Percepto's Axion Wants to Read Every Power Line
The autonomous-drone firm pivots to software, opening its AI inspection engine to any hardware — and the timing is no accident.
By The Robotics Desk
Percepto has spent years putting autonomous drones above critical infrastructure. Now it wants the data those flights produce — and data from everyone else's flights too. The company this week launched Axion, a hardware-agnostic platform that ingests feeds from third-party drones, fixed sensors, and legacy inspection archives, then runs them through more than 160 AI models to flag failures before they escalate. [1]
The move marks a deliberate strategic break. Julien Bellanger, who recently took the helm from longtime CEO Dor Abuhasira, is steering Percepto away from being a drone maker and toward being what he calls a physical AI provider. [1] The company, founded in 2014 and headquartered in Modi'in, Israel, with a U.S. base in Austin, Texas, will keep selling and supporting its existing drone-and-dock hardware — but software is now the growth vector. [1]
"When I looked at Percepto, I saw a company that had already done the hardest part: fully autonomous drones flying real missions every day at some of the most demanding industrial sites in the country," Bellanger told The Robot Report. "With AIM and Axion, we're taking that operational intelligence beyond our own fleet and opening it to the entire industry." [1]
Axion is delivered as a software-as-a-service product. It organizes inspection data by asset and applies AI that interprets equipment condition in context, so operators are not tied to a single hardware provider. The platform can identify a downed wire by its relationship to surrounding components or evaluate a thermal anomaly against expected asset behavior and environmental conditions, according to Percepto. [1] Plain-language queries through a built-in large language model let field teams ask operational questions without writing code or pulling reports. [1]
Percepto's target verticals — upstream oil and gas, power transmission and distribution, and most downstream processing facilities — share a structural problem: inspection cycles that move slower than equipment degrades. [1] The AIM (Autonomous Inspection & Monitoring) layer underpinning Axion is designed to let operators plug in whatever hardware they already own, add Percepto's autonomous systems where coverage gaps exist, and use compliance waivers to run one-to-many autonomous flight configurations. [1]
The software-first pivot is a familiar arc in hardware-born companies — but in the drone sector it arrives at a moment when policy and product strategy are converging in ways that reward exactly this kind of platform thinking.
[2] Within 120 days, the FAA was also directed to deploy AI tools to expedite the review of UAS waiver applications. [2] For a platform built around compliance waivers and autonomous long-range flights, that regulatory acceleration is structural tailwind, not background noise.
The same order mandates that all federal agencies prioritize UAS manufactured in the United States over those made abroad to the maximum extent permitted by law. [2] Percepto is threading that needle deliberately: the company says manufacturing will be located onshore in the U.S. to support the American Drone Dominance Initiatives. [1] Being hardware-agnostic on the software side while keeping domestic production on the hardware side positions the company on both sides of the policy line.
What's being built here is less a product than an industrial nervous system — one that promises operators a single view of everything that might fail, sourced from whatever sensors they already trust.
Each Axion finding is backed by inspection history and expert validation, according to Percepto — which means the system's credibility depends on the breadth and fidelity of the data operators feed it. [1] The platform opens on day one to existing inspection archives, letting teams see which assets need attention first before committing to any new hardware. [1] Whether that low-friction entry point converts infrastructure operators at scale is the real test of the software-first bet.