Borrowed Spectrum — Coverage Is Sliding From Asset to Common Component, and Exclusivity Is the Transition

Open your phone on a mountaintop and the antenna bar reads "No Service." And yet, from that very screen, a text message goes out. You didn't rent a satellite phone. You didn't unfold a dish. It is the phone that was in your pocket, unmodified. The industry calls this D2D — direct-to-device, a phone talking to a satellite with nothing in between. In the United States, this text has been an actual paid service for ordinary subscribers since the summer of 2025.

The question we answer in this piece is not "will dead zones disappear." What D2D changes is not the breadth of coverage but who owns coverage, and in what form. That the phone went unmodified does not mean the change was free — the price was paid in the switch. The differentiating asset that carriers built with their own capital — coverage — now has a path open to becoming a common component that competitors buy wholesale from the same satellite operator. When shops that each ran their own generator moved to buying electricity, the generator stopped being what set a shop apart — that is what is happening to coverage. And in a layer that has become a component, the value goes not to whoever makes the component but to the seat that switches it on and off — the phone's defaults and the service plan. This descent is not yet every market's story. Some markets are holding the layer down with exclusive contracts. Only one commercial deployment exists so far, and even that one is locked in exclusivity — so this is not a completed observation but our call on the direction, and at the end of this piece we spell out what would prove it wrong.

One more picture to carry. Hang a single lamp from the ceiling. The whole room brightens, but the light that reaches each individual desk is thinner than a desk lamp's. Hold on to this image and half the piece is already in hand. The other half is the question of whose hand is on that lamp's switch.

1. What changed this time was not the rocket — it was two permissions

Selling phone calls by satellite is not a new business. The path of using spectrum internationally allocated to satellites from the start — dedicated satellite bands — and selling dedicated handsets to match has been open for a long time — though what is new this time is not the rocket but two permissions that clear the threshold this business could never cross. That path never changed the phone in your pocket. Commercial service built on dedicated bands alone for ordinary phones remains at the scale of a few Pacific island nations, and the two exceptions that walk this path with a narrower use — emergency texting — appear in the table below.

The path this generation chose is the opposite. Instead of buying spectrum, it borrows. Every country has bands that the state allocated exclusively to carriers for terrestrial mobile service. Think of numbered parking spots in an apartment complex — lines drawn, one spot per household — except this time a car descending from the sky parks in the spot. What we call borrowed spectrum in this piece is exactly this one exception: terrestrial bands used from space.

For that exception to stand, two permissions are needed. One is the spot owner's consent. In the U.S., a satellite operator filed to lease a carrier's band before the rules even existed. The other is the regulatory regime. In March 2024 the U.S. regulator adopted rules opening terrestrial bands to supplemental use from space. The first commercial authorization came that November, and service opened the following summer.

Here we give the three roles names. Think of a landlord, a subcontractor, and a shopkeeper. The carrier — the license holder that leases out spectrum — is the landlord. The satellite operator that borrows that spectrum, builds coverage in the sky, and sells it wholesale is the subcontractor. The side that folds that coverage into a service plan and retails it to users is the storefront (the handset maker that holds the phone's defaults can also stand in this seat — we meet that case later). We use "lease" only for the one direction spectrum travels. Coverage flowing back to the carrier is not a lease; it is a wholesale purchase (the actual contract form — rates, revenue splits — is not disclosed, so "wholesale" here marks the direction money flows). And today's market has one overlap: the landlord and the storefront are the same company. This overlap is not a contradiction — it is the crux of this piece, and we will meet it again.

Each camp answered the fork differently.

The top two rows are the borrowed-spectrum path; the bottom two are the dedicated-band path. This piece's story is confined to the top two rows — the bottom two borrowed nothing, and we call on them later as counterexamples.

2. Five parts stand between an unmodified phone and orbit

Nothing was touched on the phone's side. So all the compensating has to happen on the other end.

The first part is the satellite's large antenna. A phone's transmit power is only what a palm-sized battery allows, so to hear that faint signal from hundreds of kilometers up, the ear in space has to grow.

The second part is correction in orbit. Low-orbit satellites sweep across the sky fast. Just as an ambulance siren bends in pitch as it passes (the Doppler effect), the frequency shifts, and signals take longer to make the round trip than on the ground. The satellite-side base station has to compute and cancel these distortions in advance, so the phone never knows it is talking to a satellite.

The third part is the regulatory exception from the previous section. The fourth is the standard — the mobile standard first defined a satellite segment in 2022. It works like a charging-port specification. Inside the spec, anything plugs in; outside it, you carry a proprietary cable. This non-terrestrial standard has to ride in phone chipsets before "any phone" becomes true.

The fifth is what the link actually carries — still text and low-rate data. Remove any one of the four parts and the chain breaks; the fifth is what the chain delivers. Cheap rockets appear nowhere in this chain — what opened it was two pieces of engineering, one regulatory regime, and one standard.

3. A beam that covers widely is necessarily thin

Flip the condition that made the chain work and you find the weak point. To catch faint phone signals, the antennas grew; to cover vast land with a finite number of satellites, the area each satellite illuminates — the beam — grew too. It is the ceiling-lamp condition exactly. A lamp that lights widely is thin at every desk.

We set the yardsticks apart. One commercial beam is about 20km across — more than half the area of Seoul — and every user inside it shares a combined 10MHz up and down. Independent researchers who pooled real-world usage from the fall of 2024 through the following summer measured one beam delivering a combined 3.1 megabits per second downstream. Compare that measured average with what a sparse rural cell tower theoretically handles, and it is more than twenty times thinner per unit area. Swap the measurement for the beam's theoretical maximum (18.6 megabits per second — an estimate), and about a fourfold gap still remains. The two pairings lean in different directions, so the multiple itself cannot be pinned down — but whichever pairing you take, the sign does not flip: the share reaching each spot is thinner than on the ground. Against an urban tower the gap widens by orders of magnitude, but that is an upper-bound illustration. The arena where D2D actually competes is the sparse countryside, and the fair comparison lives there.

The strongest objection stands right here: couldn't a camp that lofts a far larger antenna narrow the beam, grow the capacity, and push this ceiling up? That camp's beam area has never been disclosed, so a per-area comparison is impossible today. Only bandwidth is confirmed — take the 10MHz, split it, look at the downlink side alone, and it is 5MHz (downlink) versus 40MHz (direction undisclosed); read it either way and the gap narrows. We grant the point. But one confirmed fact remains: that camp, too, borrows its spectrum from the landlords. If the layer thickens, the product may outgrow insurance — our inference (there is no measurement of a thickened layer yet) is that what decides the value then is not thinness but the handset gate we meet in the next section. The form of ownership stays the same either way.

This weak point defines today's product. This thin layer cannot carry everyday traffic. What it sells is not bandwidth but insurance for the last cell — the mountaintop, the open sea, the city whose towers a disaster has switched off. Two special cases resolve by the same logic. A country could vertically integrate the whole stack — satellite band, handset, carrier (no real-world case that we could find; we note it only as a possibility). Even that would not refute componentization — the owner of the storefront would simply shift from a company to a state. And in a market where coverage is already dense and there seems to be nothing to sell, what shrinks is the premium, not the structure of the layer. We return to that market — our own — at the end of the next section.

4. The layer that carries the risk and the layer that holds the customer are not the same

If the thin layer is insurance, one question remains: where does insurance value attach? How a layer that underwrites only tail risk and never meets the end customer splits the value — that answer was worked out long ago, not by this industry but by insurance.

Recall the company name printed on your auto insurance policy. When you crash, that is the only company you call. That it quietly passed part of the risk to someone else is a fact most people live and die without learning. The side that contracts with you and owes the full payout is the primary insurer; the side that contracts only with the primary insurer and shoulders slices of large risks is the reinsurer. The risk-bearing layer and the customer-holding layer are different, and brand, pricing, and the power to cancel all stay with the side that holds the customer. In scale, too, the back layer today runs at about one-tenth of the front.

Push this lens to its conclusion and it actually predicts the opposite — subcontractor advantage. In insurance, what sets the direction of value is the scarcity of underwriting capacity — when capacity is abundant, as it is now, prices fall — and as we just saw, D2D's beams are structurally scarce. The reason we still judge that value stays with the storefront lies where the resemblance (the isomorphism) stops. A reinsurer never meets the end customer because of contract structure; a D2D subcontractor cannot become a storefront because its product is too thin to stand alone. On the borrowed-spectrum path, every way this insurance has been sold so far has been as an insert into a thicker product — the phone plan. And by definition a primary insurer does not lease its own assets to the reinsurer, whereas this game's subcontractor pays rent on top — one layer worse off than a reinsurer. That is as far as the lens shows; from here on, telecom's own facts.

Where in this layered structure the value could actually have been killed or crowned, a dead attempt left the coordinates. A chipset maker teamed with a satellite operator to push direct phone-to-satellite service, then folded. Both companies pointed to the same cause; the satellite operator's statement put it plainly — "the technology was developed and demonstrated, but smartphone makers did not adopt it." Not the network, not the regulator — the handset's decision was the gate.

The camp that detoured through dedicated bands completes the proof. The party that could actually walk that path — owing carriers nothing — was not a satellite operator but a handset maker. The right to put a satellite-band chip in the phone was the ticket, and one handset maker paid 95% of the new satellite network's capital expenditure to secure the feature. The path that looked like a counterexample to borrowed spectrum turns out to be the clearest evidence that value attaches to the storefront.

There is a rejoinder: isn't the landlord a gatekeeper holding the keys? A satellite needs the license holder's consent in every country, so read it as carrier leverage, not componentization. The evidence splits. There is one real case of a landlord locking a subcontractor into exclusivity — the only commercial deployment running today is exactly that form — while with the other subcontractor, two competing carriers signed side by side. We also record the condition our rejoinder stands on: it holds only while there are multiple landlords a subcontractor can switch to. And here the overlap from section 1 returns. That landlord is also the storefront. The subcontractor stands in a seat where the party it sells coverage to and the party it rents spectrum from are the same company — rent flowing up, no customers below. Value going to the storefront is not a matter of taste; it is the arithmetic of this overlap.

We also record the force pointing the other way. The sellers are few too — the borrowed-spectrum path has effectively two subcontractors, and scarce suppliers can name wholesale prices. But that power reaches only as far as wholesale. What decides componentization is not how many suppliers there are but whether buyers can beat each other with it — and as long as every competitor buys from the same subcontractor, coverage sets no one apart. The retail price tag and the cancel button remain on the storefront's screen. The subcontractor's road up into retail is blocked for the same reason — a thin product cannot anchor a standalone plan, and climbing to retail would again require the landlords' consent. On the dedicated-band path that seal breaks — and there, the seat of the storefront was taken not by a carrier but by the handset. One line we draw ourselves: the subcontractor's share of wholesale is undisclosed and we cannot see it. What this piece calls is the retail-and-defaults layer.

Two movements strike at that condition. The first is free-of-charge service. Emergency texting bundled free on a phone, standards-based texting provided at no extra charge, a satellite option folded free into premium plans. Isn't the value evaporating rather than attaching? Our judgment is absorption — and the ground is the retail facts. Every one of those freebies is either time-limited (free for two years after activation — whether charges actually begin after the period is not yet observed) or riding on the price of a premium plan, and the only real object with a retail price also sits on the storefront side: a ten-dollar-a-month satellite add-on sold even to competitors' subscribers. The claim that value vanished from the whole layer is already narrowed by that retail fact. What remains open is the size of the subcontractor's cut, which is undisclosed. We stand on the absorption side. The day it resolves is the day subcontractor receipts are disclosed — and if those receipts were in fact shrinking while the freebies spread, then evaporation was right.

The second movement is the landlords combining into a single bloc. In May 2026, the three competing U.S. carriers announced a joint venture pooling their satellite-use spectrum — still short of a definitive contract. There are two readings: the overture to carriers internalizing the satellite layer — a monopsony (few buyers, pricing power tilting to the buying side) — or a cost push-out with control retained. Under either reading, the test for componentization does not move. A differentiating asset is something you keep to yourself; you do not pool it with competitors — the moment it is co-owned, coverage sets none of the three apart. And the stronger the buying side grows, the more the value stays with the storefront, not the subcontractor. But how this venture hardens is also the falsifier of our final judgment — we meet it again in the last section.

We close this section closer to home. Korea looks, at first glance, like the "already dense" market — population-based 5G coverage runs in the 90s (a figure in the lineage of the government's 2019 target). But switch the denominator to land area and it is around 75% (a 2023 figure). We have no full map of mobile-coverage gaps, so the 5G number is only a coordinate — but that the picture flips with the denominator, that much stands. At sea, a dedicated maritime LTE network reaches 100km offshore. But turn to phones talking directly to satellites and things differ. Three low-orbit satellite supply agreements have been approved and commercial service has opened. Yet a public document that distinguishes, in writing, whether that is direct-to-phone or satellite internet requiring a separate terminal — we could not find one. That we could not find it is the full extent of our observation; we cannot say the regime makes no distinction. One question remains: if it hardens without the distinction ever being drawn, will this market's switch be set not by the regime but by the handset's defaults first?

5. Exclusivity is the transition; the component remains

Now the chain, the weak point, and the direction of value fold into one sentence. A layer built on borrowed spectrum is thin within this generation's physics; being thin, today's product is insurance; and the value goes to the storefront along all three forks — when the layer is thin, the storefront collects the premium; if it thickens, the handset gate decides (our inference); and even on the detour path, the ticket belonged to the handset. Unlike the previous generation that sold dedicated handsets, this generation left the phone alone — and handed the switch to the phone's side as the price. When coverage, once a differentiating asset, becomes a common component, carrier capital has one place left to go: outside the network — to defaults and plans. The same axis holds for readers. "Where does it get signal" is descending as a criterion for choosing a carrier; what remains is defaults and bundles.

What remains is the conditional we hung at the top. Tallying the contract facts: one deployment is commercial today, and it is exclusive. One non-exclusive contract exists but is not yet in service, and the joint pooling is an announcement short of a definitive contract. And the only time-series that shows a direction points the other way — the exclusive distribution rights in Europe and Africa were extended in 2025 out to 2034, and that carrier also holds 5% of the subcontractor. The U.K. finalized its regime more than a year and a half after the U.S.

On that record we make our call: exclusivity is the transition's form, and the equilibrium converges on non-exclusivity. As the count shows, this is not a summary of observations but a call on direction. Its basis is that the inside of exclusivity is already loose — even the exclusively bound camp sells its end product to competitors' subscribers, and the non-exclusive contract and the pooling announcement grew up beside it. But as the 2034 extension shows, the incentive to hold exclusivity is real too — which is why this call carries its falsifiers. If, in contracts signed or renewed after each market's regime is finalized — for the U.K., after December 2025 — exclusivity clauses persist and expand for years, this call is broken. If the U.S. joint venture hardens into a definitive contract that shuts out outside subcontractors or imposes worse terms on buyers outside the venture — the equilibrium being co-exclusivity, not non-exclusivity — it is broken then too. And if, within a few years, not a single non-exclusive commercial deployment ever stands, the directional call of componentization itself is broken.

Back to that screen on the mountaintop. The moment a text went out over the "No Service" icon, the lamp was already hanging from the ceiling. That light spread wide runs thin at every desk — physics had settled that too. On the day every light on the ground goes out — the day of disaster — the one thing that must still be on is that lamp. Who is positioned to answer that day is, in the end, also a question of the switch. In the years ahead, what to watch is not how many more lamps go up but how many hands hold the switches. The scorecard has four sheets: exclusivity renewal disclosures and the joint venture's definitive contract will grade the exclusivity call; subcontractor receipt disclosures will grade the absorption call; and a measurement of satellite-layer per-user throughput entering the same order of magnitude as the ground will grade this piece's entire commitment.