Extra fuel burn while Delta waits for Amazon Leo
Best-estimate dollars from the drag of older Ku/Ka radome antennas versus Starlink’s low-profile aviation terminal — accumulated since the counter window opened, ticking upward every second.
Key assumptions
Leo 2028 vs Starlink now
Delta and Amazon announced Amazon Leo (formerly Project Kuiper) for in-flight connectivity, with initial installation on about 500 aircraft beginning in 2028. United, by contrast, began installing Starlink in 2025 and by late September 2026 had equipped roughly a third of its tracked fleet — hundreds of aircraft already flying a low-profile electronically steered array.
This site does not argue that Leo will be worse connectivity. It isolates one measurable opportunity cost of the later timeline: fuel. Public statements in the 2026 Ryanair–Starlink exchange put legacy high-profile radomes near a ~2% fuel penalty and Starlink’s current aviation terminal near ~0.3% on a 737-class burn. The difference is the drag Delta continues to pay on every hour that could already have been on a Starlink-class antenna under a United-like rollout pace.
The counter starts ~six months before today (default
) and integrates that extra burn × Jet A
price × counterfactual aircraft-hours. Every input lives in
js/config.js so the model can be tightened as better data arrives.