SWITCHED ON
The daily technology series nobody asked for but everyone needed
Connected or Left Behind: Satellite Internet and the Digital Divide
About 2.6 billion people have never used the internet. Satellite constellations promise to change that. The promise is real. So are the complications nobody put in the brochure.
The internet has become so fundamental to economic participation, education, healthcare access, and civic life that being without it is no longer merely an inconvenience. It is a structural disadvantage that compounds every other disadvantage a person already has. Fixing that matters. The question is whether the way we are going about fixing it is actually going to work.
Yesterday we left the planet and went to space — reusable rockets, Starship, the genuine achievement of dramatically cheaper access to orbit, the Mars colonisation question examined without the hype, and the closing observation that the enthusiasm gap between "we're going to Mars" and "we're going to solve access to clean water" says something worth sitting with. Today we are using space to think about Earth — specifically, about the roughly 2.6 billion people who remain without reliable internet access, what satellite internet constellations are doing about that, what they are doing to the night sky and the orbital environment in the process, and whether internet access is a human right or a product, and why the answer to that question matters enormously for how we approach the problem.
Start with the scale of the gap, because it is easy to forget how large it is when you live inside the connected world.
01 — The Scale of the Disconnect
Approximately one third of the global population has never used the internet. This is not a problem that maps neatly onto poverty alone, though poverty is the dominant factor. It maps onto geography — rural and remote communities where laying fibre optic cable or building mobile towers is economically unviable under commercial models. It maps onto infrastructure — countries where the electricity grid is unreliable enough that devices cannot be consistently charged, let alone connected. It maps onto affordability — data costs that are trivial fractions of income in wealthy countries but represent meaningful portions of household budgets in lower-income ones. And it maps onto the compounding effects of historical underinvestment, where the absence of connectivity has prevented the economic development that would make connectivity commercially attractive to provide.
The consequences of this gap are not abstract. Education systems increasingly depend on digital resources — during the COVID-19 pandemic, the children of unconnected families lost educational ground that connected children did not, in ways that persist. Healthcare increasingly involves digital records, telemedicine, and access to health information. Financial services — banking, payments, credit — are increasingly digital-first in ways that exclude the unbanked and unconnected simultaneously. The job market, in almost every growing sector of every economy, rewards digital literacy and punishes its absence. Being offline is not simply being without a convenience. It is being structurally excluded from an expanding share of economic and social life.
02 — What Starlink Has Actually Done
SpaceX's Starlink is the most operationally mature low-Earth-orbit satellite internet constellation and the one that has most dramatically demonstrated what the technology can deliver in practice. As of 2026, Starlink operates over six thousand satellites in low Earth orbit — a number that was science fiction as recently as 2018 — and serves several million subscribers across more than one hundred countries. The service delivers speeds that are genuinely competitive with ground-based broadband in many use cases: download speeds typically in the range of 50 to 200 megabits per second, with latency of 20 to 40 milliseconds, compared to the 600-millisecond latency of older geostationary satellite internet that made real-time applications essentially unusable.
The difference between geostationary and low-Earth-orbit satellite internet deserves a brief explanation because it matters for understanding why Starlink is different from what came before. Geostationary satellites orbit at roughly 36,000 kilometres above Earth. At that altitude, they stay fixed over one point on the surface, which makes them useful for broadcasting but imposes unavoidable physics-based latency — light travels fast, but not instantaneously, and a round trip of 72,000 kilometres adds up. Low-Earth-orbit satellites at 500 to 600 kilometres have dramatically lower latency but must move relative to the ground, which means you need hundreds or thousands of them to provide continuous coverage rather than just a few dozen.
In remote Alaska, in rural Australia, on fishing vessels in the South Pacific, in Ukrainian military positions, and in disaster relief operations after earthquakes and hurricanes, Starlink has provided connectivity where nothing else could. This is not marketing. It is documented operational reality.
The hardware cost — the dish and router required to use the service — has fallen from $499 at launch to lower price points in some markets, though it remains a significant barrier in lower-income contexts. The monthly subscription cost similarly remains out of reach for many of the people who most need connectivity. Starlink has introduced lower-cost tiers and is working with governments and NGOs on subsidised access programs. The gap between "this technology can reach anywhere" and "this technology is affordable for the people who most need it" remains substantial and is not primarily a technical problem.
03 — The Competitors and the Constellation Race
Starlink is not alone, though it is far ahead of its competitors in deployment. Amazon's Project Kuiper has regulatory approval for over 3,200 satellites and has begun launches, with commercial service targeted for 2025 and 2026. OneWeb, now owned by Eutelsat, operates a constellation focused on enterprise and government customers rather than the consumer market. China's SatNet program is developing its own mega-constellation, with stated plans for tens of thousands of satellites. The European Union has its own constellation program, IRIS², intended to provide connectivity and secure government communications independent of US-controlled infrastructure.
The proliferation of constellation plans has produced a number that is worth pausing on. The total number of satellites currently approved and planned by various operators runs into the hundreds of thousands. The number of satellites that have ever been launched in the entire history of the space age, prior to the constellation era, was around ten thousand. The orbital environment is being transformed at a pace that the regulatory frameworks governing it were not designed to accommodate.
04 — What We Are Doing to the Sky
The astronomical community has been raising concerns about satellite constellations since Starlink's first launches, and those concerns are legitimate and increasingly well-documented. Satellites in low Earth orbit reflect sunlight and appear as bright streaks in long-exposure astronomical images. The scale of current and planned constellations means that for observatories in certain latitudes at certain times of night, a meaningful fraction of long exposures are contaminated by satellite trails. The Vera C. Rubin Observatory in Chile — designed to conduct a decade-long survey of the entire southern sky, generating the most comprehensive map of the universe ever attempted — has been specifically identified as facing significant interference from satellite constellations.
SpaceX has made genuine efforts to mitigate this, developing satellite designs with reduced reflectivity and experimenting with sunshades. The mitigations have helped. They have not eliminated the problem. And they apply only to Starlink. Amazon, OneWeb, and the Chinese constellation programs operate under their own design choices and their own degrees of engagement with the astronomical community's concerns. The cumulative effect of hundreds of thousands of satellites, operated by multiple entities with varying commitments to mitigation, has not been adequately modelled.
The orbital debris situation is the longer-term risk. More satellites mean more potential collision events. Collisions generate debris fields that can trigger cascade effects — the Kessler syndrome, in which debris from one collision strikes other satellites, generating more debris, which strikes more satellites, in a self-sustaining chain reaction that could render certain orbital altitudes unusable for generations. The probability of this outcome depends heavily on responsible deorbiting practices — ensuring defunct satellites are removed from orbit promptly rather than left to drift. Current regulations require deorbiting within five years. Compliance and enforcement are imperfect. The long-term sustainability of the orbital environment is a genuine concern that the current pace of constellation deployment is not adequately addressing.
05 — Is Internet Access a Human Right?
The United Nations Human Rights Council passed a resolution in 2016 affirming that internet access is a human right — or more precisely, that the rights people have offline must also be protected online, and that intentional disruption of internet access violates human rights. Several countries have gone further, encoding internet access as a constitutional or statutory right. The practical implications of this framing are significant, because they shift the question from "should the market provide connectivity to underserved populations" to "what obligations do states and international institutions have to ensure it."
The market answer to connectivity in unserved areas has historically been: none, because there is no profitable business model for serving them. Satellite internet changes the infrastructure constraint — it is now technically possible to provide connectivity anywhere on Earth's surface. It does not change the affordability constraint, which is economic and political rather than technical. A family in rural sub-Saharan Africa can be within the coverage footprint of a Starlink satellite and still be unable to afford the dish, the subscription, or the electricity to power them.
The digital divide, in 2026, is not primarily a technology problem. The technology exists. It is an affordability problem, a device access problem, a digital literacy problem, and a political will problem. Satellite internet has removed one significant constraint — geographic reach — while leaving the others largely intact. Celebrating the technology without engaging with those remaining constraints produces a version of progress that looks impressive from orbit and remains invisible from the ground.
Connectivity without affordability is infrastructure without access. And infrastructure without access is, ultimately, just a very expensive way of not solving the problem.
Tomorrow we are going somewhere that combines the scale of today's global connectivity story with the intimacy of Episode 12's neural interface discussion — electric vehicles and the 2035 question. The ban on new petrol and diesel car sales that multiple governments have committed to, whether the infrastructure is anywhere near ready, what the battery supply chain actually looks like, and whether the transition is happening fast enough to matter. See you then.
Switched On is a daily technology series covering AI, social media, data privacy, and the digital forces reshaping modern life — with no corporate spin, no false comfort, and absolutely no mercy for buzzwords.
