Defence Startups: Built in India, Proven in the Field

The idea

Five engineers in Coimbatore build a counter-drone system in a rented shed. They read the published trial standard and the fixed price per unit on one portal, and they book a slot at the next quarterly trial in the desert. Their system passes. The scores are online within a week, the order arrives within 30 days, and payment lands 30 days after delivery.

On a high-altitude post, a soldier's drone receives a software update every quarter, keeping it a step ahead of the latest jammer. In Lucknow, a young graduate chooses a defence startup over a job abroad, because Indian defence technology is now a thriving industry.

I believe the best way to defend a nation is to back its own inventors with clear rules, a fair price and fast payment. Ukraine had no time for long programmes, so it let frontline units buy tested drones straight from makers, and here is what happened: more than 500,000 drones ordered. When performance in the field decides who wins, talent rises quickly and our forces get the best kit, made at home.

Why now

The money is already in our budget. Setting aside 5% of defence capital spending for field-proven Indian products creates about ₹10,965 crore a year of steady demand.

Our startups are ready. iDEX has already cleared 58 startup prototypes for procurement. The next step is turning far more of them into orders at scale.

The momentum is strong. Defence exports reached ₹38,424 crore in FY26, up 62%, and Operation Sindoor made drones a frontline need.

Speed is the prize. Jammers now adapt to a new drone in 6 to 8 weeks. Quarterly upgrades let our kit keep pace.

The big moves

Here is the path I see, one step after another.

1. First, publish what we want to buy and at what price. We start with the first 10 priority products: drones, counter-drone systems, AI border surveillance and more. For each, we publish the trial standard and a fixed price per unit, set from independent cost estimates and capped at the import price.

2. Next, hold trials on a clock. Trial windows open every quarter at desert, high-altitude and coastal ranges, with results in 45 to 90 days against published pass marks. Every startup knows exactly what success looks like.

3. Then, buy what passes and pay in 30 days. Any Indian product that passes gets an order within 30 days, and payment lands 30 days after acceptance. One way to fund it is to ring-fence 5% of defence capital spending for field-proven Indian products from firms of any size or age.

4. Keep the kit improving. Drones and electronic-warfare systems are bought with software updates every quarter, so fielded kit stays as fresh as the battlefield.

5. Finally, keep classified work Indian and capital open. Classified work is done in a fully Indian-owned unit (Sovereignty first), while the rest of the company raises money under the one ownership rule in Chapter 10, and the Chapter 16 fund of funds adds Indian capital.

Today vs the new way

Here is how I see the Defence Startup Mission comparing with how we buy today.

Indian inventors get a clear path to orders and fast payment, the forces get better kit, and about 22,000 high-skill jobs follow (estimate), at about the same cost.

Proof it works

Ukraine buys from the front line. Its Brave1 Market lets frontline units order tested drones directly from makers. More than 500,000 drones have been ordered through it.

The United States moves prototypes to production. The Defense Innovation Unit's Commercial Solutions Opening lets a firm whose prototype succeeds move straight to a follow-on production contract.

India's startups already meet the bar. iDEX has cleared 58 startup prototypes worth ₹3,853 crore for procurement, proof that Indian startups can meet the standards of the forces.

What success looks like

• In the first 30 days: the Mission, the 5% earmark, fixed prices and 30-day payment are approved, and trial standards and prices are published for the first 10 priority products.
• Within 1 year: 4 trial windows are held, at least 100 products are trialled and at least 40 orders are placed within 30 days of passing.
• Within 1 year: at least ₹8,000 crore of the earmark is committed, and average payment time is 30 days or less, audited by the Controller General of Defence Accounts.
• Within 1 year: drones bought for the forces reach at least 60% Indian content by value.
• After year 1: Indian-made high-end cameras, thermal sensors and small jet engines take over from the items on the 12-month exemption list.

Guardrail ideas

Ideas to start the conversation, for the builders, experts and the public to shape together.

• Every part comes from a trusted source. Accredited labs audit the bill of materials of every system bought, with a firm lifetime ban for any fraud on origin.
• Trials are fair and open. Each trial panel includes a user officer, a DRDO scientist and an independent expert, with scores published within 7 days.
• Quality is backed by a bond. Every order carries a performance bond, with part of it forfeited for each in-service failure in the first 12 months.
• People decide on lethal force. Every use of lethal force by an autonomous system rests on a human decision, with every engagement logged.

FAQs

Q: Will guaranteed orders waste money on unproven startups?

A: Every rupee buys kit that has already proven itself. Orders follow only a passed field trial, at a price fixed in advance, and any unspent earmark returns to the capital budget at year end.

Q: How does this fit with tender rules?

A: It gives a fair, fixed price to every winner. The price is set before the trial from independent cost estimates, every product that passes is certified at that price, and two or more winners share the first order equally.

Q: What happens to defence public companies?

A: They gain a strong supplier base. They act as integrators, buying 20% of their contract value from new trial-passed suppliers, and they compete as equals in the same trials.

The detailed blueprint for this chapter, a starting draft for the mission heads, industry and the private sector to refine, is in the Implementation Annex.