Why we invested in Paar Autonomy

An autonomous machine depends on what it can perceive. A drone that cannot perceive its surroundings reliably cannot patrol a border, inspect a pipeline or follow a target, and the perception problem, sensors plus the intelligence to interpret them, is one of the harder problems in robotics. Paar Autonomy, founded in Bengaluru in 2024 by Vignesh Jayaraman, builds that perception layer for unmanned vehicles in the air, on the ground and at sea. We led the company's first institutional round in 2026.

What changed

For most of the last century, defence spending followed one rule: the platform is the product. Aircraft, ships and tanks were built in small numbers at enormous cost, and the sensors and software inside them were sub-contracted line items. The wars of the last four years reversed that. In Ukraine, in the Red Sea and in the exchanges between India and Pakistan in May 2025, cheap uncrewed systems produced in thousands did the work that expensive platforms used to do, and the value migrated from the airframe to the payload and the intelligence that runs it.

The Indian experience made this concrete. Pakistan launched 300 to 400 drones at 36 locations in a single night during Operation Sindoor, and the Defence Secretary afterwards called the conflict a reality check that exposed gaps in electronic warfare and counter-drone capability. Within a year the Defence Acquisition Council had cleared about ₹52,000 crore of procurement with counter-drone and air-defence systems as the stated tilt, and DRDO's D4 counter-drone system had been inducted with at least 72 percent indigenous content. The FY27 defence budget grew 15.2 percent to ₹7.85 lakh crore, with three-quarters of the modernisation budget reserved for domestic industry.

Procurement doctrine is changing with the budget. The shift American planners call mosaic warfare, many low-cost networked systems instead of a few expensive ones, suits companies that can deliver a payload or a software module in months for tens of crores, rather than a platform in a decade for thousands. Contracts of that size suit early-stage companies. They are also the reason we wanted a company at the payload layer rather than at the airframe layer, where listed and well-funded Indian OEMs already compete on tenders.

Where the value sits

Paar sits at the second and fourth layers. Its product is a multi-sensor, gyro-stabilised gimbal that gives a vehicle steady, high-resolution vision in difficult conditions, paired with autonomy software that lets the vehicle detect threats, navigate without GPS and carry out a mission without an operator steering every step. The same stack works on a drone, a ground robot and a surface vessel, which matters because a supplier who qualifies with one buyer can sell to all of them.

The founder

Vignesh led product at Asteria Aerospace, one of India's established drone makers, and watched what Indian platform companies relied on and where their imported payloads fell short. He is a robotics engineer by training, from Washington University in St. Louis, and the company's early angel investors and advisers come from the Indian drone and naval-systems industry, so the company's first customers already knew its founder's work.

He explains the product in the buyer's terms, which matters more in a payload company than in most, because the buyer is an engineer.

What we are betting on

• That Indian buyers will keep paying for perception and autonomy as separate line items, rather than accepting whatever the platform OEM bundles in. The D4 induction and the sovereign-architecture instinct in recent Army mandates point that way.
• That the same payload sells into industrial inspection, ports and policing, so that the company survives the gaps between defence orders.
• That the data from each deployment improves the models faster than an integrator can copy the hardware.

The path from a development contract to a production order has widened over the last two years. We will report on Paar's progress along it.