Google AI trial will reroute Atlantic jets to cut contrails
Google and the UK government just moved contrail avoidance from airline-by-airline experiments to the people who actually own the sky.
Google UK, NATS, and a UK-backed consortium launched Operation Blue Skies on Tuesday — a £5 million, 30-month trial that will use AI forecasts to nudge commercial jets around the cold, humid pockets of air that turn exhaust into lingering, heat-trapping contrails. The programme is billed as the first state-backed test at the scale of an entire oceanic airspace. It will run in Shanwick, the eastern half of the North Atlantic corridor that NATS controls, a patch of sky that atmospheric modeling credits with about 5 percent of global contrail warming. Two winter windows, 2026–2027 and 2027–2028, will carry the operational trials. Contrails.org, which is designing the trial parameters, says manoeuvres will land on roughly 20 to 40 quieter test days each winter, mostly as small altitude changes passed through ordinary air-traffic-control channels.
The climate claim is the reason anyone is bothering. Google says warming contrails — the white lines that bloom when water vapor condenses on soot in jet exhaust — account for roughly a third of aviation's total climate impact, and that they are one of the cheaper pieces of that impact to dodge. Earlier work with American Airlines, EUROCONTROL's Maastricht centre, and flight-planning firms already showed crews can make targeted deviations without leaving normal operations. What those trials could not answer is what happens when many aircraft in the same corridor move at once, and whether the forecast should live in an airline's flight plan or in the controller's display. Blue Skies is built to force that question. Google technical lead Paul Hodgson told SiliconANGLE that about 10,000 flights will pass through the corridor in the test windows, and that only 1 to 5 percent of them will be redirected. That is a small sample on purpose: enough, the consortium hopes, to see whether avoiding the worst pockets actually thins the cloud layer that satellites can measure.
The money trail is as political as it is technical. The UK Department for Transport is putting up £2.65 million through the Aerospace Technology Institute programme. Google UK is in on a pro-bono basis, contributing £1.4 million in kind — research time, engineering, and compute — while 100 percent of the government grant is reserved for academic, non-profit, and aviation partners. The Met Office will build a UK contrail forecast. Imperial College London and the University of Cambridge will independently score the climate result. Aviation minister Keir Mather framed it as British experts leading a world-first, sitting beside a much larger £219 million bet on homegrown sustainable fuel. That is the real product: not a new weather model, but a protocol that air-navigation services can adopt without waiting for every airline to volunteer.
The honest limit is scale. A few dozen winter days and a few hundred diverted flights will not clean the Atlantic. They will tell operators whether the forecast is good enough to trust when the corridor is busy, and whether altitude tweaks that look free on a slide deck stay free when traffic, fuel, and safety margins collide. If Imperial and Cambridge can see a signal from orbit, this becomes a blueprint. If they cannot, the industry has spent £5 million to learn that the easy climate win was never that easy.
What to watch: whether the first winter window actually moves flights, or whether safety assessments keep the trial on the ground.
If the forecast is good enough for a controller to act on, should airlines still get to opt out? Tell us in the comments.
