Induced seismicity is one of the main risks for gigaton-scale geological storage of carbon dioxide (CO2). Thus, passive seismic monitoring is often recommended as a necessary component of the monitoring systems for CO2 storage projects, with a particular forcus on risk mitigation. We present the first field study, CO2CRC Otway Project Stage 3 (Victoria, Australia), where distributed acoustic sensing (DAS) enabled high-precision tracking of the induced seismicity triggered by a small CO2 injection and also informed the reservoir models. In 610 days of passive seismic monitoring of the Stage 3 injection, we detected 17 microseismic events (maximum moment magnitude Mw 0.1) using five deep boreholes equipped with enhanced-sensitivity optical fiber. The DAS array has sensitivity sufficient for detection and location of induced events with Mw∼−2 in a monitoring borehole located up to 1500 m away. Thanks to the dense spatial sampling by the DAS, we were able to estimate the focal mechanisms for events with Mw>−1.5; although the monitoring boreholes provided very limited angular coverage. The main cluster of the events has the same location and source mechanism as the one triggered by the previous CO2 injection at the Otway Project site, Stage 2C. Surprizingly, the Stage 2C and Stage 3 events closely followed the actual movement of the CO2 saturation plume front (not the pressure front), as observed using controlled-source reflection seismic images. The nature of the plume-fault interaction remains unclear, but some alteration of the fault gouge by CO2 might be responsible for the faults’ reactivation by the pressure perturbation. Importantly, the seismogenic fault could not be identified in the seismic images and was only revealed by DAS observations, which also demonstrated the signature of fluid–rock interaction, that may control the CO2 flow.