DJI is doubling down on one of the biggest shifts happening in the drone industry right now: moving from simply capturing data to actually understanding it in real time.
DJI has launched its Enterprise Drone Onboard AI Challenge 2026, inviting developers, engineers, and companies to build AI-powered solutions that can run directly on drones, not just in the cloud. At its core, this challenge is about turning drones into decision-makers, not just flying cameras.
From data collection to real-time intelligence
DJI Enterprise is framing this initiative as the next evolution of drone technology. Traditionally, drones have been great at collecting high-quality aerial data. But analyzing that data often happens later, on the ground, creating delays in critical workflows.
With newer hardware like the DJI Matrice 4 drone series and the DJI Manifold 3, the tech giant is pushing toward drones that can process and interpret data mid-flight. That means identifying issues, detecting anomalies, or making decisions instantly — while still in the air.
The challenge is designed to accelerate that transition by bringing developers into the ecosystem and pushing solutions from prototype to real-world deployment.
What DJI wants developers to build
Participants in the challenge will focus on three key deployment approaches:
- AI running directly on the drone
- AI powered by the Manifold 3 onboard computer
- Cloud-based AI integrated with drone operations
The goal is simple: close the gap between “seeing” and “acting.”
DJI says these solutions could unlock major efficiency gains across industries like power line inspections, public safety, agriculture, natural resource management, and emergency response. Think drones that can autonomously detect damaged infrastructure, flag safety risks, or assist first responders without waiting for human analysis.
How to participate
Submissions are already open and will remain live until May 10, 2026.
To enter, participants must first register on DJI’s developer platform. From there, they’ll need to submit a complete solution package that includes:
- A detailed explanation of the AI model
- A working model tied to a designated test device
- A validation dataset with at least 20 samples
- A full workflow screen recording
- A declaration of independent intellectual property
This isn’t a lightweight hackathon-style entry. DJI is clearly aiming for production-ready solutions that can move beyond demos.
The competition comes with a mix of recognition and tangible rewards. Top honors include the Best Onboard AI Model Award, which will go to the most technically advanced solution, and the Industry Application Excellence Award, focused on real-world usability and deployment potential.
Winners can expect prizes like DJI hardware bundles, including the Osmo Pocket series and Manifold 3 Combo, along with official recognition and ecosystem support.
Finalists will be announced in May 2026, followed by a public testing and voting phase. Final winners are expected to be revealed in June. Interestingly, DJI is adding a public voting component alongside expert evaluation, signaling a push to validate not just technical excellence, but also real-world relevance and usability.
This challenge reflects a broader industry trend: drones are no longer just data collection tools; they’re becoming autonomous systems capable of understanding and acting on what they see. If DJI succeeds in rallying developers around onboard AI, it could significantly speed up adoption across enterprise sectors where time-sensitive decisions matter most. In other words, the future DJI is betting on isn’t just smarter drones; it’s drones that think.
More: FCC eyes sweeping reforms to boost US drone power
FTC: We use income earning auto affiliate links. More.
Comments