“More than half of our northern and central California service areas are within high fire threat areas, and a third of our assets are located in those areas,” PG&E spokesperson Paul Doherty told me. While PG&E’s service area doesn’t overlap with the L.A. fires, the growing list of gridtech and climate tech companies that it’s partnered with could serve as an example for other utilities in the state and country as a whole. In PG&E’s catalogue are vegetation management robots, power pole sensors, advanced fire detection cameras, and autonomous drones, with much of this enhanced by an artificial intelligence-powered analytics platforms.
In some ways, the 120-year-old utility is starting to act like a tech incubator. It hosted its first-ever innovation summit in 2023, where Doherty said it held a Shark Tank-style pitch fest to source ideas for a variety of grid challenges, including wildfire-related ones like system monitoring and vegetation management, ultimately receiving over 600 applications. Out of that, PG&E chose 24 concepts to move forward with in some form.
“My experience has been that they’re very focused on reducing risk,” Dave Winnacker, co-founder of the AI-powered risk visualization and mitigation platform XyloPlan, told me. “That attention is probably focused by the fact that they were held accountable and they had significant monetary losses, reputational losses.”
Last year, XyloPlan partnered with PG&E to pilot its software in the wildfire-prone Lake County, California. The platform provides insight into the areas most at risk from fast-moving fires, which Winnacker told me are much more damaging to communities and critical infrastructure than hot fires, known to be more destructive in forests. “So in our model and our future state, you can still have plenty of fire on the landscape, and you can even have plenty of fast-moving fire, but we have prioritized treatments that would disrupt those fast-moving fires that have the greatest consequences,” Winnacker, the former fire chief of the Moraga-Orinda Fire District, told me. XyloPlan’s algorithm makes recommendations on where various resiliency efforts such as vegetation management would have the greatest impact.
Winnacker acknowledges though that for utilities, “it’s really difficult and risky to take something new on.” Not only could money be wasted if it doesn’t work out, but as Winnacker told me, “It can be perceived as an admission of your doing things wrong before. The tendency to assign blame makes it harder to adopt new and innovative things.”
“I think the toughest thing for a utility is to trust a technology,” Christina Park, senior director of energy strategy at the autonomous drone company Skydio, told me. A former veteran of the utility industry herself, Park spent 15 years at the New York Power Authority and understands why utilities would be reluctant to tweak at least formerly reliable services and infrastructure that millions of households depend upon. But as climate change brings drought and more extreme weather, and as utility infrastructure ages, evolution seems like the only option. “Based on all the confluence of factors that are kind of putting their backs against the wall, they are more open to change,” Park told me. “It’s just not possible to keep doing things the old way.”
Skydio, which was last valued at $2.2 billion after its 2023 Series E funding round, operates in three main markets — defense, public safety, and utilities. PG&E has been a customer of the company since 2022, and became the first California utility to conduct fully remote drone inspections of its assets in 2023. This was made possible after the utility secured a much-coveted waiver from the Federal Aviation Administration that allows it to fly drones beyond the visual line of sight.
“An operator could fly a drone to a location that’s up over a mountain, right up over super steep, rugged terrain that would normally be really hard to access via helicopter, via foot, via vehicle, and now we have the capability to go inspect that,” Doherty told me. Six navigation cameras as well as onboard artificial intelligence and advanced computing allow Skydio drones to operate autonomously, docked and deployed at PG&E substations.
Park told me that PG&E, which has had a drone program since 2019, has used its aviation expertise to help Skydio develop key capabilities. “They have the knowledge in the drone space to really ask for more advanced features — being able to pick out when there is a zoom quality that they would really like to see or a certain lens.” After Skydio’s drones gather reams of visual data, algorithms can pinpoint the location and severity of any infrastructural defects. PG&E has developed its own A.I. model in house to do this.
PG&E is far from alone in its excitement over Skydio’s capabilities. The dronemaker has over 200 utility partnerships to date, and Park told me that across all of them she’s seeing more and more integration of new tech into the standard workflow. “Their business as usual, it just looks different than it did five years ago,” she told me. But while there might be an increased appetite in the industry for novel solutions, Winnacker warns that there are numerous logistical and financial barriers that can get in the way of promising tech moving from pilot to full-scale implementation.
“The challenge on these things always is that the benefit is very widespread, but there has to be someone who is the lead, and ultimately someone has to make the investment,” Winnacker told me. “That’s challenging, because there is a federal component, there’s a state component, there’s a local government component, there’s a non-government, land-owning agency component, and then there’s a small private property component. We have to mesh all of these.”
Sometimes, good companies with good ideas can languish as these various stakeholders with different perspectives and priorities wait for someone else to step up and foot the bill. As of now, Winnacker said he doesn’t know if PG&E is going to make a more significant investment in XyloPlan, although he said last year’s partnership proved fruitful.
But if PG&E does move forward with XyloPlan, or any other gridtech or wildfire mitigation tech for that matter, the success of that program will depend not just on the utility, but also on all the other governmental and non-governmental players that Winnacker mentioned. “There’s a need for really tight alignment, so that the work of one group compliments the other, and we don’t end up in this disjointed manner, where a lot of effort is occurring, but because it’s not coordinated, it’s not aligned, you don’t get that the reinforcing benefit of the network,” Winnacker told me.
Not to mention the fact that in rural and urban areas alike, there’s always competing demands and only so much money to go around. Especially in a state like California, which is facing a severe housing crisis, the perpetual question of prioritization looms over every budget decision. And while tech companies often promise to save utilities money in the long term — via both efficiency gains and avoided disaster costs — implementing new programs often means big upfront expenses, which typically leads to higher customer rates. And, well, everybody hates that.
Suffice it to say, there’s no perfect solution here, but inaction is the worst option of all. As Winnacker put it, “you eat an elephant one bite at a time.” So as Los Angeles recovers from some of the most destructive fires in the state’s history and utilities across the state open themselves up to new ways of doing business, “we need to start with these small bites to get moving so that we can get past the either nothing can be done, this is an act of nature discussion or this pie in the sky, oh, you know, a single tech silver bullet will just make this problem go away,” Winnacker told me.
“This is an all of the above approach, and the time is probably now, with regard to having everyone’s undivided attention on this for a very brief period of time.”
Annual CO2 concentrations and forecasts in PPM. The Keeling Curve and Met Office
Long-term CO2 concentrations dating back 2,000 years.The Keeling Curve and Met Office