Estimated extra fuel cost Estimate
Average · most likely
—
Hourly cost —
Daily cost —
Cumulative extra fuel burned
Gallons —
Pounds —
Extra CO₂ from that fuel burn
—
—
Cumulative CO₂, in everyday terms
  • —passenger cars driven for a year
  • —U.S. homes' electricity use for a year
  • —round-trip New York–London passenger flights
Since 2026-01-01 Updated —

Cost case: Average · See settings below

Fleet est. —

Extra jet fuel from keeping older satcom antennas while Delta waits for Amazon Leo, vs. a Starlink rollout that could have started earlier. Model estimate.

Projected cumulative cost by Leo rollout complete
—
Plateau when Amazon Leo finishes replacing older antennas
    • Accrued so far
    • Projected under rollout
    • Today
    • Missed rollout complete
    • Leo start
    • Leo rollout complete
    • Plateau after complete

    Extra fuel cost while Delta waits for Amazon Leo

    Older satcom antennas burn extra jet fuel versus a Starlink-class terminal. Cost rises as a missed Starlink rollout would have grown, then plateaus when Leo (~2028) finishes replacing those antennas.

    What drives the estimate

    Mainline + Delta Connection (on by default), drag, and Jet A set the burn rate. The cumulative total also depends on when Starlink installs could have started, how fast that missed rollout would have grown, and when Leo starts / how long it takes.

    Known = public filing or index · Assumed = model choice. Best / Average / Worst = lowest / most likely / highest cumulative cost presets; Advanced can exclude Connection.

    Choose a cost range (lowest / most likely / highest)

    Missed Starlink path = installs Delta could have started earlier (like United). Leo path = Amazon’s ~2028 rollout that replaces older antennas.

    Advanced — tap to tune sliders
    Projected by Leo complete: — · Case: —

    Toggle Delta Connection (~325 RJ from the 10-Q) on or off, then tune every formula input — including RJ baseline burn (gal/hr), when installs could have started, how fast that rollout would have grown, and Leo start / duration. Best / Average / Worst snap to lowest / most-likely / highest cost. Moving a control switches to Custom.

    Leo 2028 vs Starlink now

    Delta has none on Starlink today; United’s tracker count is about 607 of 1,669.

    Delta and Amazon announced Amazon Leo (formerly Project Kuiper) for in-flight connectivity, with initial installation on about 500 aircraft beginning in 2028. Delta’s mainline fleet is on the order of ~1,000 aircraft (SEC Form 10-Q reports 1,004 as of mid-2026), plus about 325 Delta Connection regional jets in the same filing. The model includes Connection by default (Advanced can exclude it). Delta has none on Starlink today — current connectivity is Viasat / Hughes.

    United, by contrast, began installing Starlink in 2025. United’s own June 2026 release put a hard floor at more than 400 Starlink-equipped planes; an independent tracker counted about 607 of 1,669 (~36%) as of late September 2026 — 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 waiting longer for Leo: fuel — and the CO₂ from burning that extra fuel. Public statements in the January 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 ~800 gal/hr 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.

    Framing for shareholders: Ed Bastian’s Leo bet trades near-term antenna-drag savings (and Starlink’s proven airline install cadence) for a 2028 partner stack. This page tracks the fuel side of that trade — and the emissions that come with it — empirically, labeled as estimates, with Best / Average / Worst as lowest / most likely / highest cost.

    Is anything legally stopping Delta from installing Starlink?

    No hard legal ban is apparent from public FAA records.

    No hard legal ban is apparent from public FAA records. Installing a new satcom antenna on a U.S.-registered airliner requires an FAA Supplemental Type Certificate (STC) (or equivalent approved data) for that aircraft type — a certification pathway, not a prohibition. SpaceX / Starlink Aviation already holds current FAA STCs covering installation of Starlink aviation terminals on multiple commercial types that overlap Delta’s fleet mix, including Boeing 737-800 / 737-900 / 737-900ER (e.g. ST09945AC), Boeing 777 variants, and various Airbus A320 / A321 / A330 families (additional STCs on public FAA DRS listings). An STC authorizes the modification; it does not by itself mean Delta has contracted for kits, data rights, or installs.

    What keeps Starlink off Delta today looks commercial and contractual — Delta’s Amazon Leo partnership (~500 aircraft beginning ~2028), existing Viasat / Hughes connectivity, and airline choice — rather than a statute or FAA rule forbidding Starlink. Whether Delta could obtain install authority for every exact fleet subtype, serial range, and configuration is an engineering / certification detail beyond this page. This is factual context for shareholders, not legal advice.

    Model note: averages for fleet size, flight hours, and Jet A for now. The estimate will get more complex as Delta adds aircraft and Jet A moves; a later version can track fleet growth and a Jet A path explicitly.