Field over demo
A system that only works in controlled conditions does not work. We design against wind, vibration, GPS degradation, and lossy links from day one, and we treat failed tests as primary data.
Company
A research company, on purpose
Brobot is a research-stage robotics company. We build autonomous machines — starting with vertical take-off aerial platforms and the coordination software that lets many of them act as one system.
- Founded
- 2025
- Stage
- Research & Development
- Base
- Haridwar, Uttarakhand, IN
- Focus
- Autonomous aerial systems
- Team
- Founder-led, hiring later
- Funding
- Pre-seed, bootstrapped
The story
Where this comes from
Brobot started with a simple observation: the hardest problems in robotics are not in the demo, they are in the field. Machines that work on a bench fail in wind, dust, degraded GPS, and unreliable networks. We are building for the field first.
We are currently in our research and development phase. That means our output today is prototypes, test data, working software, and published engineering notes — not revenue. We think the honest way to build a deep-tech company is to say exactly where we are and show the work.
Our near-term focus is a VTOL flight testbed and a multi-agent coordination stack built on ROS 2. Our long-term ambition is full-stack autonomy: airframes, embedded intelligence, and fleet coordination designed together rather than bolted together.
We are based in Haridwar, Uttarakhand — close to the terrain we want our machines to handle. India's manufacturing base, engineering talent, and open airspace-adjacent research corridors make it the right place to build this company.
Principles
How we decide
Four commitments that settle most arguments before they start.
Show the work
We publish engineering notes, test results, and roadmap updates as we go. Claims on this site describe what we have actually built or are actually building — nothing is inflated for effect.
Software-defined, hardware-serious
Autonomy is a software problem, but it lives or dies on hardware discipline: power budgets, thermal limits, EMI, and mass. We do both, and we prototype in-house.
Build for fleets, not units
One capable machine is a product; many coordinated machines are an infrastructure. Every architectural decision assumes multi-agent operation from the start.
Operating system
How we work
Research loop
Read, prototype, test, write up. Each research thread ends in a working artifact and a public note, so progress is legible to outsiders.
Simulation to field
Everything is containerised and simulated first — Docker-based ROS 2 environments give us reproducible builds — then validated on physical hardware.
Long horizon
We plan in multi-year phases, not sprints toward a demo day. The roadmap on this site is the actual plan we operate against, and we update it when reality disagrees.