EVOLVE is Sand’s annual summit bringing together technology leaders, policymakers and practitioners working at the intersection of AI and critical infrastructure.
If Physical AI depends on understanding the real world, the first challenge isn’t intelligence—it’s visibility.
This showcase brought together three founders solving a common problem in very different environments: making previously invisible systems measurable. From underground water networks to remote infrastructure and wildlife ecosystems, each demonstrated how new sensing technologies are expanding the data available for AI to analyze and act upon.
Across every demonstration, the discussion returned to the same principle: AI cannot optimize what it cannot observe. Expanding the world's ability to sense and measure physical systems is becoming just as important as improving the models themselves.
Whether inspecting aging water infrastructure, monitoring pipelines or reducing wildlife collisions, each company demonstrated how richer operational data enables organizations to move from reactive responses to earlier, more informed intervention.
Although the founders tackled very different challenges, they shared a common goal: capturing data that previously didn't exist or was too difficult, expensive or infrequent to collect. From underground pipes to remote infrastructure and wildlife behaviour, each innovation expands AI's understanding of the physical world—and, with it, its ability to drive better outcomes.
underground water pipe in the US
Estimated investment needed to replace aging water infrastructure over the next 20 years
of the world imaged at drone resolution
Wildlife encounters prevented by Flox Intelligence
Much of today’s AI conversation focuses on models—the algorithms that interpret data and generate decisions. This showcase highlighted a different reality: before AI can become more intelligent, the physical world must become more observable.
Across water infrastructure, aerial inspection and environmental monitoring, each founder demonstrated that the real bottleneck often isn’t analysis—it’s access to reliable, high-quality operational data. Organizations can’t predict failures in pipes they have never inspected. They can’t monitor infrastructure they rarely see. And they can’t manage ecosystems they don’t understand.
That perspective also reinforced one of the central themes of Evolve 2026: Physical AI is built through ecosystems rather than standalone technologies. Sensors, platforms and intelligent models each play different roles, but their value grows when they operate together. The technologies showcased weren’t competing approaches to Physical AI—they were complementary ways of expanding what AI can sense before it analyzes or acts.
Perhaps the session’s most important insight was this: the future of AI won’t simply depend on building smarter models. It will depend on building richer connections to the physical world those models are designed to understand.
Vice President Partnerships & Marketing, Sand
Moderator
CEO & Co-Founder, Flox Intelligence
Co-Founder & Chief Technology Officer, Motmot
Founder & CEO, Cloudline
