Analysis
Sponsored
AI

In a crowded market, firetech veteran Technosylva is betting on utilities

The 27-year-old data company is embracing AI — and finding new customers.

|
Published
October 1, 2024
Listen to the episode on:
Apple Podcast LogoSpotify Logo

Photo credit: Cal Fire

Photo credit: Cal Fire

In early August, while the Park Fire was consuming hundreds of thousands of acres in northern California, Cal Fire's staff chief for fire intelligence testified before a joint hearing of the California legislature about the agency’s use of artificial intelligence.

Since 2019, Phillip SeLegue said, the agency has been using a suite of AI-powered prediction tools to model the behavior of fires, including the Park Fire.

“Traditional methods of wildfire prediction have relied heavily on historical data and basic models,” he said, “which often fall short of accounting for the complex and dynamic nature of fire behavior.”

Today, though, Cal Fire uses algorithms developed by firetech company Technosylva to process data including weather conditions, topography, fuel characteristics, and satellite imagery to identify patterns and correlations that might not be visible through more conventional analysis, and to make predictions up to five days in advance.

For example, he added, the models can be trained to recognize the influence of specific variables of fire behavior, such as the effect of wind on fire spread or the impact of humidity levels on fuel moistures. For the Park fire — which burned 429,603 acres before being fully contained — a team of four analysts were assigned specifically to use those models to analyze and predict the fire’s movements.

Bryan Spear, Technosylva’s CEO, said the market for this kind of data has only started to take off in the last year, as utilities recognize the “risk at the asset level” that wildfires pose. But as a slew of startups with millions of dollars in venture funding in hand enter the space that Technosylva has worked in for nearly three decades, the company’s longevity means it has data that no one else does.

“When we put AI and machine learning on top of these sophisticated models, and all these 27 years of data that we have, that's when it really started to take the next level,” he told Latitude Media, “and started to expedite the decision making in the field.” 

Listen to the episode on:
Apple Podcast LogoSpotify Logo
WHITE PAPER
The future of flexibility: Unlocking the full value of VPPs for a changing grid

Brought to you by EnergyHub: Learn about the untapped potential of VPPs and how utilities can scale these technologies to deliver a more reliable, resilient, and flexible grid.

DOWNLOAD NOW
WHITE PAPER
The future of flexibility: Unlocking the full value of VPPs for a changing grid

Brought to you by EnergyHub: Learn about the untapped potential of VPPs and how utilities can scale these technologies to deliver a more reliable, resilient, and flexible grid.

DOWNLOAD NOW
WHITE PAPER
The future of flexibility: Unlocking the full value of VPPs for a changing grid

Brought to you by EnergyHub: Learn about the untapped potential of VPPs and how utilities can scale these technologies to deliver a more reliable, resilient, and flexible grid.

DOWNLOAD NOW
WHITE PAPER
The future of flexibility: Unlocking the full value of VPPs for a changing grid

Brought to you by EnergyHub: Learn about the untapped potential of VPPs and how utilities can scale these technologies to deliver a more reliable, resilient, and flexible grid.

DOWNLOAD NOW
Get in-depth coverage of the energy transition with Latitude Media newsletters

In some ways, Technosylva looks very similar to many of those newer companies leveraging AI for wildfire management. For one thing, the company primarily relies on the software-as-a-service model that’s extremely common in the firetech segment. For another, Technosylva isn’t relying solely on the utility sector to support it — though it’s certainly a key customer.

However, the company’s only outside investor is TA, a private equity firm specializing in growing already-profitable companies. And, given the company’s longevity and its partnership with Cal Fire, Technosylva also seems to have answered a question that many companies in the field are still grappling with: how to make money in the long-term.

The new utility customer 

Technosylva was founded in 1997, years before wildfires became a way of life in most of the West. The company got its start developing geographic information systems technology to help fire agencies model active fires to determine where they’d move. 

Today, though, the company’s largest customer segment is no longer those agencies, but rather electric utilities. Spear said that utilities use their Technosylva’s AI tools to predict ignition conditions and, ideally, head fires off before they start. 

It wasn’t until last year that Technosylva’s utility customer segment really started to take off, Spear said, but utility interest in the tech dates back to 2007, when San Diego Gas and Electric was found and held liable for the ignition of the Witch Creek Fire, and was ultimately forced to pay millions of dollars in damages to private and federal property.

Technosylva currently provides its wildfire risk platform to utilities in 15 states, Spear said, but the segment emerged first in California, thanks in part to a clause in the state’s Constitution that allows citizens to sue public utilities and other government entities for property damage.

San Diego Gas and Electric ended up being Technosylva’s first utility customer, and the two partnered to adapt prediction tools to the needs of utilities in 2017. Southern California Edison and Pacific Gas & Electric signed on a few years later.

“At that point it was still seen as a California problem,” Spear said. But Boulder’s Marshall Fire in 2021 was something of a turning point: “That’s when it started to become apparent that this is…more of an industry problem.”

The fire in Lahaina, Hawaii last year “woke up the utility industry,” Spear added, because Hawaiian Electric Company, which acknowledged its power lines had likely caused the blaze, “lost 75% of their market capitalization overnight — and it was an area that was not widely expected to be at high risk.”

In 2024, Technosylva is running half a billion simulations every day in California alone, using inputs like building locations, critical facilities, and utility asset data to make localized predictions.

No items found.