Solar Cell Gigafactory Risk Profile in India 2026: Diffusion Furnace Fire, Silane Handling and the ALMM List-II Capacity Crunch

From June 1, 2026, modules supplied to government-backed and ALMM-linked projects must use cells sourced from manufacturers on ALMM List-II, the Ministry of New and Renewable Energy's approved list for domestically produced solar cells. The rule was notified in December 2024, so the deadline did not arrive as a surprise. The arithmetic behind it did.

India's ALMM List-I module capacity sits in the range of 190 GW or more, while List-II enlisted cell capacity reached only about 30 GW by the April 2026 revision, including a fresh heterojunction (HJT) line from Reliance. That mismatch, roughly six units of module capacity for every one unit of compliant cell capacity, leaves most standalone module makers without a domestic cell to put inside the laminate. The market response is a rush of cell-line construction. Through 2026 and 2027 a string of TOPCon gigafactory projects has been announced across Gujarat, Madhya Pradesh, Uttar Pradesh and other states, several by groups whose existing footprint was module assembly rather than cell fabrication.

For brokers, the important point is not the policy debate. It is that a wave of high-hazard process plants is being commissioned fast, often by groups whose prior insurance history is a module-assembly shed, not a chemical-grade fab. A module assembly line is essentially stringing, lamination and framing: a moderate fire load and a clean placement. A cell line is a different animal. It runs pyrophoric gas, high-temperature furnaces and concentrated acids in one building. Underwriters who priced the module sheds at module rates will badly misprice the cell lines if they reuse the same mental model. This post sets out where the hazard actually sits and how to place it.

The published solar rooftop installer and rooftop EPC contractor risk profiles deal with installation and balance-of-system work. The solar float glass furnace post deals with one upstream input. A cell gigafactory is upstream of all of these and carries a hazard set closer to a semiconductor fab than to anything in the rooftop chain.

The cell process moves a polished silicon wafer through texturing, diffusion or deposition to form a junction, anti-reflective coating, passivation and metallisation, then test and sort. The TOPCon and HJT architectures that dominate the 2026 build wave lean heavily on plasma-enhanced chemical vapour deposition (PECVD) and high-temperature diffusion or annealing furnaces. New high-throughput PECVD tools entering the market in 2026 process several thousand wafers an hour, which concentrates enormous replacement value and downtime exposure into a single tool string.

Three hazard families sit inside that line and define the risk:

• Pyrophoric and flammable process gas, principally silane, used in deposition, alongside ammonia, nitrogen trifluoride and other specialty gases that carry their own toxicity and reactivity.
• High-temperature furnaces running continuously at several hundred degrees Celsius, a classic ignition and machinery-breakdown source whose internal heating elements, quartz tubes and vacuum systems can also fail without any external fire.
• Wet chemistry using hydrofluoric, nitric and other concentrated acids for texturing and cleaning, a corrosion, toxic-release and personal-injury exposure that pulls liability and workers-compensation considerations into the account.

There is also a cleanroom dimension. Cell yield depends on particulate control, and a smoke or sprinkler-water event that contaminates a cleanroom can write off work-in-progress and force a deep clean even where the structural damage looks minor. That makes the gap between physical repair time and yield-recovery time wider than a conventional factory, which feeds straight into the business-interruption number.

None of these features appears in a module-assembly placement. Treating the cell line as "solar manufacturing" and stopping there is the single most common mispricing error a broker will see this year. The correct reference class is process manufacturing, and the closest published analogue in the Sarvada library is the semiconductor estate, not the rooftop chain. A broker who internalises that one reframing will ask better questions and protect both the rate and the coverage.

Silane (SiH4) is the gas that should reset the underwriter's anchor. It is pyrophoric: it ignites spontaneously on contact with air, with no external spark needed, at concentrations reported as low as the low single-digit percent range and a lower flammable limit near 1.37 percent by volume. Worse for loss control, it burns with an almost invisible flame, so a leak can be burning before anyone sees it.

The loss history is real, not theoretical. India has seen fatal silane-related industrial incidents, and overseas PV and semiconductor plants have lost workers and months of production to silane events. A silane release does not behave like an ordinary flammable-gas leak, which may disperse before finding an ignition source. Silane finds its own ignition source instantly, and a delayed-ignition release can deflagrate. That combination is why this single gas can move a placement from an ordinary property rate to a heavily engineered, surveyed and possibly facultatively-supported one.

Underwriting silane exposure is a gas-handling-systems conversation, not a generic factory conversation. The questions that matter are whether the plant uses gas cabinets with excess-flow valves and automatic isolation, whether cylinders sit in a dedicated detached gas yard with blast relief, the density and placement of silane-specific flame and gas detection, the integrity of the purge and abatement system, and the bund and separation distances between the gas yard and the main process hall. A broker who walks into the renewal with those answers already gathered will hold the pricing conversation. One who treats silane as a footnote will hand the underwriter a reason to load the whole account.

The furnace and deposition tools sit at the intersection of fire and machinery breakdown, which is exactly where Indian property placements tend to leak coverage. A diffusion or annealing furnace running continuously at high temperature is both an ignition source for any nearby flammable and a high-value asset whose own electrical, heating or vacuum systems can fail internally. The first peril belongs to the fire and special-perils section. The second belongs to a machinery-breakdown extension. If the placement carries one and not the other, a furnace burnout that does not produce an external fire can fall into a coverage gap.

PECVD tools add electronic and control-system fragility on top of the thermal load. They embed sensitive controllers, RF generators and vacuum subsystems that are better matched to electronic equipment insurance than to a blanket fire policy. The practical placement question is how the three covers, fire and special perils, machinery breakdown, and electronic equipment, interlock so that a single furnace or tool loss is unambiguously covered under one of them with no argument over which.

Under the material damage section, the reinstatement value basis is non-negotiable for this asset class. A high-throughput PECVD string or diffusion furnace bank is imported, long-lead and frequently quoted in dollars or euros. An indemnity (depreciated) basis would leave the insured unable to replace the tool after a loss. Brokers should also stress-test the sum insured against current landed replacement cost, including installation and commissioning, because a tool ordered eighteen months ago at one price will not reinstate at that figure today. Get the average clause wrong here and underinsurance will bite hard on a partial loss.

On a cell gigafactory, the business interruption exposure usually dwarfs the material damage figure, and the ALMM List-II shortage makes that worse, not better. The logic runs in two directions.

First, the replacement-time problem. The critical tools are imported and long-lead. A serious furnace or PECVD loss can mean a procurement and recommissioning cycle measured in many months, not weeks. The maximum indemnity period has to be set against that real lead time, and 12 months is frequently too short for a tool-destroying loss. Brokers should push for an indemnity period that survives a long-lead reorder, then build in extended-period-of-indemnity thinking for the ramp back to yield.

Second, the scarcity premium. Because compliant domestic cell capacity is so tight relative to module demand, an operating cell line that goes down is not easily substituted. The insured loses gross profit, and in many cases the downstream module business loses its only ALMM-compliant cell source. That can pull in consequential loss arguments and supplier or customer extensions that brokers must scope deliberately rather than leave to standard wording.

Two placement actions follow. One, contingent business interruption and supplier-extension cover deserve explicit attention where a single cell line feeds a captive module plant, because an own-damage event upstream can strand the whole group. The standard fire-policy BI extension will not always respond to a loss that hits an associated company's premises, so the wording needs checking line by line. Two, for projects still in construction, the delay in start-up and advance loss of profits angle matters: a fire or tool loss during commissioning of a plant racing to hit the ALMM window does financial damage well beyond the physical repair, and DSU/ALOP is the only cover that responds to it.

There is a quieter point on the gross-profit definition itself. A cell line ramping toward design yield is not yet earning at full rate, so a loss early in the ramp can be argued down using a trailing turnover that understates the plant's earning power. Brokers should engage with how the BI declaration is built for a plant still climbing the yield curve, and consider whether a turnover or output basis better reflects the real exposure than a backward-looking profit figure. Getting this right at placement avoids a painful argument with the loss adjuster after a claim.

Because this is a build wave, a large share of the 2026 to 2027 exposure is construction-phase, not operational. Erection of furnace banks, gas yards and cleanroom tooling is precisely when first-loss events cluster. The relevant cover is erection all risks within the engineering insurance family, and the way it is structured will decide whether a commissioning loss is recoverable.

The parallels to the semiconductor fab construction-phase placement are direct, and brokers placing cell gigafactories should borrow from that playbook. Testing and commissioning is the highest-hazard window: the first time silane flows, the first furnace ramp, the first wet-bench acid charge. Standard erection-all-risks testing periods are often too short for a process plant of this complexity, so the testing-and-commissioning sub-limit and duration need explicit extension.

The other construction-phase trap is marine and transit. The PECVD strings, furnaces and gas-handling skids are imported, high-value and fragile, and a single shipment can represent a meaningful fraction of the plant's tool value. Marine cargo and inland-transit cover, with a DSU extension that responds when a damaged shipment delays the whole commissioning timeline, belongs in the conversation from day one, not bolted on after the order is placed. Sea-fastening, shock and tilt-indicator conditions, and proper handling clauses for sensitive tools all belong in that cargo wording, because a furnace or PECVD chamber damaged in transit can quietly delay the entire build.

One more construction-phase item is the contractor mix. A cell gigafactory is built by a blend of civil contractors, cleanroom specialists and foreign tool-installation engineers, and the cross-liability and principal-controlled aspects of the engineering programme need to be set so that a fire started by a sub-contractor during fit-out does not trigger a recovery fight between the insured and its own builders. Aligning the erection-all-risks third-party section with the project's contracts at the outset saves a great deal of trouble later.

The underwriting of a cell gigafactory should look like the underwriting of a chemical or semiconductor process plant, and the broker who frames it that way controls the account. The decisive questions are not generic.

1. Silane and process-gas system: detached gas yard with blast relief, excess-flow valves, automatic isolation, silane-specific detection density, purge and abatement integrity, and separation distances to the process hall.
2. Furnace and PECVD protection: detection and suppression around continuously hot tools, electrical and thermal interlocks, and whether machinery-breakdown and electronic-equipment cover sit alongside fire so no furnace loss falls into a gap.
3. Wet chemistry: hydrofluoric and acid storage, bunding, scrubbers, spill control and the toxic-release and liability exposure that flows from it.
4. Compartmentation and fire separation: walls and detection between the gas yard, the high-temperature furnace bay, the wet benches and the warehouse, so one peril does not consume the whole plant.
5. Values and indemnity period: reinstatement-value sums insured stress-tested against current landed tool cost, and a BI indemnity period set against real import lead times.

The broker's edge in 2026 is simple. Many of these plants are being built by groups new to process-grade risk, and many Indian property underwriters have not yet seen a domestic cell line up close. The broker who arrives with a process-risk survey, a clean silane and furnace narrative, and a property and BI structure that mirrors fab placements will set the terms. The one who treats it as another solar account will leave value, and coverage gaps, on the table.