Insights
  • SpaceX

Starlink Adopters Kept Legacy Satellite at 31.3% but Cable at 69.4%

Starlink adopters kept legacy satellite at 31.3% versus a 67.6% non-adopter baseline; cable retention was 69.4% versus 86.0%.

Starlink adopters were much less likely to keep legacy satellite than cable

Share of prior-provider customers with a payment 1–7 months after the first Starlink charge, compared with calendar-matched non-adopters; Starlink starts Jul. 2024–Feb. 2026

Source: YipitData U.S. debit/credit card panel. Starlink adopters: legacy satellite n=1,487, cable n=2,924. Non-adopter controls: legacy satellite n=15,854, cable n=558,679. Data through Aug. 15, 2026. Retention is a card-observed lower bound; results are associations, not causal estimates.

Chart data

Starlink adopters were much less likely to keep legacy satellite than cable
SeriesShare retaining the prior internet provider
Legacy satellite — Starlink adopters31.3%
Legacy satellite — non-adopters67.6%
Cable — Starlink adopters69.4%
Cable — non-adopters86%

How much larger was the prior-provider retention gap for legacy satellite?

Legacy-satellite retention was 36.3 percentage points lower among Starlink adopters than non-adopters, versus 16.6 points lower for cable—a gap 2.2 times as large. The calendar-matched control reduces timing effects but does not prove Starlink caused the difference because adopters may already have been more likely to leave.

Assumptions & Methodologies

We analyzed YipitData’s U.S. debit/credit card panel from Jan. 2021–Aug. 15, 2026. The adopter cohort included 32,760 users whose first observed Starlink charge fell between Jul. 2024 and Feb. 2026. For adopters with a prior-provider payment in the preceding 180 days, retention required at least one payment 30–210 days after the first Starlink charge; the first 30 days were skipped to reduce billing overlap. Legacy satellite covered Viasat and HughesNet, while cable covered four wireline cable operators. The control used customers who never adopted Starlink and the same calendar windows. Fixed-wireless internet was excluded because home and mobile billing could not be separated. Unobserved cards make retention a lower bound, and the control does not eliminate adopter self-selection. Results do not establish causation, permanence beyond seven months, or eventual cable cancellation.

Keep exploring

Ask a follow-up question

Take this research one step further with YipitData.

Did legacy-satellite lines remain canceled beyond seven months after Starlink adoption?

Did Starlink adopters who previously paid for cable cancel cable after a longer follow-up period?

How did prior-provider retention differ by geography or Starlink plan?

Data sources

Credit & Debit PanelTransaction-level spending across the US economy