Over 300 EVs Burnt in One Bengaluru Yard. Nobody Priced That Yard as a Single Location

Deccan Herald reported on 12 August 2026 that a fire in a parking yard in Bengaluru's Ramamurthy Nagar gutted over 300 electric vehicles. One compound, one ignition event, one loss date.

That sentence describes an accumulation loss. The insurance arrangements sitting behind it almost certainly did not. Shared-mobility and last-mile operators buy motor cover the way every fleet in India buys it: vehicle by vehicle, each with its own registration number, its own Insured Declared Value, its own own-damage section and its own deductible. The policy schedule is a list of assets. It is not a statement of where those assets sleep.

An operator running 3,000 vehicles across a city does not hold 3,000 independent risks overnight. It holds perhaps eight to fifteen risks, because the vehicles are gathered into eight to fifteen leased compounds for charging and dispatch. Between roughly 11 pm and 5 am, the entire fleet is stacked into a handful of postcodes. Nothing in a per-vehicle motor schedule tells an underwriter that, and nothing in the pricing reflects it.

The timing makes this worse rather than academic. Livemint reported on 12 August 2026 that Yulu had raised a $93 million round of equity and debt for expansion and new business lines, and Zag Daily reported on 13 August 2026 that the round was intended to quadruple the company's e-mobility delivery fleet in India. CNBC-TV18 reported on 20 August 2026 that Mahindra Last Mile Mobility had crossed 4 lakh EV sales and was targeting 10 lakh electric vehicles by 2031. The vehicle count is rising quickly. The number of yards available to park them in a city like Bengaluru is not.

A property fire that destroys a warehouse produces one claim. A yard fire that destroys 300 vehicles insured under motor own damage produces 300 claims, and the difference is not cosmetic.

Each vehicle carries its own intimation, its own registration certificate, its own policy line and its own damage assessment. Under Section 64UM of the Insurance Act, 1938, losses above the prescribed threshold cannot be settled without a licensed surveyor's report, and a burnt-out EV is never below that threshold. The surveyor has to establish, for each vehicle, that this specific registration number was in this specific yard on the night in question and is now a total loss. That is 300 separate evidentiary exercises conducted on a debris field where number plates have melted.

What actually decides the file is identification, not quantum. Chassis numbers stamped on the frame survive a fire more often than plates do, and battery pack serial numbers may be recoverable from the pack casing. Neither is much use unless the operator can produce, from its own systems, a list of exactly which vehicle IDs were parked in that compound on that date, and unless vehicle ID maps cleanly to registration number and to policy line.

Expect the file count itself to become the bottleneck. Surveyor capacity in one city is finite, insurers will want a joint inspection before debris is moved, and municipal and fire-services clearance controls when the site can be entered at all. An operator planning around a four-week settlement on a 300-vehicle event is planning optimistically.

The own-damage settlement on a destroyed vehicle is not its replacement cost. It is the Insured Declared Value at policy inception, which is the manufacturer's listed selling price of that model less a depreciation percentage tied to vehicle age, and it is then reduced further by the applicable deductible. On a single vehicle this is a familiar haircut. Applied simultaneously to an entire yard, it becomes a balance-sheet event.

Run the arithmetic in the shape it actually arrives:

1. Each vehicle settles at its own IDV, so a yard holding a mix of two-year-old and four-month-old units settles at a mix of depreciated values, not at fleet average.
2. Each vehicle absorbs its own deductible, so the deductible is charged 300 times rather than once. A per-event deductible would be charged once. Motor own damage has no per-event concept.
3. Vehicles under hypothecation settle with the financier's interest noted, so part of the money may never reach the operator's account at all.
4. Replacement vehicles must be bought at current prices while settlements arrive at depreciated values, and the gap is funded from working capital.
5. Every day a replaced vehicle is not on road is revenue lost, and motor own damage does not pay for downtime.

That last point is the one operators consistently underestimate. There is no business-interruption element inside a motor package policy. Business interruption cover attaches to material damage under a property policy, and a vehicle destroyed under motor own damage produces no loss-of-earnings indemnity at all unless the operator has bought something separately for it.

The cash position, then, looks like this: outflow starts immediately, in the form of replacement vehicle orders and continued yard rent and driver retention costs, while inflow arrives file by file over months, net of 300 deductibles and depreciation. An operator with thin working capital can be solvent on paper and stalled in practice.

The second question after the motor question is whether the compound itself was insured, and by whom.

Last-mile and shared-mobility yards are usually leased plots rather than owned premises. Many are open ground with a boundary wall or fencing, a site office, a metering point and a run of charging bays. In insurance terms this is a location, and a location has to appear on a policy schedule with an address before anything at it is covered. The recurring failure is that the operator's fire policy, if it holds one, schedules the head office and perhaps a central workshop, while the overnight yards leased on short tenures were never added by endorsement.

A standard Indian fire wording responds to property at the situations described in the schedule. Property at an unnamed location is not covered, and no amount of arguing about intent fixes that after the event. Where an operator expands quickly, the endorsement discipline usually lags the property team by a quarter or more, which is precisely the period in which new yards fill up with vehicles.

There is a second gap behind the first. Even where the yard is scheduled, the sum insured is often set against the visible fixed assets, the site office, the charging infrastructure, the transformer and the fencing. The vehicles standing on the ground are far more valuable than everything else at that address combined, and they are usually not contemplated at all in the property sum insured because the operator assumed motor cover handled them. Motor cover does handle the vehicles. It handles them one at a time, with no aggregate, which is exactly the problem.

Yard addresses are what a property underwriter needs first, ahead of vehicle counts. Before the next renewal, produce a list of every compound the fleet uses overnight, with its address, its lease status, its charging capacity and its peak overnight vehicle count. Most operators discover during this exercise that they cannot name all their yards.

Suppose the yard is scheduled and the operator wants the vehicles themselves considered under property cover. The classification question then decides the outcome, and it is genuinely unsettled in practice.

A vehicle held for hire is not stock in trade, because the operator is not selling it. It is not conventional plant and machinery either, since plant sits at a fixed situation while these assets leave the premises every morning and return every night. Registered vehicles are also commonly excluded from property wordings on the reasoning that motor insurance is their proper home, which is the same reasoning that produced the per-vehicle structure with no location aggregate in the first place.

The practical consequences are worth stating plainly:

• Property cover on vehicles at a yard, where it can be arranged, generally needs a specific extension naming the class of vehicle and the situations at which they are held.
• Where such an extension exists, it typically carries its own sub-limit and its own basis of settlement, which may differ from the motor IDV basis.
• Double cover on the same asset raises contribution between the property insurer and the motor insurer, so the extension has to be drafted with the motor policy in view rather than in isolation.
• Vehicle dealers face a related version of this problem with unregistered stock, discussed in our note on flooded commercial fleets and dealer stock.

The honest position for most operators today is that neither policy was written with a 300-vehicle single-location fire in mind. Motor own damage will pay, slowly and per vehicle, and the property policy will contribute little or nothing. Fixing that requires a deliberate conversation at placement, not a claim-stage argument.

The spacing assumptions used for vehicle parking come from a world of petrol and diesel. A liquid-fuel vehicle fire has a fuel tank with a finite quantity, an ignition sequence that usually gives some warning, and a plume that responds to water. Rows of vehicles a metre apart, with an access lane every few rows, reflect that behaviour.

Lithium-ion propagation behaves differently. A cell in thermal runaway generates its own heat and its own oxidiser, which is why the fire continues without an external air supply and why water functions mainly as a cooling measure rather than as an extinguishing one. Heat radiating from one pack raises the temperature of the pack in the next vehicle, and the failure walks down the row. On tightly packed two-wheelers and three-wheelers, where vehicles stand handlebar to handlebar to maximise yard capacity, there is no separation for the propagation to cross. The economics of yard rent push directly against the physics of separation.

Charging compounds the exposure because it concentrates energy and electrical faults in the same place. A yard where 300 vehicles charge overnight on a shared bank is an accumulation of stored energy, of charging electronics and of cabling, all within the same boundary wall. Our earlier piece on EV battery fire claims covers the investigation side of a single-vehicle event; the multi-vehicle case adds the question of which vehicle started it, which matters for subrogation against a manufacturer or a charger supplier and rarely gets answered on a site burnt to that extent.

The risk-engineering implications are unglamorous and specific. Physical separation between charging bays and parked rows. Blocks of limited vehicle count with real fire lanes between them rather than nominal ones. Charging on a separate electrical circuit with monitoring. A yard population cap that is set by fire separation rather than by how many vehicles physically fit.

The pressure is seasonal and it is arriving now. Moneycontrol reported on 19 August 2026 that quick-commerce firms were bracing for a festive worker crunch as hiring demand surged, and The Economic Times reported on 17 August 2026 that festive-season hiring was expected to generate up to 2.7 lakh gig jobs, concentrated in e-commerce and retail delivery. More riders means more vehicles, and more vehicles means yards running above their normal overnight population for eight to ten weeks.

Work the placement before the vehicles arrive, not after:

1. Declare every yard address. Every compound used for overnight parking, with lease status and duration, added to the property schedule by endorsement as it comes into use rather than at the next renewal.
2. Declare peak overnight vehicle count per yard, not average. The exposure is the maximum population on the worst night of the festive peak, not the twelve-month mean a fleet report produces by default.
3. Ask for a per-location aggregate and price it. A named cap per yard, sitting above or alongside the per-vehicle motor cover, is the only structure that responds to a single-event, multi-hundred-vehicle loss as one event. If no insurer will write it at a sensible price, that is itself information about how the exposure is viewed.
4. Set the cap against the real number. A cap equal to the summed IDV of the peak overnight population at the largest yard is the honest figure. A cap set at a comfortable round number that no one has reconciled to a vehicle count is decoration.
5. Separate charging from parking in the yard layout, and record the layout, because the underwriter's questions after Ramamurthy Nagar will be about separation distances and charging bank location.
6. Build the yard-population export now. Telematics, gate logs and charging records reconciled to registration numbers, retained daily. This is the evidence pack that determines how fast 300 files move.

The wider discipline here is the one multi-site corporates already apply to natural catastrophe exposure, set out in our note on accumulation management across multiple locations. Fleet operators have the same problem in a different shape. Their locations move every morning and reassemble every night, and the reassembly is the part nobody has priced.