EV Battery Cell Gigafactory Operational Risk Profile in India 2026: Formation Fire, Electrolyte Inventory and the High-PML Property Programme

Two large lithium-ion cell gigafactories, each planned around 40 GWh of capacity, are crossing from construction into operation in India this year. Tata Group's Agratas is advancing its Sanand plant with imported process equipment being installed through 2026, and Reliance Industries has signalled that its battery gigafactory begins operations later in the year, scaling toward a much larger figure over time. These are not the swapping stations, charging depots or recycling yards brokers have already learned to place. They are very large, single-site chemical-process plants making an electrochemically active product.

The placement problem is concentration combined with no loss record. Independent trackers of the Advanced Chemistry Cell (ACC) Production Linked Incentive scheme reported through early 2026 that only a small fraction of the targeted cell capacity had actually been commissioned, with disbursement still lagging the announced ambition. The practical effect for an insurance buyer is that India's cell-making capacity, and therefore its insured property value, is bunched into a handful of first-of-a-kind sites rather than spread across many plants with comparable hazards.

That changes how the risk underwrites. A single declared value at one Sanand or Jamnagar campus can run into several thousand crore once buildings, dry rooms, coating and calendering lines, formation and aging racks, and electrolyte and solvent inventory are added together. There is no Indian gigafactory fire from which to read a frequency. Reinsurers price this off overseas plant experience, engineering surveys and the named perils they fear, not off a domestic burning cost.

A broker who can walk the process front to back will place this account far better than one who treats it as a single factory shed. The line splits into distinct hazard zones, and the property programme should reflect that.

Electrode manufacturing is where slurry is mixed and coated onto foil, then dried and calendered. The cathode side uses N-methyl-2-pyrrolidone (NMP) as the coating solvent. NMP is a combustible liquid with a high boiling point that is recovered and recycled, so the plant runs a continuous solvent loop with vapour, storage and distillation hazards attached. This is a chemical-handling exposure sitting inside what looks like an electronics plant.

Cell assembly happens inside dry rooms held at very low dew points, often around minus 40 degrees Celsius and lower for high-nickel chemistries. The dry room is both a protection feature and a single point of failure: lose dehumidification and you can lose a production campaign to moisture ingress without any fire at all, which is a machinery-breakdown and business-interruption event rather than a material-damage one.

Formation and aging is the zone underwriters fear most. Cells are charged for the first time, filled with electrolyte and held on racks for days. The electrolyte is flammable, the cells are electrochemically live, and a single defective cell can go into thermal runaway and propagate down a rack. International fire engineering consistently identifies cell finishing and formation as the highest-probability ignition source on the line.

A workable zoning for the slip

• Electrode and solvent area: NMP fire and explosion, solvent recovery breakdown.
• Dry rooms: dehumidification failure, contamination of work in progress.
• Formation and aging: thermal runaway, rack-to-rack propagation, the controlling PML scenario.
• Warehousing of finished cells and electrolyte: stored-energy fire load, separation distances.

Write the survey and the sum insured against these zones, not against one blanket figure.

Underwriting this account comes down to one question above all others: how bad does the worst credible formation-area fire get before it is stopped, and how much of the plant and the production schedule does it take with it.

The physics are unforgiving. A lithium-ion cell in thermal runaway vents flammable and toxic gas, generates its own oxygen pathway and resists conventional suppression because water cools but does not always halt the internal reaction. Pack hundreds of live cells onto formation racks in a single hall and you have a fuel load that can self-propagate. The probable maximum loss is therefore not a small bench fire. It is the loss of a formation hall, the contamination of adjacent clean zones by combustion products, and a long outage while charred high-value equipment is replaced on lead times measured in many months from overseas vendors.

This is why the business interruption exposure usually dwarfs the material damage exposure. A formation line is purpose-built and single-sourced. Replacement coaters, calenders and formation cyclers are not held in Indian inventory. An indemnity period of 18 to 24 months is the realistic floor for this class, and many reinsurers will want to see 24 to 36 months before they are comfortable, because the bottleneck is equipment manufacturing slots abroad, not local reconstruction.

The practitioner takeaway is to insist on a fire and explosion engineering report that explicitly models formation-area propagation and the suppression response time, and to let that report, rather than an arbitrary percentage, set the PML on which the reinsurance is structured.

Getting the sum insured right on a first-of-a-kind plant is harder than on a mature manufacturing risk, and getting it wrong invites the average clause at the worst possible moment.

The values are dominated by imported process equipment. Coating, calendering, slitting, stacking, formation and aging machinery is largely procured from Korean, Japanese, Chinese and European vendors, priced in foreign currency, and exposed to exchange rate movement between order and any future reinstatement. A material damage sum insured fixed at historic landed cost will be inadequate if a claim falls two years later after rupee depreciation and equipment price escalation. Build the declared value on current reinstatement cost, refreshed annually, and confirm the basis is reinstatement value rather than indemnity so depreciation is not deducted from a near-new line.

Stock and inventory need their own treatment. Electrolyte, NMP and solvent stocks, plus work in progress sitting in dry rooms and on formation racks, are both a fire load and a value at risk. Declare them honestly, because an under-declared inventory understates the very fuel that drives the PML.

On a gigafactory the largest dependency is rarely the building. It is a single line of imported equipment with a long replacement lead time. Read calculating an adequate sum insured for commercial property alongside the engineering report before you fix the declared value.

For business interruption, the gross profit figure must be projected, not historical, because the plant ramps from low utilisation toward 40 GWh over the policy period. A BI sum insured anchored to a low first-year output will leave the insured badly short once the line is running at scale. Quantify the gross profit on the projected ramp and revisit it at every renewal. The mechanics of getting this right are set out in our guide to business interruption for Indian manufacturers.

The wording on this account will be more conditional than a routine fire policy, and the broker's job is to negotiate workable warranties rather than accept whatever the survey throws up.

Expect the placement to hinge on a risk engineering survey, typically from an international firm, that drives specific protection warranties. The recurring themes are fire detection and suppression in formation and aging halls, gas detection for electrolyte and solvent vapour, separation distances between formation racks and between cell storage stacks, dry-room integrity, and a hot-work and impairment management regime. Each of these can become a warranty or a condition precedent, and a breach can be fatal to a claim, so the broker must confirm the insured can actually maintain compliance day to day before the wording is bound.

Watch the exclusions and sub-limits closely. Insurers will probe contamination and clean-up of combustion residue, defective-cell and serial-defect questions that overlap with product liability, and any attempt to carve out thermal runaway as a distinct peril. The product side matters because a cell defect that causes a downstream EV or storage fire is a product liability and recall exposure, not a property loss, and the two programmes must be read together so a defect-driven event does not fall between them.

Machinery breakdown and electronic equipment cover deserve explicit attention. Dry-room dehumidification, formation cyclers, environmental controls and the plant's own utilities are breakdown-exposed, and a dehumidification failure that ruins work in progress is a machinery-breakdown plus BI claim with no fire at all. The underwriting logic for that cover is closer to process industries than to consumer electronics, and our note on machinery breakdown underwriting for process industries frames the questions to ask. Finally, pin down de-link or aggregation language so material damage, machinery breakdown and BI respond coherently to a single formation fire rather than arguing over which section pays.

A risk this large and this novel does not stay in one insurer's net retention. The broker is really arranging a reinsurance-led programme, and understanding where the capacity sits changes how the submission is built.

Indian insurers will front the policy and cede the bulk of it. GIC Re takes its obligatory cession, but the controlling capacity for a several-thousand-crore first-of-a-kind cell plant comes from facultative reinsurance placed into international markets that have actually underwritten gigafactories abroad. Those reinsurers price off overseas plant experience and their own engineering view, which is exactly why the engineering survey and the formation-area PML modelling are the documents that move terms. A thin submission gets a defensive quote loaded with conditions; a submission with a credible fire engineering report, clear zoning and an honest inventory declaration gets engaged underwriting.

Because capacity is syndicated, the broker should expect higher deductibles and possibly a layered structure rather than a single ground-up policy. A meaningful material damage deductible and a time-based BI deductible aligned to the indemnity period are normal here and are the price of getting the line filled. Trying to force a low deductible on a no-loss-history risk usually costs more in rate than it saves.

What strengthens the submission

1. An international risk engineering survey that models formation-area fire and propagation.
2. Zone-by-zone sums insured on current reinstatement cost, refreshed for currency.
3. A BI projection built on the production ramp, with an 18-to-36-month indemnity period justified by equipment lead times.
4. A protection and impairment compliance plan the insured can demonstrably maintain.
5. Read-across to the product liability and recall programme so a cell defect does not fall into a gap.

For the wider lithium value chain, our profiles on lithium refinery property underwriting and EV battery storage and BESS underwriting give brokers the upstream and downstream context the same reinsurers will be weighing.

It helps to rehearse the claim before the line is even bound, because the recovery mechanics expose every weakness in the placement.

The first test is causation and section response. A formation fire that starts with a single defective cell, propagates down a rack and damages the hall raises an immediate question: is this a material damage fire loss, a machinery-breakdown event, or a product-defect matter? The honest answer is often that it engages more than one section, which is why de-linked but coordinated wordings matter and why a single appointed surveyor who understands the whole programme is worth insisting on at placement.

The second test is contamination and clean-up. Lithium-ion combustion deposits toxic and corrosive residue across clean zones well beyond the seat of the fire. Cells, electrodes and electronics exposed to that residue may have to be scrapped even if they never burned, and the clean-up and decontamination cost can rival the direct fire damage. If the wording sub-limits contamination or excludes pollution clean-up, the insured discovers the gap at the worst time. Negotiate that exposure explicitly at placement.

The third test is the length of the outage and the BI quantification. Because the bottleneck is overseas equipment replacement, the BI claim will run for many months, and the loss adjuster will scrutinise the projected gross profit, the ramp assumptions and any mitigation such as outsourcing or partial running. A plant that has documented its production plan and its single-point dependencies will quantify its loss far more credibly than one that has not. Our walkthrough on business interruption claim quantification sets out the evidence a surveyor expects.

Finally, subrogation. If a fire traces to defective imported equipment or a vendor's design fault, the insurer's recovery rights against that supplier can be significant, and contracts signed during construction will determine whether those rights survive. Brokers should flag this at placement so commercial waivers do not quietly extinguish a future recovery.