VECROS Unveils Jetcore, an India-Designed 'Brain' for Drones That Fly Without GPS

Most drones stop being clever the moment they lose their satellite signal. Fly one into a warehouse, a mine shaft, a collapsed building or a jammed battlefield, and the GPS-dependent autopilot that guided it outdoors is suddenly blind. In late July 2026, Delhi-founded, Bengaluru-based deep-tech startup VECROS moved to sell the answer to that problem as a product in its own right, unveiling Jetcore — an India-designed AI autonomy compute platform for drones and robots built to keep thinking when GPS goes dark.

From finished drone to the module inside it

VECROS is best known for ATHERA, its flagship autonomous drone that navigates GPS-denied, signal-denied and high-risk environments using onboard cameras and sensors rather than external positioning. Jetcore takes the compute-and-autonomy core that powers ATHERA and packages it as a standalone platform other manufacturers can build around.

According to the company, Jetcore integrates high-performance AI compute, real-time flight and motion control, multi-sensor connectivity, power architecture, embedded firmware and VECROS's proprietary Jetpix autonomy stack into a single module. In other words, it bundles the hardware "brain" and the software that makes a machine autonomous into one board that a drone or ground robot builder can drop into their own airframe or chassis.

What "autonomy without GPS" actually means

The technical heart of Jetcore is a set of capabilities that let a machine understand where it is and what is around it using only what it carries. The platform is designed to run visual-inertial odometry, GPS-denied localisation, SLAM (simultaneous localisation and mapping), obstacle avoidance, multi-camera perception, sensor fusion, 3D mapping and full autonomous mission execution — all onboard, without continuous cloud connectivity and without a pilot in the loop.

That combination is what separates a genuinely autonomous system from a remotely-piloted one. A drone that can localise and map from vision and inertial data can operate indoors, underground, inside metal structures that block radio, or in contested airspace where signals are deliberately jammed — precisely the settings where conventional platforms fail.

Built around NVIDIA Jetson

Jetcore is built around NVIDIA's Jetson edge-AI hardware, the compute platform that has become a default choice for robotics developers who need to run neural networks on the machine itself. Pairing Jetson-class compute with VECROS's Jetpix stack and flight-control electronics is what allows the heavy perception workloads — reading a dozen camera feeds, fusing them with inertial and depth data, and turning that into navigation decisions in real time — to happen at the edge rather than on a distant server.

VECROS says the platform arrives after more than 5,000 hours of autonomous flight and early commercial validation, meaning the autonomy software has been hardened in the field before being offered to third parties.

A wide addressable market

By selling autonomy as a module rather than only as a finished drone, VECROS is aiming at a much larger set of customers. The company is pitching Jetcore to drone and robotics manufacturers across defence, warehousing, industrial inspection, infrastructure monitoring, mining, utilities, public safety and agriculture — sectors where GPS-denied navigation, obstacle avoidance and onboard intelligence are the difference between a demo and a deployable system.

The strategy mirrors a broader pattern in robotics: the highest-value, hardest-to-copy layer is often the autonomy stack, not the airframe. A company that can license that stack to many hardware makers can spread its engineering investment across a far wider base than one selling a single product.

The company behind it

VECROS — a contraction of "vector" and ROS, the Robot Operating System — was founded in 2021, with roots in aeromodelling and research work at IIT Delhi. It was co-founded by chief executive Besta Prem Sai, an engineering-physics graduate of IIT Delhi who ran the institute's aeromodelling club, and chief technology officer Rajeshree Deotalu, a robotics specialist from NIT Nagpur. The startup has drawn backing from early investors and programmes including NVIDIA's Inception initiative and the defence-innovation ecosystem around iDEX, reflecting both its AI-hardware focus and its relevance to Indian security use cases.

Why it matters

India has a fast-growing drone industry, but much of it still assembles airframes around imported flight controllers and foreign autonomy software. A domestically designed autonomy compute platform — one that handles perception and navigation onboard, works where GPS does not, and can be embedded by many manufacturers — chips away at that dependence in one of the most strategically sensitive layers of the stack.

Whether Jetcore becomes a standard building block for Indian drone makers will depend on price, reliability and how openly it can be integrated. But the launch is a notable marker: an Indian deep-tech startup is now trying to sell the intelligence that goes inside autonomous machines, not just the machines themselves.

Sources

• VECROS Unveils Jetcore, an India-Designed AI Autonomy Compute Platform for Drones and Robots — BISinfotech
• Startup news and updates: Daily roundup (July 30, 2026) — YourStory
• VECROS builds AI-powered, GPS-free autonomous drones — Electronics For You
• How VECROS plans to dominate international skies with GPS-defying drones — Inc42