[Adobe Stock]
In a study released March 19, Google said its contrail forecasts were integrated directly into American’s flight planning software and tested across 2,400 transatlantic flights between January and May 2025. Half were given an option to avoid contrail-prone air; the rest served as controls. Among the 112 flights that actually flew the avoidance plan, contrail formation fell 62% relative to controls, with an estimated 69% reduction in climatological warming. The associated paper also reported an 11.6% reduction in contrail formation across the full group of flights marked eligible for avoidance, with no statistically significant difference in fuel use between groups.
But only 112 of the roughly 1,200 flights offered the avoidance option actually flew it. That circa 9% gap between modeled opportunity and what airline operations could actually absorb is arguably the most important number in the study.
Why contrails matter, and why the science is unsettled
Contrails are widely viewed as a major aviation warming factor, but the size of that effect — and the trade-offs involved in mitigating it — remain under active debate. A 2022 IPCC report noted that clouds created by contrails account for roughly 35% of aviation’s global warming impact, over half the impact of the world’s jet fuel. According to Contrails.org, a nonprofit under Breakthrough Energy, contrails are responsible for an estimated 1–2% of total human-caused climate change.
A 2025 Nature Communications paper led by atmospheric scientist Andreas Petzold found that more than 80% of long-lived contrails form within pre-existing cirrus clouds, where their added radiative effect may be weaker and, in some conditions, could tilt toward cooling rather than additional warming. And David S. Lee, author of the most influential 2021 assessment of aviation’s climate impact and chair of the IPCC’s aviation working group, has argued not that contrails don’t matter, but that aviation’s non-CO₂ mitigation still carries enough uncertainty and trade-offs that it should not be treated as simple low-hanging fruit.
The chosen metric and time horizon can dramatically shift the comparison — some frameworks put contrail warming at multiples of aviation CO₂, while others reduce it to a fraction — which is why the debate over whether to spend fuel avoiding contrails remains unresolved.
From research demo to operational tool
The project builds on a smaller 2023 effort by Google, American and Breakthrough Energy that showed a 54% reduction across 70 test flights. That earlier phase required hours of manual coordination to identify candidate flights. What changed this time was operational integration: instead of manual selection, contrail forecasts were embedded in American’s standard planning workflow alongside Flightkeys, the airline’s flight planning software provider.
That shift moves the question from whether contrails can be avoided in principle to whether avoidance can be folded into production systems at scale. The AI analyzes weather and satellite data to generate forecast maps identifying atmospheric regions where contrails are most likely to form. Dispatchers can then suggest minor altitude adjustments, often just a few thousand feet, before departure. Because a relatively small share of flights accounts for the majority of contrail warming — one recent study found roughly 2–3% of flights caused 80% of annual contrail energy forcing — Google estimates the fleet-wide fuel cost of avoidance could be as low as 0.3%. The company’s models put the climate return on that investment at 20 times the warming caused by the additional fuel burned.
Wider industry momentum, and new partnerships
A visit to a Google Research event last year revealed that the company’s contrail ambitions extend well beyond a single American Airlines partnership. A member of Google’s climate AI team disclosed that the company is working with additional airline partners to build global demonstrations of contrail avoidance, deliberately seeking carriers across different regions to test the approach under varied operational and atmospheric conditions. Those partnerships have not yet been publicly announced.
The broader industry is moving in the same direction. Eurocontrol’s Maastricht Upper Area Control Centre conducted the first operational contrail prevention trial in 2021 and has worked with Google since 2022. The Contrail Impact Task Force, convened by RMI and Breakthrough Energy, now includes Alaska, American, Southwest, United, and Virgin Atlantic alongside Airbus, Boeing, and Google Research. In Europe, French carrier Amelia and Thales recently reported that their own contrail-avoidance deployment avoided more than 2,000 tonnes of CO₂ equivalent in 2025 by modifying just 59 flights out of more than 6,400 operated.
Meanwhile, Alaska Airlines has pursued a parallel track: AI-driven route optimization that has delivered 3–5% fuel savings on long-haul flights and saved more than 1.2 million gallons of fuel in 2023 alone through its partnership with Air Space Intelligence.
The scalability question
The 112-of-1,200 compliance rate is the number that should focus the industry’s attention. Thomas Walker of the Clean Air Task Force told the Associated Press that international coordination remains a challenge, and that CATF has encountered “a little bit of pushback” in conversations with other major airlines. Air traffic control constraints, routing limitations, and operational practicality all limit what pilots can actually do with a contrail forecast, however accurate the AI behind it may be.
American Airlines VP of Sustainability Jill Blickstein said the trial showed it was not difficult for dispatchers and pilots to file alternative plans, but the airline is not yet making contrail avoidance a routine part of its regular flight planning process. Google’s Dinesh Sanekommu said the next step is a larger trial and integration with additional flight planning software providers.
