Evolve 2026/ Session notes

Embodied Intelligence at Work

EVOLVE 2026 · MIAMI · BREAKOUT SHOWCASE · SESSION 05

Chris Suzdak (moderator) · Dr. Demetrius Coleman · Spencer Horne · Sára Nožková · 6 min read

Watch Full Session

THE 60-SECOND READ

Three founders, three sensors, one idea. Physical AI is only as good as the data feeding it, and most of the physical world is still unmeasured. A robot that inspects water pipes from the inside. An autonomous airship built to close the gap between drones and satellites. A field device that talks to wildlife in their own language. All three plug straight into the Sense layer of SandOS.

Key Highlights

By the numbers (data points sidebar)

underground water pipe in the US
0 M mi
needed to replace aging pipes, 20 yrs
$ 0 B
of the world imaged at drone resolution
< 0 %
animals deterred by Flox
0 +

In their words

“What if instead of fighting wildlife, we could communicate with them in a language that they actually understand and use already?”
Sára Nožková, CEO, Flox Intelligence

Notes

Sand’s Chris Suzdak framed the session before any founder spoke. These three companies are not pitching for investment. They are showing the kind of sensor and data partners that plug directly into the SandOS Marketplace unveiled in the keynote, the physical layer that lets AI sense, analyze, and act in the real world.

Dr. Demetrius Coleman walked through the economics of reactive maintenance. Most water utilities fix pipes once they break, or plan replacements on engineering guesswork based on a pipe’s age and material. MotMot’s robots replace that guesswork with recorded data, cameras, audio, and pressure sensors moving through the pipe itself. The current generation is tethered for safety. Full autonomy is next.

Spencer Horne’s pitch was about resolution and reach as a trade-off. Satellites see everything but resolve little. Crewed aircraft resolve well but are too expensive to fly often. Drones resolve best of all but can’t cover distance. Cloudline’s airship is built to sit in the middle of that triangle, flying far enough and often enough to build the kind of longitudinal dataset that makes change detection possible over time.

Sára Nožková closed with the case for Flox. Legacy deterrents like pyrotechnics and scarecrows fail because animals habituate to them, and because they take constant human labor to run. Her field unit, called Edge, is built for remote and harsh environments, works across species, and learns from every interaction. Every encounter feeds a platform that can also generate the regulatory reports authorities require.

From the room (session gallery)

Get started

Physical AI for Critical Infrastructure.