Cathay flew more than 80 flights on Google contrail forecasts sent over in-flight Wi-Fi

Most flights never need this. Research collated by Resources for the Future finds that 2% to 3% of flights produce around 80% of contrail warming ^[6]. On North Atlantic routes, roughly 38% of flights made warming contrails and 14% made contrails that were cooling on balance, with the other 48% leaving nothing that persisted ^[7]. So 62% of that sample either did no harm or helped ^[8]. The 2% to 3% figure and the 38% figure answer different questions rather than fighting each other: a bit over a third of flights make a warming contrail at all, and a much thinner slice inside that group does nearly all of the damage.

That thinness is why the Cathay number should be read as a route result. One regional corridor carried most of the estimated benefit, while the trial was sold as the first test of the technique on services over 16 hours ^[3]. The account does not reconcile those two things, and until somebody does, the headline reduction describes a particular set of routes in a particular season rather than an efficacy rate an operations team can apply across a schedule.

The fuel objection has two answers on the record and they do not match. A 2023 study across 70 flights with American Airlines cut contrails by 54% at about 2% extra fuel on the rerouted flights, which the researchers argued would fall to as little as 0.3% spread across a whole airline schedule ^[9]. A larger transatlantic trial of 112 flights between January and May 2025 reported a 62% reduction against a control group and found no statistically significant difference in fuel burn ^[10]. Google's own calculation puts warming avoided at roughly 20 times the warming added by the extra fuel ^[11]. Note that its abatement range tops out at five times its floor ^[12], which is the honest measure of how precisely this can be priced. Add the three trials together and the entire operational record here comes to about 262 flights ^[13].

The European Commission puts non-CO2 effects at two thirds of aviation's climate impact and contrails alone at about a third ^[14], which makes contrails roughly half of the non-CO2 problem ^[15]. Of the £5m going into Operation Blue Skies, £2.65m comes from the Department for Transport through the Aerospace Technology Institute and £1.4m in kind from Google UK, with the Met Office, NATS, Imperial College London and the University of Cambridge in the consortium and 20 to 40 test days planned per winter season ^[16]. That leaves about £950,000 of the headline figure unattributed in the announcement ^[17].

Cathay's director of digital and IT, Lawrence Fong, said in a statement that aviation needs solutions to address climate change and that AI is accelerating that progress ^[18]. The forecasting is where the AI sits, and the deployment is a file-delivery problem the airline owns rather than the vendor. Two conditions sort operators here. The first is whether a document already reaches the flight deck in cruise. The second is whether the network crosses regions where persistent contrails form often. With both, this runs on existing plumbing at a fuel penalty inside normal variance. With the channel but not the corridors, crew attention goes to flights that were never going to matter. With the corridors but not the channel, connectivity is the project and the forecast is the easy part. On this evidence the corridor selects the airline rather than the reverse ^[6], and a carrier whose network misses the humid bands has nothing here to buy at any forecast quality.