As extreme weather becomes the norm, utilities are scrambling to improve the grid’s resilience, aiming to prevent the types of outages and infrastructure damage that often magnify the impact of already disastrous weather events. Those events cost the U.S. $182 billion in damages last year alone.
With the intensity of storms, heat waves, droughts, and wildfires growing every year, some utilities are now turning to artificial intelligence in their quest to adapt to new climate realities. Rhizome, which just announced a $6.5 million seed round, uses AI to help assess and prevent climate change-induced grid infrastructure vulnerabilities. It’s already working with utilities such as Avangrid, Seattle City Light, and Vermont Electric Power Company to do so.
“With a combination of utility system data and historical weather and hazard information, and then climate projection information, we can build a full profile of likelihood and consequence of failure at a very high resolution,” Rhizome co-founder and CEO Mish Thadani told me.
While utilities often have lots of data about the history of their assets and the surrounding landscape, there’s no real holistic system to bring together these disparate datasets and provide a simple overview of systemic risk across a range of different scenarios. Utilities usually rely on historical data to make decisions about their assets — a practice that’s increasingly unhelpful as climate change makes previously rare extreme weather events more likely.
Rhizome aims to solve both problems, serving as an integrated platform for risk assessment and mitigation that incorporates forward-looking climate modeling into its projections. The company measures its success against modeled counterfactuals that determine avoided power outages and the economic losses associated with these hypothetical blackouts. “So we can say the anticipated failure rate across the system for a Category 1 hurricane was X, and after you invest in the system, it will be Y,” Thadani told me. “Or if you’ve made a bunch of investments in the system, and you do experience a Category 1 hurricane, what would have been the failure rate had those investments not been made?”
This allows utilities to provide regulators with much more robust data to back up their funding requests. So while Thadani expects electricity prices to continue to rise and ratepayers to bear the burden, he told me that Rhizome can ultimately help regulators and utilities keep costs in check by making sure that every dollar spent on risk mitigation goes as far as possible.
Rhizome’s seed round, which came in oversubscribed, was led by the early-stage tech-focused venture firm Base10 Partners, which aims to automate traditional sectors of the economy. Additional funders include climate investors MCJ and CLAI, as well as the wildfire-focused venture firm Convective Capital. In addition to its standard risk assessment system, Rhizome has also developed a wildfire-specific risk mitigation tool. This quantifies not only how likely a hazard is to occur and its potential impact on utility infrastructure, but also the probability that an equipment failure would spark a wildfire, based on the geography of the area and historical ignition data.
Thadani told me that he considers evaluating wildfire risk “to be the next step in a sequence” as a utility evaluates the threats to its system overall. So while customers can choose to adopt either the standard product or the wildfire-specific product, many could gain utility from both, he said. The company has also developed a third offering specifically tailored for municipal and cooperative utilities. This more affordable system doesn’t provide the same machine learning-powered cost-benefit metrics, but can still help these smaller entities evaluate their infrastructure’s vulnerability.
Right now, Rhizome has a “lean and mighty” team of just 11 people, Thadani told me. With this latest raise, he said that the company will immediately hire five or six engineers, primarily to do further research and development. As Rhizome looks to onboard more and larger customers, it’s planning to incorporate more advanced modeling features into its platform and operate it increasingly autonomously, such that the model can retrain itself as new weather, climate, and utility data becomes available.
The company is out of the pilot phase with most of its customers, Thadani said, having signed multiple enterprise software contracts. That’s big, as utilities have gained a reputation for showing an initial appetite for testing innovative technologies, only to balk at the cost of full-scale deployment. Thadani told me Rhizome has been able to avoid this so-called “pilot purgatory” by making a point to engage with senior-level stakeholders at utilities — not just the innovation teams — to “graduate from that pilot ecosystem more quickly.”
