Sponsored by
When GPS disappears, most drones fail
They slow down, lose direction, or return home.
Some now keep going.
A new class of drones, powered by onboard AI from companies like NVIDIA and tested in programs by DARPA, can fly and complete missions without any satellite signal.
No remote pilot.
No coordinates.
They rely entirely on what they see.
Traditional drones depend on GPS to know their position.
These drones do not.
They use cameras and onboard processing to estimate position relative to the environment. Instead of asking “Where am I on Earth?”, they continuously answer “Where am I inside this space?”
This approach is known as Visual SLAM.
Navigation using vision instead of GPS
Story 1: Flying through forests without crashing
Skydio builds drones that can move through dense environments without human control.
Once a subject is selected, the drone follows while avoiding obstacles in real time.
What makes this work:
Real time obstacle avoidance using multiple cameras in all directions
Onboard AI processing instead of relying on cloud or remote control
Continuous path adjustment as the subject moves unpredictably
3D mapping on the fly while navigating complex terrain
Stable tracking even when the subject goes behind objects briefly
👉 Product link: https://www.skydio.com
Future implication: This level of autonomy reduces the need for skilled pilots in dynamic environments.
Stop re-prompting. Say it right the first time.
Voice-first prompts preserve the nuance you cut when typing. Speak once, paste into any AI tool, get results that don't need a follow-up. 89% of messages sent with zero edits.
In the DARPA Subterranean Challenge, drones were tested in underground environments with no GPS and no external guidance.
They had to explore, map, and move through tight spaces on their own.
What made this possible:
Vision based localization without any external reference
Simultaneous mapping and navigation in unknown environments
Sensor fusion combining cameras and inertial data
Autonomous decision making when paths are blocked
Operation in complete signal denial conditions
👉 Program link: https://www.darpa.mil/research/challenges/subterranean
Limitation: Performance drops in low visibility environments like dust or darkness.
What the drone is actually doing
It is not just capturing video.
It is extracting structure from the scene. Edges, depth, and motion are used to estimate movement. As the drone moves, it builds a live 3D map and positions itself inside that map.
In simple terms, it is flying while constructing its own understanding of space.
Why this matters
GPS is not always available or reliable.
It can be blocked, jammed, or spoofed.
A system that does not depend on GPS can operate in places where traditional drones fail. This includes indoor environments, industrial sites, tunnels, and disaster zones.
This makes autonomy more practical in real conditions.
Drone inside tunnel
Forget the hype. Here's what's actually working in AI.
90% of AI content is noise. The AI Report is the 10%.
We cover real enterprise deployments, actual business outcomes, and the AI strategies leaders are betting on right now — not lab experiments, not demos, not speculation.
400,000+ executives, operators, and founders read us every weekday to cut through the clutter and make faster, smarter decisions about AI before their competitors do.
No hype. No fluff. Just the signal.
See what's actually working in AI across every industry right now — free, in 5 minutes a day.
The bigger shift
This is not only about drones.
It shows how machines are changing.
Older systems follow external signals.
Newer systems build their own understanding and act on it.
This pattern is now visible in robotics, self driving systems, and automation.
What comes next
As onboard AI improves, these systems will become faster and more reliable in difficult environments.
Interaction will also change.
Instead of controlling movement, you give intent.
“Inspect this structure.”
“Map this area.”
The system handles the rest.
Final thought
GPS told machines where they are.
Vision lets them understand what’s around them.
That shift is what makes real autonomy possible.