Defence autonomy
The economics of defence have inverted. Cheap, autonomous, networked systems now beat expensive platforms, and India's budget is moving that way.
What we believe
The segment is defined by a cost ratio. A one-way attack drone costs tens of thousands of dollars; the missile that has traditionally shot it down costs a million or more. Closing that gap is now the main priority of procurement, and the fix is a stack of cheap sensors, autonomy software and interceptors that cost about as much as the threat. This is what American planners call mosaic warfare: many low-cost systems, networked, instead of a few expensive ones. It suits startups because the contracts are the size startups can deliver.
Interceptor economics
The cost of the shot against the cost of the threat, multiplied by how many shots a raid needs. Change the assumptions to see how the ratio moves.
Reference points: one-way attack drones of the Shahed-136 class are estimated at $20,000 to $50,000 (CSIS and RUSI, cited by open-source trackers); a PAC-3 MSE interceptor is listed at about $5.3 million and older Patriot rounds at about $1 million (JINSA cost tables, July 2025); Ukrainian FPV interceptors cost $1,000 to $2,500 and hit rates of 80 to 90 percent are reported by trade press; Pakistan launched 300 to 400 drones at 36 locations in one night during May 2025 (Ministry of External Affairs briefing). Wartime production and hit-rate figures are directional.
We invest at the layers primes integrate around rather than at the layers they own: perception and autonomy (Paar Autonomy), high-speed unmanned aircraft and the targets used to train against them (Naveria Space), and the components inside both.
Source: Union Budget documents; FY27 analysis by MP-IDSA (total ₹7,84,678 crore, up 15.2 percent; capital outlay ₹2,19,306 crore).
What must happen next
The channel from grant to order is the binding constraint. iDEX has engaged 676 startups and signed 551 contracts, and the armed forces have procured more than ₹2,400 crore from iDEX innovators, but the step from a development contract to an acceptance of necessity still takes years. We underwrite procurement literacy in a founding team as seriously as the technology.
Source: Department of Defence Production, iDEX programme data, March 2026.
Source: JINSA cost tables (July 2025) for missiles; US Army testimony and trade press for interceptor drones (Merops about $15,000; Coyote about $100,000); Ukrainian FPV interceptor costs reported by trade press. Directional.
What we look for
- Payload rather than platform: sensors, seekers, EW and mission software that sell to every airframe.
- Counter-UAS layers that survive threats that cannot be jammed: passive detection and kinetic autonomy.
- Founders who have sold to a service or a DPSU before, or who bring someone who has.
- Dual-use revenue that keeps the company alive between defence orders.
What we pass on
Companies whose only route to revenue is a lowest-price tender against a prime. Jammer-only counter-drone stacks. Payloads that are more than 70 percent imported with no substitution roadmap.