The best way to spot a drone, according to Xubin Aerospace, is to watch the air move. The Oakville-based startup is building a long-range detection system that doesn't listen for radio signals or ping the sky with radar. Instead, its "Eagle Vision" technology uses off-the-shelf electro-optical and infrared cameras, paired with proprietary AI, to look for the subtle aerodynamic disturbances a drone leaves in its wake. It is a quiet, passive approach for a market that is increasingly loud and contested.
A wedge in the counter-drone stack
Most counter-unmanned aerial system (CUAS) technologies rely on active sensing. Radar emits energy; radio-frequency (RF) detectors listen for control signals. These methods work, but they have blind spots. Radar can struggle with small, slow-moving objects at low altitude. RF detection fails against pre-programmed or autonomous drones. Xubin's bet is that a purely optical layer, supercharged by computer vision trained on fluid dynamics, can fill those gaps. The system is designed to provide continuous situational awareness over sprawling, complex sites like rail yards, bridges, and energy facilities where other modalities are limited [xubinaerospace.com, retrieved 2024].
The company claims its AI doesn't just look for the drone's shape, but classifies targets by their unique "flow signature",the way they disturb the air around them. This aerodynamics-informed approach is the core of its differentiation. It has already demonstrated a Technology Readiness Level (TRL) of 5, achieving a 1 km detection range with what it calls "primitive hardware" [10]. The roadmap aims for 2-5 km with higher accuracy.
Validation from defense, capital from aviation
For a hardware and AI startup founded in 2023, Xubin has secured the kind of early validation that opens doors in the defense and critical infrastructure sectors. In 2025, it was among ten winners of NATO's DIANA Innovation Hackathon and was selected for Canada's IDEaS Counter-UAS Sandbox. These are not revenue signals, but they are powerful credibility signals for a company courting government and industrial buyers.
Financing is following the technical milestones. The company is currently seeking a CAD $3.5 million pre-seed round (estimated) [The Icebreaker]. It has already secured an early-stage minority investment from Xegasus Aviation Investments, a Netherlands-based firm [LinkedIn]. The table below outlines the known competitive landscape Xubin is entering.
| Competitor | Core Technology | Notes |
|---|---|---|
| DroneShield | RF, acoustic, radar | Publicly listed, broad product suite. |
| Robin Radar Systems | Radar-focused | Specializes in bird and drone detection. |
| Teledyne FLIR | EO/IR, thermal imaging | Defense incumbent with extensive hardware. |
| Sentrycs | RF, AI | Focus on adaptive RF detection and mitigation. |
Xubin's position is distinct: it is not trying to out-radar the radar companies. Its wedge is a passive, camera-based system that claims to see what others miss by understanding physics, not just pixels.
Where the wheels could come off
The ambition is clear, but the path from hackathon winner to deployed system is paved with hard, expensive problems. Three stand out.
- The environmental variable. An AI trained to see aerodynamic signatures must contend with wind, weather, and the visual noise of a cluttered environment,swaying trees, moving vehicles, flocks of birds. Proving high accuracy and low false-alarm rates in real-world conditions, beyond a controlled demo, is the non-negotiable next step.
- The unit economics of sight. Long-range EO/IR systems capable of 5 km detection are not cheap, even with off-the-shelf cameras. The AI software must deliver enough marginal detection capability to justify the total system cost over a competing radar or layered solution. The business case rests on being a necessary complementary layer, not just a novel one.
- The solo-founder scale. The company is led by CEO and co-founder Mostafa Najafiyazdi [CIX Summit]. Building, selling, and deploying dual-use hardware systems for defense and infrastructure is a multi-disciplinary marathon requiring expertise in hardware engineering, AI research, enterprise sales, and regulatory compliance. Assembling that team will be as critical as refining the algorithm.
The next twelve months
The coming year is about moving from prototype to pilot. Success looks like securing one or two paid deployments with a critical infrastructure operator,a rail network or port authority,to generate real-world performance data. That data would be the strongest possible argument for the next funding round and for challenging the deeper-pocketed incumbents.
On paper, the energy required is not trivial. A typical EO/IR sensor for this range might consume 50-100 watts. Running multiple sensors, plus the AI compute, at a remote site could demand a few hundred watts continuously. That is the equivalent of leaving a dozen old incandescent light bulbs on, forever, to watch the sky. The trade is straightforward: if the system prevents a single drone-based disruption to a freight corridor or energy substation, the electricity bill becomes a rounding error.
The incumbent Xubin must ultimately beat is not a startup, but a line item. It is the default, budget-approved radar system from a supplier like Teledyne FLIR. To displace it, Xubin needs to prove its optical layer doesn't just add cost, but adds indispensable sight.
Sources
- [xubinaerospace.com, retrieved 2024] Eagle Vision, See What Others Can't | https://xubinaerospace.com/
- [The Icebreaker] The new arsenal of war | https://www.theicebreaker.ca/p/the-new-arsenal-of-war
- [10] NATO DIANA Hackathon 2025 win and IDEaS Sandbox selection
- [LinkedIn] XEGASUS PARTICIPATES IN DUAL USE START-UP XUBIN AEROSPACE | https://www.linkedin.com/posts/xegasus-investments_xegasus-participates-in-dual-use-start-up-activity-7371553541552771072-IlzO
- [CIX Summit] Mostafa Najafiyazdi speaker profile | https://cixsummit.com/speakers/mostafa-najafiyazdi