Orvakal Drone City Is Being Built Ahead of the Certification Regime. What a Drone OEM Should Insure in the Meantime

The New Indian Express reported on 19 August 2026 that Union Civil Aviation Minister K Rammohan Naidu had announced Orvakal Drone City as India's premier integrated hub for drone manufacturing, research, testing, certification and training. The same report puts potential investments at up to Rs 5,000 crore, with 5,000 direct and 25,000 indirect jobs, all plots in the 600-acre facility to be allotted, and 8 of the first 12 plots covering 55 acres already allotted against Rs 130 crore invested. Manufacturing is expected to begin within two to three years.

The detail that matters to an underwriter sits further down the same report. A special committee drawn from the Civil Aviation Ministry, the DGCA, the Quality Council of India and other agencies has been given a three-month deadline to work through testing, certification, sandbox facilities and airspace permissions. The minister also said the government would focus on BVLOS technology, eGCA 2.0, skill development and pilot training, alongside the Centre's Rs 120-crore PLI scheme and the Drone Rules, with national demand projected to exceed six lakh drones.

The sequencing is unusual. Plots are allotted and capital is committed now, while the certification regime that will define what an airworthy product looks like, and therefore what a defect looks like, is drafted on a three-month clock in parallel. An OEM signing a lease at Orvakal this quarter is buying land for a factory whose output specification is still being written.

Most drone cover sold in India today is operator cover. It is written for a company that flies drones: third-party liability by weight class, hull cover on the airframe, payload liability, and remote pilot exposures, all priced against the Drone Rules and the operator's Digital Sky registrations. Our earlier note on DGCA drone operator insurance walks that structure in detail.

A manufacturer sits elsewhere. The OEM does not fly the aircraft that causes the loss. Its exposure is that an airframe it designed, built and shipped two years ago comes down on a third party, and the claimant, the operator's insurer, or both, come back up the chain. That is a product liability claim, and it is where standard market wording starts working against the buyer. Three mismatches show up in almost every submission we have seen:

1. The aviation exclusion. General product liability wordings carry a broad exclusion for products used in or incorporated into aircraft. Where the wording defines "aircraft" without limitation, an unmanned airframe is caught, and the entire manufacturing exposure sits outside the policy the OEM thinks it bought.
2. Design cover. Many Indian product liability slips confine cover to manufacturing defect and exclude design defect, or restrict design to an errors-and-omissions sub-limit. For a drone maker, design is the exposure. A flight-control law that mishandles a wind gust is a design failure replicated across every unit shipped.
3. Occurrence definition. Product liability sold on a claims-made basis with a short retroactive date does not reach the units already flying. An airframe delivered in the first year of the cluster's operation may still be in service well past the retroactive date on a renewal three years later.

An OEM that buys general product liability with an unqualified aircraft exclusion and no design cover is uninsured for the two loss scenarios most likely to end the business. Check the exclusion wording before the limit.

The fix is not a different policy class. It is a negotiated carve-back to the aviation exclusion plus an affirmative design grant, both written into the slip rather than assumed from the policy wording template.

What the carve-back needs to say, in substance:

• The aircraft exclusion does not apply to unmanned aircraft systems designed, manufactured or supplied by the insured and operated within the categories permitted under the Drone Rules and any airworthiness or type certification issued by the DGCA.
• Cover extends to defect in design, specification, formulation, plans, instructions for use, and software supplied as part of the product, not manufacturing defect alone.
• Airworthiness is defined by reference to the certification in force at the date of supply of the individual unit, not the date of loss. Without this, a rule change mid-service silently strips cover from a fleet that was compliant when it left the factory.
• The definition of "product" captures firmware and over-the-air software updates. A drone whose flight behaviour changes after a remote update is a different product from the one that shipped, and a wording drafted for physical goods will argue that point at claim stage.

The last two points are the ones brokers most often let slide. They are also the ones a three-month certification sprint makes urgent. If the DGCA and the Quality Council land a testing and certification regime this quarter, every airframe designed before it exists will be assessed after the fact against a standard it was not built to. The date-of-supply anchor is what stops that becoming a coverage argument.

Standard product recall cover in the Indian market responds to an accidental defect that makes the product unfit or unsafe, usually with a bodily injury or property damage trigger. That trigger is the wrong shape for this cluster.

The realistic recall scenario at Orvakal is regulatory. The special committee publishes a certification standard. An airframe already in service does not meet it, either because a detect-and-avoid capability becomes mandatory for the operating category, or because a component sourcing or software assurance requirement is added. No unit has failed. No injury has occurred. The DGCA nonetheless restricts the type, and the OEM has to retrofit or repurchase a shipped fleet.

That is a regulatory recall or government-mandated recall trigger, and it has to be bought explicitly. Points to press:

• The trigger should be a written order, direction or condition imposed by the DGCA or a competent authority affecting the type or category, not only a finding of defect.
• Costs covered should include retrieval logistics, retrofit labour and parts, replacement airframes where retrofit is not feasible, and the OEM's own cost of the airworthiness re-submission.
• Third-party recall liability matters if the OEM supplies subassemblies into another manufacturer's airframe. The primary recall cost then sits with the customer and comes back as a claim.

Manufacturing at Orvakal is expected to begin within two to three years. Between now and then the exposure is construction and installation, not product. An OEM that only starts thinking about insurance when the first airframe ships will have run the most capital-intensive phase of its life bare.

The build-out stack:

• Erection All Risks (EAR) for the plant and the installation of production and test equipment. See our primer on erection all risks for how the cover sits across the testing and commissioning period. Composite manufacturing, autoclaves, environmental test chambers and anechoic or EMI test facilities are all installation risks with their own commissioning failure modes.
• Marine cargo and delay in start-up on imported capital equipment. Drone production lines pull in imported test rigs, motion systems and inspection equipment. A marine cargo loss on a single long-lead item pushes the whole commissioning date, and the financial consequence sits in delay in start-up cover rather than in the cargo policy.
• Stock throughput once production begins. Drone OEMs hold high-value, low-weight inventory: batteries, flight controllers, gimbals, sensors, RF modules. Lithium batteries in bulk storage are a fire risk with their own warranties and their own storage conditions, and a conventional fire policy declaration basis handles a fast-turning imported inventory badly.
• Machinery breakdown on the test infrastructure. In a cluster whose stated purpose includes testing and certification, the test bench is production equipment. Its failure stops airworthiness submissions, not just production.

The minister named BVLOS specifically as a government focus. It is also the single biggest coverage gap in the drone book today.

Operator policies in the Indian market are written and rated against the Drone Rules as they apply to line-of-sight operation, assuming a remote pilot who can see the aircraft and intervene. Once the aircraft flies beyond visual line of sight over third parties, the loss distribution changes in three ways at once: the exposed third-party population is larger and unknown, the pilot's ability to prevent an accident is mediated by a command-and-control link that can fail, and the detect-and-avoid function moves from human eyes to a sensor and software stack that the OEM built.

That last shift is the one that moves risk onto the manufacturer's balance sheet. In visual operation, an accident is usually pilot error and the operator's third-party liability cover answers. In BVLOS, an accident is often a failure of the detect-and-avoid system, which is a product, and the claim routes to the OEM.

What an OEM should require of its customers, contractually:

1. Operator third-party liability specifically endorsed for BVLOS operation, not a VLOS policy assumed to extend.
2. The OEM named as additional insured on the operator's liability policy for vicarious exposures arising out of the operator's flying.
3. A waiver of subrogation in favour of the OEM from the operator's insurers, subject to the OEM's own products cover responding to genuine product defect.
4. Evidence of the operator's DGCA permissions for the specific airspace and operating category, refreshed at each renewal, so that an operator flying outside its permission does not drag the OEM into an uninsured claim.

There is no credible Indian loss experience for BVLOS product failure, because the operating mode is new and the fleet is small. Limits therefore have to be built from exposure. A workable method for an OEM at Orvakal:

1. Start from the worst credible single event. One airframe of the OEM's heaviest class, at its maximum permitted operating altitude, coming down on a populated area during a BVLOS mission. Estimate third-party bodily injury and property damage on Indian court award benchmarks for comparable aviation and public liability claims, and treat that as the single-occurrence anchor.
2. Multiply for the production run, not for the unit. A design or software defect is not one event. Size the aggregate against the number of units of the affected type in service, since one flight-control law reaches every airframe carrying it.
3. Add the recall cost separately. Retrofit or repurchase of a shipped fleet is a balance-sheet number the OEM can calculate precisely from its own bill of materials and unit count. It does not need to be estimated.
4. Layer. No single Indian insurer will write the aggregate an OEM needs on an untested exposure. Expect a primary of Rs 10 crore to Rs 25 crore from a domestic insurer with excess layers placed into the international aviation market through reinsurance facilities.
5. Reprice annually against the certification regime. As the DGCA and Quality Council standard lands, the OEM should expect the design carve-back to become cheaper once compliance is demonstrable.

The commercial context supports the effort. Analytics India Magazine reported on 11 August 2026 that ideaForge posted Q1 FY27 revenue of Rs 68.6 crore and now has a drone taking off every two minutes, while Inc42 reported on 10 August that the same quarter took the company to a Rs 2.3 crore loss. Financial Express reported on 18 August that six drone stocks were in focus as the Centre fast-tracked Rs 52,000 crore of defence orders covering jammers, radars and anti-drone systems. Revenue growth on thin or negative margins, against a large public order book, is exactly the profile where a single uninsured product recall is existential rather than expensive.

The three-month committee deadline is the planning horizon. An OEM that lands its cover after the certification standard publishes will be buying into a market that has already repriced. The order of work:

1. Now, before ground-breaking. EAR on the plant and equipment installation, marine cargo with delay in start-up on imported capital equipment, and contractors' liability. These attach on the first equipment purchase order.
2. Now, on the design side. Get the products liability carve-back agreed in principle while the airframe is still on the bench. It is far easier to negotiate design cover on a product under development, where the OEM can show its verification and validation regime, than after a fleet is flying.
3. At first delivery. Products liability incepts, with the aviation carve-back, design grant, date-of-supply airworthiness anchor and software-inclusive product definition all in the slip.
4. At BVLOS type approval. Aviation products liability replaces or sits above the general products slip, and the customer contract terms on operator cover, additional insured status and waiver of subrogation become conditions of sale.
5. Continuously. A register of every unit shipped, its configuration, its firmware version and its certification basis. This is the document that decides a recall's cost and a products claim's outcome, and no insurer will build it for the OEM.

Our manufacturers' guide to product liability in India covers the base wording issues that apply to any manufacturer. The drone-specific work sits on top of that foundation, not instead of it.