Maxwell Brodie vividly recalls the destructive wildfire he experienced as a kid growing up in the interior of British Columbia.
One night in 2003, lightning struck a tree at around 4 a.m., sparking a massive blaze that scorched Okanagan Mountain Park. Winds picked up, the skies turned orange and more than 30,000 people evacuated from his hometown. Brodie remembers helping his dad attach a soaker hose to protect their cedar roof from falling ash.
The experience would inspire Brodie nearly two decades later to launch a software startup that gives autonomous helicopters and other aircraft the capability to perceive and suppress wildfires.
“That is just something that, as a child, you don’t forget,” said Brodie, co-founder and chief executive of Alameda-based business Rain. “As we experience these more frequent and severe fires, expanding response capacity to include being able to respond at night in smoky conditions, and in high winds, becomes more important.”
Brodie is among a small but growing cadre of entrepreneurs in California promising new technology — much of it powered by artificial intelligence — that could dramatically change how firefighters prevent and fight wildfires.
Confronting budget shortfalls, fire departments have traditionally been cautious about embracing costly and often experimental firefighting technology that hasn’t been proven in the field. But the magnitude of the unprecedented L.A. fires that destroyed thousands of structures and killed at least 27 people has brought new interest and urgency to finding more effective ways to combat wildfires.
“It’s just a completely different scale…We’re gonna have to come up with new ways to fight [fires],” said Josh Wilkins, a retired San Bernardino County Fire Department fire captain.
In Silicon Valley, major tech companies including Google and AI-giant Nvidia have been investing in research that could help firefighters better detect and track wildfires.
Nvidia announced it teamed up with Lockheed Martin in 2021 and the U.S. Department of Agriculture Forest Service and Colorado’s Division of Fire Prevention and Control to create a digital version of a fire that allows firefighters and incident commanders to better understand how a fire spreads and suggest more informed ways to suppress it.
“The 21st century security technologies that we’re developing to respond to security threats are directly applicable to the complex environment of a wildland fire,” said Dan Lordan, senior program manager at Lockheed Martin Artificial Intelligence Center in Connecticut.
AI-enabled decision aids may soon be able to support first responder command decisions but are dependent on the availability of data and how close it is to real time, Lordan said.
Space agency NASA also is working on technology that could make it possible for drones and remotely piloted helicopters to fly at the same time to address wildfires even when there’s low visibility.
Fire departments across the state already use an AI tool, run by UC San Diego, that can detect fires in video footage so they can respond quickly to flames. Known as ALERTCalifornia, the program deploys more than 1,144 cameras and sensor arrays that capture live video around the clock.
The California Department of Forestry and Fire Protection worked with ALERTCalifornia and DigitalPath to develop the AI tool.
“It creates a network that watches over California,” said Cal Fire Battalion Chief David Acuña.
There have been some successes. Last month , for example, ALERTCalifornia’s AI system detected a fire in Black Star Canyon and alerted the Orange County Fire Authority at 2 a.m. Firefighters doused the fire and contained it to less than a quarter-acre.
Nonetheless, while ALERTCalifornia has helped save lives, its limitations were also exposed during the L.A. fires, in which powerful winds fueled flames that spread so fast that firefighters couldn’t keep up.
To improve its capabilities, Cal Fire is testing new equipment with BurnBot, a South San Francisco company that operates large vehicles that can do controlled burns with little or no smoke. The state-of-the-art vehicles, called RX, are equipped with propane torches that allow operators to control the length and temperature of flames. They also have water spray nozzles and a heavy roller to extinguish flames.
Wilkins, who advises BurnBot and other wildfire prevention startups, believes the vehicles could have slowed the spread of the L.A. fires if they had been deployed.
“Once we get to wind-driven fires, you’re fighting embers,” Wilkins said. “It’s basically millions and millions of matches flying through the air and one big bush on fire can transmit thousands of embers, and each one of those embers has the potential of igniting anything it lands on.”
Acuña said the agency is still evaluating BurnBot’s vehicles and awaiting data to help determine how or whether they will be used.
One obstacle to the more widespread use of these futuristic firefighting tools, including sensors that can detect smoke and fire, is a dearth of private and public sector funding.
“It’s been a wake-up call to all of us of what we’re up against,” said Sonia Kastner, the co-founder and chief executive of San Francisco-based Pano AI. “We need a radical shift in how we approach firefighting and natural resource management.”
Kastner knows the challenges firsthand. She started Pano AI, which built an AI-powered platform to detect fires and alert emergency responders, after the 2018 Camp fire that left 85 people dead, burned 153,336 acres and caused an estimated $16.5 billion in losses.
Pano AI relies on cameras, placed on high vantage points like cell towers, to scan the surrounding area and relay video images to emergency personnel. They have been used in Ukiah and Rancho Palos Verdes in California and in other states.
The Department of Homeland Security operates a technology center within its Washington-based Science and Technology Directorate that has supported the development of sensors to detect fire and toxic chemicals.
S&T and N5 Sensors conducted a controlled burn in Stafford, Va., ahead of the 2023 wildfire season. Data collected were used to enhance the sensors and their detection capabilities.
(N5 Sensors)
About 450 so-called Alpha and Beta sensors, which can cost a few thousand dollars each, have been deployed to areas including Orange County, Bay Area cities and have helped to detect fires in Hawaii, Colorado and Oakland, Calif.
To support the initiative, Homeland Security received $4 million in funding over four years, but the agency hasn’t been able to secure more federal money, said Jeff Booth, director of the Sensors and Platforms Technology Center for the department’s Science and Technology Directorate.
“I have no further federal funding to take this a step further,” Booth said. “Maybe with the new administration, they could see the value of deploying this even further.”
Teams prepare for a SuperVolo XL flight at the Monterey Bay Academy Airport near Watsonville, Calif.
(Don Richey / NASA Ames)
For startups like Rain, getting buy-in from investors and fire departments is key.
Founded in 2019, Rain operates out of an old traffic control tower in the former Naval Air Station Alameda. The company, which has 15 employees, raised $9.7 million in seed financing led by venture capital firm DBL Partners.
Rain has worked with Lockheed Martin company Sikorsky and with fire officials in Orange County in the hopes of bringing its technology into operational use.
“When there’s that partnership between the innovators in the fire community and technologists, that’s what opens up entirely new tools, technologies and markets,” chief executive Brodie said.
