Coverage is going everywhere. The money won't follow. · Analyst edition

In April 2026, Amazon announced it was buying Globalstar, the satellite company that handles Apple's Emergency SOS calls for iPhone users off the grid. The deal puts a hyperscaler one layer below one of the most consequential user experiences in consumer electronics.

The weighted answer:

The direction, however, is not in doubt. Coverage is commoditizing. The question is who captures the margin above the commodity layer, and the five frameworks this panel ran have a mostly consistent answer: not the carriers.

How we got here

BlueWalker 3 validated the LEO-to-handset phased-array approach in 2020: a satellite with sufficient antenna aperture to communicate directly with unmodified handsets using 3GPP cellular protocols. Two years of regulatory waiting followed, as the FCC had not yet authorized commercial satellites to share licensed terrestrial spectrum.

FCC supplemental coverage from space (SCS) rules cleared in March 2024. AST SpaceMobile's BlueBird Block 1 launched in September 2024, making it the second competing implementation of the 3GPP NTN interface and moving coverage from Custom-Built to Product stage. By April 2026, Amazon had acquired Globalstar, placing a hyperscaler directly into the satellite layer that Apple depends on.

The deeper pattern

T-Mobile, AT&T, and Verizon have all signed satellite coverage partnerships. Each made rational sense at the time. The collective result is that they are jointly commoditizing geographic coverage, the single most defensible pricing advantage US carriers have held for decades. It justified thirty years of tower builds, capex cycles, and spectrum auctions.

Disruptive Innovation (Christensen) calls this the profit paradox: reported metrics improve while the structural basis of pricing power erodes. Wardley Mapping (Wardley) adds the Red Queen Effect: once T-Mobile adds satellite backup, AT&T must follow, then Verizon, until all three have it and none earns a dollar more from it. The treadmill runs faster. Nobody gets ahead.

The carriers cannot win this game by participating. They cannot afford to sit it out.

The five frameworks

Throughput Accounting (Bragg): the bottleneck is phased-array manufacturing

A system's output is governed by its binding constraint. For AST SpaceMobile, the binding constraint is the AST5000 ASIC production rate. Throughput Accounting, Bragg's framework for scoring decisions by revenue minus truly variable cost against the binding constraint rather than by standard cost allocation, makes the logic direct: until the phased-array manufacturing bottleneck is elevated, AST's throughput grows slowly regardless of demand or commercial agreements. Block 1 supplier failures in 2023 seeded this constraint; it compounds through the Block 2 phase, and the analysis treats it as persistent through at least 2027.

SpaceX avoids this constraint through vertical integration of satellite production and launch. That asymmetry gives SpaceX a durable cost-of-provision advantage. It is the primary structural reason the fragmented scenario holds 40% weight.

Throughput Accounting also partially dissents from the OEM-capture-only reading: once a constellation is built, the truly variable cost per additional subscriber is near zero. An efficient-scale operator like SpaceX earns real volume margin on the coverage layer even as per-unit prices fall. Two margin pools exist, OEM above and efficient-scale operator below. Carriers are squeezed on both sides.

Modularity / Design Rules (Baldwin and Clark): the substitutable interface

Modularity / Design Rules is the panel's structural outlier: it dissented from all four scenarios. Its case against fragmentation: the visible interface (the published 3GPP NTN Release 17/18 standard) is already enabling multi-provider entry faster than a stable-fragmentation picture requires. Its case against OEM capture: two competing OEM nodes (Apple on consumer safety, Amazon on enterprise logistics) mean the architectural apex is contested, not controlled by a single player. Its case against SpaceX Direct: 3GPP NR authentication requires a mobile network operator; the interface standard prohibits carrier bypass without an international standards change SpaceX cannot engineer unilaterally. Its case against carrier integration: working against an already-modularized standard adds complexity without delivering architectural control.

The dissent from all four scenarios reflects a coherent position: the modular architecture is already in place and advancing, and none of the four proposed end-states fully accounts for that fact.

Complexity Economics (Arthur): where the increasing returns sit

The coverage layer faces diminishing returns. Adding satellites past the point of full geographic coverage provides no additional value to the phone user. Operators are building infrastructure, and infrastructure gets priced like infrastructure.

The OEM layer is structurally different. Apple's more than 100 million US iPhones represent a platform where satellite capability can be delivered at near-zero marginal cost per user. Each new safety feature makes the ecosystem stickier and Apple's negotiating leverage against any satellite provider stronger. Once coverage is interchangeable, the satellite provider needs Apple's distribution. Apple, past that point, needs no specific satellite provider. Increasing returns compound the advantage as coverage commoditizes.

Amazon, through Globalstar, is building a second distinct node targeting enterprise logistics, agricultural IoT, and maritime monitoring: the application niches that cheap universal coverage enables. Complexity Economics predicts this pattern. Commoditizing substrates open new niches above them, populated by agents who couldn't exist before.

Disruptive Innovation (Christensen): the S-curve and what happens at performance adequacy

The LEO phased-array technology S-curve (tracking improvement rate over time) is in its steepest phase since the 2022 architecture validation. Performance adequacy, the point where the technology is good enough for routine, non-emergency use, arrives in 2026-27 on current trajectory.

This is why Disruptive Innovation dissents from the Fragmented Multi-Stable State (40%) while still assigning it the largest weight: the S-curve inflection tends to break fragmented equilibria, and the fragmentation thesis requires the inflection to stall. Disruptive Innovation gives OEM Vertical Capture (28%) and SpaceX Direct (17%) high confidence, gives Carrier Vertical Integration (15%) medium confidence as a rational but fraught defensive move, and explicitly opposes fragmentation as the technology maturation curve advances.

Wardley Mapping (Wardley): component positions and movement

The coverage layer sits at approximately 0.45 on the evolution axis as of mid-2026, confirmed mid-Product by pricing opacity (no published commodity tariffs). The value-chain apex, Device OEM OS Integration, is the margin concentration point.

Wardley Mapping favors both fragmentation (pricing opacity confirms mid-Product stage) and OEM capture (the Innovate-Leverage-Commoditise pattern, seeding applications while commoditizing the coverage layer beneath, is already operating at the apex). It opposes SpaceX Direct at the structural level: SpaceX's spectrum access depends on T-Mobile's FCC SCS license, making the partnership a constraint rather than merely a business preference. It identifies the one genuine benefit of Carrier Vertical Integration: MNO capital injection could elevate the phased-array manufacturing constraint faster than AST could alone.

The four futures

40%: Fragmented Multi-Stable State. Three parallel tracks co-exist through end-2028: Starlink/T-Mobile, BlueBird/AT&T+Verizon, Amazon-Globalstar for enterprise. Wholesale pricing stays negotiated. No open tariff appears. Coverage improves without completing the Product-to-Commodity transition. Supported by Complexity Economics (no adoption threshold crossed), Throughput Accounting (manufacturing bottleneck persists), and Wardley Mapping (mid-Product confirmed). Opposed by Disruptive Innovation (S-curve inflection will break this equilibrium) and Modularity / Design Rules (visible interface enables faster multi-provider entry).

28%: OEM Vertical Capture. Coverage commoditizes by 2027. Apple and Amazon lock distinct verticals above the commodity layer. Satellite operators become thin-margin wholesale providers, analogous to tower companies. Carriers retain distribution but earn no new per-subscriber D2D margin. Supported by Complexity Economics, Disruptive Innovation, and Wardley Mapping. Partially opposed by Modularity / Design Rules (two-OEM apex is contested, not captured by a single architecture owner) and Throughput Accounting (efficient-scale satellite operator also retains margin; this is not a complete OEM takeover).

17%: SpaceX Direct-to-Consumer Bypass. SpaceX migrates to direct consumer subscription once performance adequacy clears. T-Mobile deal renegotiated or ended. Supported by Complexity Economics, Disruptive Innovation, and Throughput Accounting. Opposed by Wardley Mapping (spectrum dependency is structural) and Modularity / Design Rules (interface standard requires carrier authentication).

15%: Carrier Vertical Integration. A major MNO acquires controlling equity in AST SpaceMobile before capital-cycle distress. Favors: Disruptive Innovation at medium confidence: one MNO will attempt this as a rational defensive move. Against: every other lens. Complexity Economics, Wardley Mapping (Red Queen and Commoditization Conflict operate independently of ownership), Throughput Accounting (fixed cost internalized at wrong point), and Modularity / Design Rules all oppose.

Milestones

Confidence summary: Coverage commoditization is directionally high-confidence (all five frameworks agree on direction). Timing is medium-confidence; a catalytic event, first open tariff publication, Apple feature expansion beyond SOS, SpaceX direct launch announcement, or MNO acquisition, would shift the weights materially. The fragmentation scenario is stable, not robust: it requires the S-curve inflection to stall past 2028.

On track record: This is the panel's first analysis of this topic. The scorecard shows zero resolved calls across all five lenses. Hit rate, calibration score, and predictive edge are undefined. These are opening positions.

The map

The value chain runs from deep infrastructure at the bottom (launch services, now effectively commodity) through spectrum licensing and phased-array manufacturing, through the coverage radio link, through carrier billing and distribution, through OEM device integration, and up to the emergency and consumer applications users actually interact with. Visibility increases going up the axis; evolution runs from Genesis (left, custom-built) to Commodity (right, interchangeable).

Component movements by scenario:

What this means

The coverage race is not really about coverage. It is about who controls the user relationship when connectivity becomes universal. For carriers, the structural position is uncomfortable in a precise way: investing in infrastructure that serves customers today while accelerating the erosion of the pricing power that justified that investment for thirty years.

For anyone building or investing above the coverage commodity, the structural signal across three frameworks is consistent. This is a once-in-a-generation infrastructure transition being funded by other people's capital. Emergency and safety applications are already staked. Agricultural IoT, maritime monitoring, and industrial location services are the next tier of niches that cheap universal coverage enables. The substrate is being built. The new margin will sit wherever the application lives.

The coverage layer is becoming plumbing. The margin is moving up, not sideways.