The First Decade of Breakthrough Listen

Abstract

Breakthrough Listen (BL), launched in July 2015 with a 10-year, \$100M commitment from the Breakthrough Prize Foundation, is the most comprehensive search for technosignatures ever conducted. This paper reviews the program's first decade from the facilities and instrumentation perspective: the digital backends and data recorders, the signal-processing pipelines, the data archive, and the observatory partnerships that made the science possible. We organize the decade into three architectural eras. In Generation 1 (2016--2019), BL deployed packetized raw-voltage wideband recorders on the 100-m Green Bank Telescope and the 64-m Parkes telescope, enabling targeted surveys of thousands of nearby stars across $\sim$0.7--8 GHz. In a network-expansion phase (2020--2023), BL extended its instrumentation to partner facilities including the Sardinia Radio Telescope and international LOFAR stations, and developed verification frameworks that proved essential when the program's highest-profile signal candidates emerged. Since 2023, BL has operated in a fully commensal era: Ethernet-based, commodity-GPU backends riding on routine observatory operations: COSMIC at the Karl G. Jansky VLA, a revived Allen Telescope Array with deployed real-time deep-learning pipelines, and an autonomous multi-beam commensal survey at MeerKAT (BLUSE; $>$1.2 million pointings since mid-2022). Over the decade, capturing data at ever higher rates forced a systematic approach to compression and archiving; the storage estate grew to $\sim$42 PB of capacity ($\sim$31 PB in use) across seven systems on four continents by March 2026. We conclude with engineering lessons (commodity hardware over bespoke hardware, open data as a first-class requirement, commensal science as backend validation, and verification frameworks built before they are needed) and with the outlook for SKA-era commensal technosignature science.