Satellite Internet Timeline 1990–2026: Starlink, Kuiper & the Space Broadband Race

The two animated pulses above are illustrative, not a live measurement — but the ratio is real. A GEO satellite sits so far out that it takes roughly a quarter of a second for a radio signal to make the round trip at the speed of light, before any equipment even processes it. A LEO satellite a few hundred kilometres up cuts that same trip down to a few milliseconds. That is the entire physics story behind why Starlink feels faster than the satellite internet people remember from the 2000s: it isn’t a better satellite, it’s a much shorter commute.

The catch is what a GEO satellite gives up by moving closer. Because GEO orbits at exactly the altitude where a satellite’s orbital period matches Earth’s rotation, it appears to hang in one spot in the sky — one dish, pointed once, keeps working indefinitely. A LEO satellite has no such luxury: it races across the sky and disappears over the horizon within minutes, so a LEO network only works at all if there are enough satellites, phased-array antennas that can track a moving target, and software that hands a connection to the next satellite before the current one drops out of range. LEO broadband is not simply “satellites, but closer” — it is a completely different engineering problem that trades one satellite’s simplicity for a fleet’s complexity.

The FCC placed a draft order on the agenda for its 30 September 2026 open meeting that would unlock more than 1,000 MHz of spectrum across the 12 GHz and 42 GHz bands. According to the commission’s own agenda materials and multiple trade-press reports, the order is framed around boosting capacity for high-speed satellite broadband delivered to homes, in-flight Wi-Fi, on-ship connectivity, and the traffic-routing functions that satellite ground networks rely on internally.

This sits alongside, but is distinct from, two other 2026 FCC spectrum items worth knowing about: a “Spectrum Abundance for Weird Space Stuff” rulemaking process aimed at next-generation orbital missions more broadly (an earlier-stage proceeding, published in the Federal Register in April 2026), and a separate “Unleashing Unlicensed Spectrum for Direct-to-Device” notice (Federal Register, 8 September 2026) focused on unlicensed spectrum for IoT-style device-to-satellite links, not carrier-grade phone service. It is easy to conflate these three actions in headlines; they are related but separate regulatory tracks.

What still has to happen: a scheduled agenda vote is not automatic approval — commissioners can amend, delay or vote against an item. Even if adopted on 30 September 2026, an FCC order typically still needs to be published in the Federal Register and survive its effective-date and any legal challenges before operators can actually use the new spectrum commercially. The 12 GHz band in particular has been the subject of a long-running dispute between satellite operators and terrestrial 5G proponents over interference risk, so opposition or conditions attached during implementation are realistic possibilities, not just formalities.

Why it matters regardless: spectrum is the raw material every satellite broadband and direct-to-device service ultimately depends on. Regardless of the exact final shape of this order, its direction signals that US regulators see satellite capacity as a growing, contested resource worth actively expanding — not a settled, secondary technology.

🔭 Astronomy Interference

Astronomers have raised sustained, documented concerns about satellite constellations. Research published in 2026 in the Monthly Notices of the Royal Astronomical Society found that nearly all spacecraft from major constellations exceed the 7th-magnitude brightness limit set as a guideline for professional astronomical research by the International Astronomical Union’s Centre for the Protection of the Dark and Quiet Sky (IAU CPS). Bright, moving satellites can leave streaks across long-exposure telescope images and complicate radio astronomy. Mitigation is real but partial: SpaceX’s newer “Starlink Mini” satellites at 550 km are measurably fainter than earlier Starlink models, and coordination between operators and astronomy bodies continues, but the underlying tension between more satellites and darker skies has not been resolved.

🛰 Orbital Congestion and Collision Risk

Space is vast, but the useful low-orbit shells satellites actually want to occupy are not infinite in practical terms. SpaceX reported that Starlink satellites performed roughly 355,000 collision-avoidance manoeuvres over a recent 12-month period — more than three times the number performed in 2024 — and researchers have documented multiple-day windows in 2026 where conjunction (close-approach) risk assessments briefly spiked. ESA’s own 2026 Space Environment Report documented a roughly 20 percent year-on-year rise in collision probability in the busiest low-orbit bands. None of this means an uncontrolled cascade of debris (sometimes called Kessler syndrome) is imminent or inevitable — automated collision-avoidance systems have kept major constellations’ safety records clean so far — but researchers describe the trend as a rising, not yet catastrophic, operational risk that deserves continued monitoring rather than either alarm or dismissal.

📡 Spectrum Competition

Two satellites, or a satellite and a terrestrial network, can occupy entirely different physical locations and still interfere with each other if they try to use overlapping radio frequencies without proper coordination. This is exactly the tension behind the long-running 12 GHz dispute between satellite operators and terrestrial 5G proponents, and it is a core reason the FCC’s September 2026 spectrum order (above) is contested rather than routine. Direct-to-cell services add another layer of complexity, since they typically borrow a mobile carrier’s own existing terrestrial spectrum rather than using dedicated satellite frequencies, requiring close technical coordination to avoid interference with normal cellular service.

🚀 Launch Frequency and Space Traffic

Mega-constellations require sustained, frequent launches, not a one-time deployment — both to complete a constellation and to replace satellites that reach the end of their roughly 5–7 year design life. That steady cadence is itself a form of traffic growth in low orbit that regulators and operators are still building the coordination tools to manage at scale.

Across all three scenarios, one constraint holds: dense cities will very likely keep relying heavily on terrestrial networks, because towers and fibre offer far greater local capacity, better indoor coverage, denser spectrum reuse, and a lower infrastructure cost per delivered bit in areas with enough people to share the cost. Orbital and spectrum regulation, not just satellite technology, will likely decide which of these scenarios ends up closest to reality.