Engineering Global Resilience

From parched landscapes in California to submerged streets in Missouri, 2024 was a brutal reminder that climate volatility isn’t a distant threat—it’s a balance sheet risk today. In the U.S. alone, the NOAA recorded 27 separate billion-dollar climate disasters in 2024, totaling over $184 billion in damages. Wildfires, hurricanes, and flooding events didn’t just torch forests or knock out power—they shredded supply chains, displaced communities, and triggered cascading insurance failures.

The real cost isn’t just economic. It’s human. It’s systemic. And it’s growing.

At SF Climate Week 2025, this issue was top of mind for us at Wireframe - prompting us to convene this panel on "Engineering Global Resilience", with experts from Google, Gridware, Understory and Nephila Advisors. The stakes are clear: How do we design, fund, and deploy the tools that keep critical systems functional when nature doesn’t play nice?

Katie Fehrenbacher of Axios set the stage by framing the urgency. Climate events have moved from rare shocks to recurring disruptions. The question facing public and private actors isn’t whether to respond, but how fast and where first.

That question is already shaping how businesses allocate risk. Alex Kubicek, CEO of insurtech startup Understory, pointed to Hurricane Milton as a turning point: “We had $400 million in auto inventory exposed. Standard models projected $11 million in losses. Because our clients followed storm action plans, we had zero.” These weren’t just lucky calls—they were coordinated actions enabled by Understory’s proprietary weather-sensing network and real-time insurance modeling. The company is part of a new breed of insurtechs that don't just price risk—they actively reduce it.

On the electrical grid, the problems are just as urgent. Gridware COO Meghan Reibstein explained how her startup is embedding intelligence into some of the most overlooked assets in energy: wooden utility poles. “We’re turning poles into IoT devices. That allows utilities to detect wind events, falling trees, and wildfire conditions in real time. And utilities aren’t slow adopters anymore—they’re calling us.” The startup’s Gridscope sensors provide granular context that lets utilities act before a fire starts, not just after one ignites.

Pascal Karsenti of Nephila Advisors expanded the lens to the broader reinsurance market. Modeling wind and flood is table stakes; wildfire is the monster under the bed. "Wildfires spread structure to structure. To assess risk, I don't just need your roof condition—I need every building between you and the wildlands. That’s a geospatial data problem at massive scale." Karsenti emphasized that tech like aerial imagery and AI overlays can help, but only if local governments standardize and open their inspection data. "If your community can prove it's done the work, it should be more insurable."

That interplay between private tech and public capacity was a recurring theme. Google’s Mansi Kansal described their work expanding flood forecasts to over 100 countries and helping cities simulate the impact of reflective roofs and urban tree canopies. "These tools let cities advocate for funding, plan interventions, and understand the ROI of passive cooling," she explained. She also added “and there’s even more we can do to tie risk data to real business decisions.”

When the topic turned to generative AI, speakers noted that LLMs aren’t magic—but they can compress complex geospatial forecasts into action plans that non-experts can act on in minutes. Kubicek pointed out how that speed translates to saved assets: “We use generative AI to turn a meteorologist’s ten-page storm report into two sentences for a service manager. Move the cars now. That’s the difference between $10 million in damage and zero." Kansal also highlighted Google’s recently announced geospatial reasoning framework that could allow developers to build bespoke workflows to handle intricate geospatial analyses using LLMs.

Yet the panel didn’t shy away from the structural barriers. Building codes lag. Wildfire mitigation funding is patchy. And misinformation or political inertia can delay interventions. "You shouldn’t be allowed to build homes with flammable siding and shredded wood landscaping in fire zones," Karsenti said bluntly. "That’s not freedom. That’s negligence."

Reibstein tied it back to economics: "After a billion-dollar fire, suddenly our solution seems cheap. Calls spike. Interest surges. It shouldn’t take a catastrophe to realize the ROI on resilience."

So where does this leave founders and innovators in climate tech?

It leaves them staring at a multi-trillion dollar opportunity. Insurance, utilities, urban planning, public safety—all are overdue for resilience upgrades. But this isn’t about selling dashboards. It’s about full-stack solutions that detect, predict, and act. The winners won’t just model risk. They’ll compress time between signal and response.

Startups that can integrate domain expertise, sensor tech, real-time modeling, and pragmatic go-to-market execution have a shot at reshaping the landscape. But they’ll also need strong partners, smart capital, and customers who don’t just see climate tech as insurance—but as essential infrastructure.

At Wireframe Ventures, we’re backing the builders of that infrastructure. If you’re developing solutions to climate volatility—not just software, but systems that make communities stronger, faster, and safer—we'd love to hear from you! Let’s engineer global resilience, together.