Palomar Mountain is getting a new job as San Diego County's high-altitude lookout, with San Diego Gas & Electric teaming up with Qualcomm and UC San Diego's Scripps Institution of Oceanography on an edge AI pilot meant to spot wildfire and extreme-weather trouble as it starts, not after the fact. The project, called Edge Alert Sentinel, will post rugged hardware on the mountain and run models right where the weather is happening so operators can get near-real-time alerts. The idea is to shave precious minutes off the gap between sudden shifts in wind or flash flooding and a response from utility crews or emergency officials.
How the technology works
In a press release via San Diego Gas & Electric, the partners describe Edge Alert Sentinel as a ruggedized edge AI gateway powered by Qualcomm's Dragonwing IQ9 processor and an MLOps platform from Edge Impulse. The on-device neural processor is designed to perform up to 100 trillion operations per second, running models that look at immediate conditions and flag emerging risks right at the collection point. Predictive alerts and localized telemetry will travel over SDG&E's private cellular network to the utility's control center so operators can make faster calls on the grid and public safety.
Pilot timeline and scale
According to Sempra, the first Edge Alert Sentinel unit is being installed on Mt. Palomar and will be put through its paces during the upcoming Public Safety Power Shutoff season. The companies plan to use what they learn there to expand to other high-risk sites starting next year, with a broader regional rollout eyed for 2027. Over time, the collaboration also envisions applying the same kind of on-device AI to automated inspections and autonomous aerial work on grid infrastructure.
Built on long-running local data and upgrades
SDG&E points to years of preparation behind the pilot. The utility's website notes it has invested nearly $6 billion in wildfire mitigation efforts since 2007, and SDG&E highlights long-running partnerships with UC San Diego and research networks that feed atmospheric observations into models. Local reporting says Edge Alert Sentinel will tap into UC San Diego's existing web of cameras and stations. The San Diego Union-Tribune reports that network includes roughly 130 cameras and about 225 automated weather stations that provide regional situational awareness.
Why this matters for San Diego
The push for on-site analytics is meant to cut the lag time that comes with sending everything to the cloud, especially when Santa Ana winds, drought conditions and sudden storms can flip fire risk in a matter of minutes. The state Office of Energy Infrastructure Safety recently signed off on SDG&E's 2026-2028 Wildfire Mitigation Base Plan, which gives regulatory backing to the utility's expanded monitoring and hardening work. At the federal level, the White House's June 2, 2026 executive order on advanced AI directs agencies to prioritize cyber defenses and set up an AI cybersecurity clearinghouse, a move that will influence how critical-infrastructure operators secure deployments like these at the edge of the grid.
What to watch next
Project success will hinge on how fast and reliably the system can flag danger while keeping false alarms in check and staying online when connectivity is stressed. Nakul Duggal of Qualcomm said the effort aims to deliver faster, more reliable insights where conditions are changing, according to the companies' release, and SDG&E says it will use results from the Palomar pilot to tune its models, training practices and handoffs with emergency responders. For San Diego residents, the key question is whether those hyperlocal alerts actually shrink response times during the most dangerous weather and PSPS events.
