San Diego's New Sentinel: Can AI at the Edge Tame Wildfire Risk?

SAN DIEGO, CA – June 08, 2026 – In a region defined by its complex relationship with fire and weather, a landmark collaboration is taking shape. San Diego Gas & Electric (SDG&E), semiconductor giant Qualcomm Technologies, Inc., and UC San Diego's renowned Scripps Institution of Oceanography have unveiled Edge Alert Sentinel (EAS), a project that moves artificial intelligence from the cloud to the cliffside. By placing high-powered AI processing directly at the point of risk, this initiative aims to revolutionize how we predict and respond to wildfires and extreme weather, making decisions in seconds, not minutes.

The first deployment is being installed on the high slopes of Mt. Palomar, a critical vantage point for monitoring the volatile conditions that fuel Southern California's disasters. This is not just another sensor network; it's a fundamental shift in environmental intelligence, designed to build resilience into the very fabric of the region's infrastructure.

"For nearly two decades, our region has avoided a catastrophic electrically caused wildfire because we chose to lead early and never stop looking ahead," said Scott Crider, President of SDG&E. "Edge Alert Sentinel reflects that same mindset. By working with Qualcomm Technologies and UC San Diego, we're bringing world-class technology and science together, so intelligence lives where the risk lives — on the front lines — and communities are safer because of it."

From Cloud to Cliffside: A New Paradigm in Intelligence

For years, the paradigm for large-scale data analysis has been the cloud. Data is collected by sensors in the field and transmitted to remote data centers for processing. But in an emergency where seconds count—like a nascent wildfire fanned by Santa Ana winds—that round trip introduces potentially catastrophic latency. Furthermore, the very emergencies that demand a response can cripple the connectivity needed to get data to the cloud in the first place.

Edge Alert Sentinel flips this model on its head. The system processes data locally, or "at the edge," enabling near-instant analysis. At its core is a ruggedized edge AI gateway powered by the Qualcomm Dragonwing™ IQ9 processor. This is not a standard computer chip; it's an industrial-grade workhorse designed to operate in harsh environments, from -40°C to over 100°C, while delivering up to 100 trillion operations per second (TOPS). This immense computational power allows the system to run complex AI models directly on-site.

"Through this collaboration, we're intending to bring real-time intelligence directly to the front lines of wildfire response," noted Nakul Duggal, a Qualcomm Technologies executive vice president. "By combining on-site AI with advanced sensing and connectivity, we're helping deliver faster, more reliable insights where conditions are changing — so responders can assess risk and act with greater speed and confidence."

Deploying and maintaining such sophisticated technology in remote, rugged terrain presents significant challenges, from ensuring a stable power supply to managing software updates with limited connectivity. However, the potential payoff—resilient, real-time intelligence that works even when networks are down—is a prize worth pursuing for utilities operating in high-risk zones.

Forged in Fire: An Answer to a Decades-Old Threat

The urgency behind this innovation is etched into San Diego's recent history. The devastating 2007 wildfires, linked to SDG&E's equipment, served as a brutal catalyst for change. In the years since, the utility has invested nearly $6 billion into becoming a leader in wildfire mitigation, building the nation's largest utility-owned weather network, hardening its grid with covered conductors and undergrounded lines, and deploying drones for inspections.

One of its most visible, and controversial, tools has been the Public Safety Power Shutoff (PSPS). While effective as a last-resort measure to prevent ignitions during extreme weather, these shutoffs cause profound disruption for communities. A key promise of EAS is its potential to refine this blunt instrument. By providing a more granular and immediate understanding of on-the-ground conditions, the system could help SDG&E make more targeted decisions, potentially reducing the scope and duration of PSPS events and restoring power more quickly once the threat has passed.

EAS represents the next logical evolution in this two-decade-long journey, moving from broad-based infrastructure hardening to hyper-local, intelligent risk assessment. It's a testament to the idea that true resilience is not just about building stronger poles and wires, but about creating a smarter, more responsive system.

The Power of Three: Uniting Industry, Science, and Operations

The strength of Edge Alert Sentinel lies in its unique trifecta of collaborators, each providing an indispensable piece of the puzzle.

SDG&E brings the operational expertise, the physical infrastructure, and the urgent, real-world problem that needs solving. Qualcomm provides the advanced AI processing and edge-computing architecture that serves as the project's technological engine. But the secret ingredient may be the contribution from the Scripps Institution of Oceanography.

For decades, Scripps has been meticulously observing the region's environment. Through networks like the High Performance Wireless Research and Education Network (HPWREN) and its successor, ALERTCalifornia, Scripps has amassed an unparalleled dataset on Southern California's atmospheric conditions. This rich historical data is the lifeblood for training the AI models, teaching them to recognize the subtle patterns that precede a major weather event or signal a heightened fire risk.

"With this new onsite AI capability, we're moving beyond observation to predicting impact in real time — at the exact moment and place where danger emerges," explained Frank Vernon, director of HPWREN. This collaboration allows Scripps to transform its deep scientific understanding and observational record into an actionable tool for community safety.

A Blueprint for Resilience Beyond California

While born from Southern California's specific challenges with wildfire, the EAS model is designed for scalability. The core concept—deploying resilient, low-latency AI to monitor and predict environmental threats—is applicable to a host of climate-driven risks facing communities worldwide. The same architecture could be adapted to provide early warnings for floods by monitoring water levels in real time, or for extreme storms by analyzing wind and pressure data on-site.

The project's pilot phase on Mt. Palomar will be closely watched during the upcoming fire season. Insights gathered will inform an expansion to additional sites, with a wider rollout targeted for 2027. This methodical approach, backed by Sempra's significant capital investment plans and a supportive regulatory environment in California that encourages innovation, provides a potential blueprint for how public-private partnerships can accelerate the adoption of life-saving climate technologies. The initiative demonstrates a powerful strategy for building permanence in an age of performance, proving that the most durable defense is one that is intelligent, agile, and always learning.