Fiber Castles, Cable Forts, FWA Camps, and Satellite Warbands: The Home Internet Battlefield Is Drawing Territorial Lines

Introduction

The US home internet competitive landscape is fragmenting into four distinct territorial formations, and these formations are reshaping wireless competition through convergence. We call them Fiber Castles, Cable Forts, FWA Camps, and Satellite Warbands. None of them are invincible. Castles have cracks. Forts have moats that can be crossed. Camps have poor defenses. Warbands can be run over. Every formation must be maintained and guarded, or it falls.

This report draws on Recon Analytics’ continuous consumer telecom survey covering both mobile and home internet (January 2025 to March 2026, 682,988 total responses of which 343,176 are home internet), measuring ISP cNPS (complete experience) across broadband technologies and providers. The survey data is supplemented by Q4 2025 carrier earnings disclosures, FCC BEAD program preliminary awards data, FCC spectrum allocation proceedings, and Census Bureau American Community Survey (ACS) mobility data.

The central finding: convergence, bundling wireless and broadband from a single provider, is the structural force that determines which formations hold and which fall. But convergence built on switching friction rather than customer satisfaction is a structurally weaker formation that erodes where a superior alternative arrives. The carriers that win are the ones that build formations their customers actually want to stay in, not just formations that are expensive to leave.

Fiber, FWA, and satellite broadband all score cNPS of approximately 27, while cable sits at 2.6, a 24.5-point gap. But the headline parity between fiber, FWA, and satellite is misleading. FWA and satellite cNPS are inflated by reference-point bias: customers who escaped bad prior experiences rate the same product dramatically higher. FWA-to-FWA switchers, who have no “escape from misery” uplift, rate FWA at cNPS 15.9, nearly half the headline number.

AT&T’s fiber convergence strategy has a visible crack: AT&T wireless carries the lowest cNPS of the three Mobile Network Operators (MNOs), and when bundled with fiber, it drags down the broadband experience. Converged fiber cNPS is 22.3 versus 28.8 standalone. Happy people choose to bundle; bundling does not make them happier.

Cable operators are the fastest-growing wireless providers in the US but have the worst broadband cNPS of any formation. Cable is losing subscribers to both fiber (37.4% of departures) and FWA (14.3%). The wireless bundle slows the bleeding but does not stop it.

FWA Camps are drying up across the country as fallow spectrum capacity tightens. The capacity shortage is most acute in urban and suburban markets where mobile traffic density is highest. In urban and suburban markets, degraded FWA customers return to cable or move forward to fiber where available. Unlike rural markets, urban FWA refugees have terrestrial alternatives. The Camp erodes toward the Fort and the Castle, not toward the Warband. In rural markets, where capacity constraints emerge later, Starlink picks up the disaffected FWA subscribers until fiber arrives and displaces both.

Starlink and Amazon Leo together won over $700 million in preliminary BEAD awards covering more than 750,000 locations across the US. In these government-subsidized markets, the Warbands are entrenched by federal funding, feeding them like manna from the heavens. Amazon Leo is the largest BEAD winner by number of locations despite not yet having a single commercial broadband customer.

Table of Contents

1. Executive Summary 2
2. The Convergence Engine 3
3. Churn Intent Is Not Churn 5
4. Fiber Castles 7
5. Cable Forts 9
6. FWA Camps 11
7. T-Mobile: Fiber Towers, Not Fiber Castles 13
8. Starlink Warbands 14
9. The Spectrum Wars 16
10. The “When the Beatings Stop” Effect 17
11. No Formation Is Permanent 18