Every commute, delivery, and business trip burns fuel—and much of the carbon released into the atmosphere is a byproduct of exactly that movement. Reporting in The Innovation, researchers turn large-scale mobility data into a near-real-time carbon meter. Daily intercity mobility networks, built from mobile-phone signaling data covering 318 million users in China, tracked national CO2 emissions in lockstep (r = 0.89) through the COVID-19 lockdown upheaval of early 2020. Combined with population and location data, machine-learning models predicted city-level emissions across 366 cities with R2 up to 1.00, and generalized to Italy, the US, and Mexico with R2 of 0.91–0.99 across two independent emission inventories. Mobility networks also reveal functional city clusters—regions defined by real movement rather than administrative borders—offering new units for coordinated climate governance.