Newcomers to the space economy tend to treat orbit and spectrum as separate decisions: where the satellites fly is an engineering question, what frequencies they use a regulatory one. The industry’s veterans know better. The two are so tightly coupled — by physics first, and then, more rigidly still, by seventy years of regulation that froze the physics of earlier eras into law — that every satellite product that has ever succeeded is best understood as a pairing: direct-to-home television is GEO-plus-Ku, satellite navigation is MEO-plus-L, broadband constellations are LEO-plus-Ku/Ka, and the direct-to-device gold rush is, as we shall see, the violent consequence of a pairing the old map never anticipated. Understanding why the pairs form as they do is the fastest route to reading the industry — and to designing anything new within it.
The physics that binds them
Five physical couplings do most of the work.
Distance sets the power budget. A geostationary satellite sits at 35,786 kilometres; a low-orbit broadband satellite at a few hundred. Radio power thins with the square of distance, so the GEO path is weaker by a factor of thousands — tens of decibels — before anything else is considered. And because free-space loss also grows with frequency, distance and band compound: high frequencies from high orbits demand large antennas, high power, or both. This single equation explains half the map: it is why GEO systems historically favoured big dishes and rain-resilient lower bands for wide coverage, why closing a high-frequency link to a small cheap terminal pushes you downward in altitude, and why the tiny antenna in a mobile phone — the weakest terminal of all — can realistically be served only from low orbit.
Distance sets the clock. Light takes roughly a quarter of a second to make the GEO round trip — perceptible in conversation, fatal to interactive applications — while LEO round trips run in the tens of milliseconds. Latency is a service characteristic, services map to spectrum allocations, and so the latency of an orbit quietly constrains which regulatory services, and hence which bands, make commercial sense from it. Broadcast, which tolerates any latency, married GEO; interactive broadband’s move to LEO is, at bottom, a latency migration that dragged whole bands along with it.
Motion sets the terminal. A GEO satellite hangs stationary in the sky, so its ground antenna can be a fixed dish installed once by anyone — the property that made satellite television a mass consumer product. Low-orbit satellites sweep the sky in minutes, demanding tracking: steered dishes, electronically scanned arrays, or antennas so unfussy about direction that tracking is unnecessary. That last option only exists at lower frequencies, where near-omnidirectional antennas work — which is why LEO Internet-of-things systems cluster in VHF, UHF, L and S bands, and why LEO broadband had to wait for affordable phased arrays before Ku- and Ka-band constellations could reach consumers.
Motion also smears the signal. Fast-moving satellites impose Doppler shifts and rapid handovers that complicate receivers — trivial for modern wideband systems, but historically another reason simple, cheap receiving equipment paired with the motionless GEO arc.
Geometry sets the map. One GEO satellite sees a third of the Earth but the poles poorly or not at all; continuous high-latitude service belongs to inclined and highly elliptical orbits — the Molniya-and-Tundra family that Russian communications and polar-focused systems have always used — or to LEO constellations, whose global lattice covers everywhere as a by-product. Meanwhile the low-elevation, long-atmosphere paths that plague high latitudes punish high frequencies hardest, coupling geography, orbit and band in one knot. And in the opposite direction, constellations changed the arithmetic of spectrum itself: a LEO system reuses the same frequencies thousands of times across its moving cells, extracting vastly more capacity per megahertz than a single GEO beam ever could — which is precisely why the constellations’ operators covet wide bands and why the incumbents of the arc resist their terms of entry.
The regulation that binds them harder
If physics proposes the pairings, regulation ratifies them — and then outlives the physics.
The ITU’s table of allocations is organised by service — fixed-satellite, broadcasting-satellite, mobile-satellite, Earth-exploration, radionavigation — and the services carry orbital assumptions in their bones. The broadcasting plans that guarantee every state a position are GEO plans. The allotment endowments are GEO endowments. The coordination arc — the mechanism deciding who must negotiate with whom — is defined around the geostationary arc and exists for no other orbit. Most consequentially, Article 22 of the Radio Regulations makes seniority explicit: in the core Ku- and Ka-bands, non-geostationary systems must protect geostationary networks within hard power limits, a hierarchy written in the 1990s as the price of letting the first NGSO constellations into the incumbents’ bands at all. The result is that an orbit choice is simultaneously a rank choice: the same megahertz that make a LEO system junior and constrained in one band leave it unencumbered in another, and a constellation’s regulatory strategy is largely the art of picking which seniority structures to live inside. Add the machinery that attaches to orbits specifically — the milestone regime exists for constellations, not for single GEO satellites — and the register itself enforces the marriage: you do not file for frequencies, you file for frequencies in an orbital configuration, and the pair is what the queue records.
The critical practical insight is that this legal map encodes the physics of its drafting eras. GEO seniority in Ku-band reflects the 1990s balance of power; the plans reflect 1970s equity politics; the terrestrial mobile allocations now coveted for direct-to-device were drawn when no satellite could conceivably speak to a phone. When technology moves — phased arrays, mass-manufactured constellations, sensitive receivers — the physics decouples from the law, pressure builds along the seam, and the great regulatory fights of each decade erupt exactly there. Which brings us to the present.
The pairings that run the industry — and the one breaking it
Read as orbit-frequency pairs, the industry becomes legible at a glance. Each row of the table below is a product the market has already validated. In every case the orbit supplies the geometry and the economics, the band supplies the link and the regulatory position — and the pairing, not either half alone, is the business.
| Pairing | The product | Why the physics pairs them | The regulatory & commercial position |
|---|---|---|---|
| GEO + C/Ku | Television distribution; legacy connectivity | A motionless satellite allows cheap fixed dishes; C-band shrugs off rain across wide beams | The GEO plans and the arc’s seniority protect incumbents — a mature, well-defended estate |
| GEO + L/S | Classic mobile-satellite service (maritime, aviation) | Low frequencies reach small mobile terminals; one satellite covers an ocean | Scarce MSS allocations held by long-standing incumbents — now the edge of the D2D battleground |
| MEO + L | Satellite navigation (every GNSS) | Medium orbit gives whole-Earth visibility from ~24–30 satellites; L-band penetrates weather and foliage to cheap receivers | State-run systems in protected navigation allocations; not a commercial filing target |
| MEO + Ka | Equatorial data constellations | The latency-capacity compromise: closer than GEO, far fewer satellites than LEO, Ka-band for capacity | A niche pioneered by one operator family; lightly congested |
| LEO + UHF/L/S (narrowband) | Satellite Internet of things | Tiny messages to near-omnidirectional terminals; proximity substitutes for power | A crowded space of small allocations; WRC-27 is studying new narrowband slivers |
| LEO (sun-synchronous) + X/Ka downlink | Earth observation | The imaging orbit paired with high-capacity data-dump bands; polar passes are why polar ground stations exist | Shared EESS allocations, congested at the classic X-band downlink |
| LEO + Ku/Ka | Broadband mega-constellations | Massive frequency reuse across moving cells; needed phased arrays to reach consumers | Junior to GEO under Article 22 power limits — a seniority now being litigated jurisdiction by jurisdiction |
| LEO + terrestrial mobile bands | Direct-to-device | The phone’s antenna forces the satellite low and large; the phone’s radio forces the spectrum terrestrial | The frame-breaker: no settled regime — improvised authorisations, licence mega-deals, WRC-27 studies |
The last row deserves its own paragraph, because it is breaking the map that produced all the others. In direct-to-device, physics dictates both halves of the pairing and leaves nothing to negotiate. On one side of the link, the terminal is a phone: its antenna is tiny and its transmit power minuscule, so the satellite must fly low and grow large to close the link. On the other side, that phone’s radio speaks only the bands terrestrial networks use — so the spectrum must be the very allocations the treaty map handed to ground-based mobile decades ago, when no satellite could conceivably join the conversation. An orbit compelled by physics is thereby forced into frequencies forbidden by history. Every headline fight of the current cycle — satellite operators paying tens of billions for terrestrial licences, regulators improvising national authorisations, WRC-27 studying satellite use of the mobile bands — is this single pairing forcing the law to catch up with the link budget. Nothing demonstrates better that the orbit-frequency relationship is not a background technicality but the industry’s tectonics.
The hands-on sequence for anyone designing a system
For an operator at the drawing board, the couplings above compress into a decision sequence that should run in this order — because each step constrains the next.
Start from the terminal, not the satellite. What antenna can your user realistically host — a fixed dish, a tracking array, a palm-sized patch, a phone? The terminal’s size, power and steering ability set the joint envelope of viable orbits and bands more tightly than any other factor.
Let the service set the clock. If your application tolerates latency, the whole map is open, GEO’s economics included; if it doesn’t, you are constrained low, with everything that implies for constellation size and capital.
Let geography vote. Serving high latitudes, or the whole globe continuously? Certain orbits are eliminated before spectrum is even discussed.
Then choose the band as a regulatory position, not just a link budget. Among the physically viable bands, compare the seniority structures: where would you rank, whom must you protect, how congested is the queue, what does coordination history in that band look like? A band that closes the link but buries you behind a decade of hostile coordination may be worth less than a physically inferior band where you would file senior and alone — this is the point where the engineering analysis and the filing strategy from the earlier guides in this series become one conversation.
Finally, stress the pair against the future. The map is moving: the power-limit rules that define LEO’s rank beneath GEO are being rewritten jurisdiction by jurisdiction, the mobile bands are being pried open to satellites, and each WRC redraws seams. A pairing chosen for today’s law alone is a position taken on yesterday’s physics.
The one-sentence version, for the front of the notebook: in this industry you never own frequencies and you never own an orbit — you hold a position in a pairing of the two, and the pairing, with its physics, its rank and its clocks, is the asset.