings with time-matched weather records. The value is in the duration: satellite services have historically looked strong in short demos but degraded during sustained rain or snow, exactly the conditions that matter during a three-day storm when a rancher needs to reach a veterinarian or a volunteer fire crew needs situational updates. Because the raw data are public, any researcher can check whether link stability improved as SpaceX launched additional satellites and pushed firmware updates.
The second study tackled mobility. A team drove a Starlink terminal across Central Europe on a defined set of routes, logging throughput, latency, and outage events as the dish handed off between satellite beams and coverage cells. The results are directly relevant to long-haul truckers, rural ambulance services, and anyone who needs connectivity while moving through areas with no cell towers. The researchers recorded where the link dropped and how quickly it recovered at coverage boundaries, producing a route-level picture that a single speed test screenshot cannot provide.
Together, the two papers establish a measurement baseline. The WetLinks data can be revisited in future years to detect genuine improvement or regression. The mobility study sets a benchmark for Starlink’s growing vehicle, marine, and aviation product lines. Both datasets are freely available, meaning a rural school board weighing a terminal purchase can consult the same evidence as a telecom regulator.
Where SpaceX’s claims outrun the evidence
SpaceX has not published the methodology behind its 5.5 million user figure. It is unclear whether “active users” means subscribers with a current billing relationship, terminals that transmitted data in a given month, or something else. The distinction is practical, not pedantic. A count inflated by dormant accounts or trial users would overstate the service’s real reach in underserved areas, while a device-based metric might undercount shared connections at schools, clinics, or community Wi-Fi hubs that serve dozens of people from a single dish.
The claim that dead zones are shrinking “faster than expected” raises a similar transparency problem. SpaceX has not released the internal forecasts that define its baseline expectation. Without that reference point, outside analysts cannot judge whether the acceleration is meaningful or whether the original projections were set conservatively to guarantee a good headline later. The phrase functions as marketing until the underlying models become available for review. Regulators and development agencies considering terminal subsidies for remote communities have no independent way to confirm whether SpaceX is genuinely outpacing its plans.
Neither arXiv study was designed to validate country-level coverage claims. The WetLinks dataset can show whether outage frequency declined over its collection window, and the mobility study can reveal whether edge-of-cell performance was usable, but neither maps dead-zone closure at the county or census-block level. In practice, a village seeing new dishes appear on rooftops can infer local adoption but cannot know whether the next valley over remains offline.
Context the numbers leave out
Several factors complicate any blanket statement about Starlink closing the digital divide.
Geography. The mobility study was conducted in Central Europe, a region with relatively dense orbital coverage and moderate terrain. Performance in equatorial zones, mountainous areas, or the high-latitude communities where many indigenous populations live could differ substantially. Snow load on terminals, vegetation, and horizon obstructions all affect link quality. Generalizing from Central European road tests to global dead-zone reduction requires caution.
Congestion. As more users sign up within a single satellite cell, per-user capacity can drop unless the constellation and ground station network scale in parallel. The current independent datasets capture performance during specific windows of adoption. They do not yet show what happens when a rural region transitions from a handful of early adopters to widespread use, or when a wildfire evacuation suddenly pushes hundreds of users onto the network at once.
Cost. Starlink’s standard residential plan runs roughly $120 per month in the United States, with a terminal that costs $499 upfront. In lower-income countries, SpaceX has introduced reduced pricing, but the hardware cost alone can exceed a month’s household income in parts of sub-Saharan Africa and South Asia. Expansion to 118 countries does not automatically mean affordability in 118 countries, and the company’s coverage map does not distinguish between areas where service is available and areas where people can realistically pay for it.
Competition and alternatives. Amazon’s Project Kuiper has begun prototype satellite launches and aims to offer a competing LEO broadband service. Eutelsat’s OneWeb constellation is operational and targeting enterprise and government contracts. Meanwhile, terrestrial fixed-wireless providers and government-funded fiber buildouts continue to reach new rural areas. Starlink’s growth is impressive in isolation, but the broader question is whether satellite broadband is the most cost-effective path to closing a given community’s connectivity gap, or whether it serves best as a stopgap until ground infrastructure arrives.
How to weigh the evidence before buying
For anyone making a practical decision about Starlink, whether for a remote homestead, a school, or a mobile operation, the most useful approach is to separate the evidence into tiers.
The strongest tier is primary measurement data from independent researchers with transparent methods. The WetLinks dataset and the Central Europe mobility study both qualify. They provide raw numbers that can be checked, reproduced, and extended. When these studies report usable speeds or stable handoffs, the findings carry weight because the methods are open and the data can be reanalyzed as the constellation evolves.
The second tier is SpaceX’s own announcements. Corporate claims about user counts and geographic reach are informative but self-reported. SpaceX files with the FCC for spectrum licenses and participated in the Rural Digital Opportunity Fund, but it is not required to disclose granular subscriber data the way a publicly traded carrier reports to the SEC. Until the company opens its internal metrics to independent audit, its figures should be treated as directional rather than precise.
The weakest tier is anecdotal: user testimonials, Reddit threads, and social media speed-test screenshots. These can surface trends worth investigating, but a farmer in Wyoming posting about faster downloads does not tell us whether a clinic in rural Peru experienced the same gain, or whether those speeds hold up during storms and peak evening hours.
What SpaceX still owes the public
Starlink has expanded its constellation and user base at a pace no other satellite operator has matched, and early independent measurements suggest the system can deliver broadband-class performance under many conditions. But the specific claim that dead zones are shrinking faster than expected remains unverifiable until SpaceX publishes the forecasts it is measuring against or allows third-party audits of its coverage models.
Independent researchers have built the tools to test that claim. The next step belongs to SpaceX: release the baseline data, or accept that outside analysts will define the benchmark on their own terms, using open datasets and transparent methods to map where satellite internet is genuinely closing the gap and where it still falls short.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
