Google and American Airlines Use AI to Reduce Contrails 62%

Google and American Airlines are showing that AI-guided contrail avoidance achieved a 62% reduction in contrail formation rates across a trial of 2,400 transatlantic flights.

Contrails and Why They Matter

Contrails are the thin white streaks formed at cruising altitude when water vapor condenses onto jet exhaust particles and freezes into ice crystals. They are visible in the sky over most major air corridors, but their climate impact is disproportionate to their appearance.

According to research published by Google, contrails account for roughly 35% of aviation’s total contribution to atmospheric warming. Unlike CO2, which disperses across decades, contrail warming is concentrated in the hours following a flight, making it both significant and, in principle, actionable.

The problem has historically resisted intervention because identifying which flights will produce the most warming contrails requires integrating real-time weather data, satellite imagery, atmospheric humidity profiles, and flight-path variables. Manual coordination proved too slow to operationalize at a commercial scale.

How Google’s AI Models Work

Google partnered with Breakthrough Energy to build AI systems that analyze weather and satellite data to map where contrails are likeliest to form before a flight departs. When a high-risk zone is identified, the system recommends a small altitude adjustment to avoid it. Because a relatively small share of flights accounts for most of the contrail warming, Google estimates that rerouting about 15% of departures would be enough to yield a significant climate benefit across an airline’s entire operation.

In an early test covering roughly 70 flights, American Airlines pilots who followed the AI recommendations cut contrail formation by 54%. A larger trial, in which Google’s forecasts were built directly into American Airlines’ existing flight-planning software, covered 2,400 transatlantic flights and achieved a 62% reduction in contrail formation compared to flights that made no adjustments, according to Google.

The cost is real but small. Planes that change altitude to avoid contrail-prone air burn slightly more fuel in the process. Across American’s full fleet, Google calculates that fuel use would rise by 0.3%. The company’s models put the climate return on that investment at 20 times the warming caused by the additional fuel burned.

That math sets contrail avoidance apart from the other tools airlines have for cutting emissions. Sustainable aviation fuel is expensive and in short supply. New, more efficient planes take years to procure and deploy. Contrail avoidance asks for none of that. It works with the aircraft already flying and the software already in use.

Alaska Airlines and AI-Driven Operations

Alaska Airlines has pursued a parallel track, deploying AI not for contrail-specific avoidance but for systemic route optimization that reduces fuel burn across its network. Through a renewed partnership with Air Space Intelligence, Alaska uses the Flyways AI Platform, which ingests weather patterns, wind conditions, turbulence forecasts, airspace constraints, and air traffic volume to generate real-time optimized routing recommendations for dispatchers and pilots.

Over four years of deployment, the platform has identified optimization opportunities on 55% of Alaska’s flights. For flights longer than four hours, Flyways has delivered fuel savings and emissions reductions of 3% to 5%.

In 2023 alone, the optimized routes saved more than 1.2 million gallons of fuel, reducing emissions by approximately 11,958 metric tons of CO2, according to Alaska Airlines. The airline has set a near-term target to become the most fuel-efficient U.S. carrier and a long-term goal of achieving net-zero emissions by 2040.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.