Green Hydrogen and Electrolyser Startups: Insuring Performance Warranties and First-of-a-Kind Plants Past the SECI Tender Stage

For most of the last three years, a green hydrogen founder in India spent insurance budget on the wrong things. The instinct was to protect the prototype: the stack, the test rig, the patents, the small team. That made sense while the company was burning grant money and proving a cell could split water without degrading. It stops making sense the day the company signs an offtake with the Solar Energy Corporation of India (SECI).

The tender activity through 2025 and into 2026 has been heavy. Under the Strategic Interventions for Green Hydrogen Transition (SIGHT) scheme, SECI ran green ammonia supply tenders running to several hundred thousand tonnes per annum, and offtake agreements covering a large share of that volume have been signed by a mix of climate-tech and established energy players. Individual winners committed to facilities producing hundreds of thousands of tonnes a year, typically anchored to fertiliser offtake, so the contracted volumes are large enough that a missed delivery date carries real penalties.

The moment that ink dries, the risk profile inverts. The startup is no longer exposed to whether the technology works in principle. It is exposed to a written promise that a specific plant will produce a specific volume of hydrogen at a specific energy efficiency, on a specific date, for a specific buyer, with liquidated damages if it does not. That is a performance guarantee, and it sits inside a project finance structure that lenders will not fund unless the guarantee is credibly backstopped.

Brokers who treat this as a slightly larger property programme will lose the account, or worse, place cover that pays nothing when the plant underperforms. The exposure that decides whether the company survives the scale-up is contractual, and the structure that backstops it is still uncharted territory for most Indian intermediaries.

A first-of-a-kind (FOAK) green hydrogen or green ammonia plant carries a risk shape that standard engineering underwriting was not built for. Understanding the layers is the difference between a placeable programme and a declined one.

The construction phase looks familiar but behaves differently. Erection All Risks (EAR) covers the physical build, but the electrolyser stacks, balance-of-plant, rectifiers and the ammonia synthesis loop are often imported, partly prototype, and have no loss history an underwriter can price against. A stack failure during commissioning is not a simple machinery claim; it can stall the entire plant ramp and trip the offtake timeline.

The commissioning and testing phase is where FOAK plants actually fail. The technology may work at pilot scale and behave unpredictably at commercial scale. Hydrogen embrittlement, membrane degradation, rectifier harmonics and ammonia loop instability tend to surface here. Many standard EAR wordings restrict or exclude cover during extended testing of prototype plant, which is precisely when this asset is most fragile.

The performance phase introduces an exposure that has nothing to do with damage. The plant can be entirely undamaged and still breach its guarantee because the electrolyser consumes more kWh per kg of hydrogen than promised, or the stack degrades faster than the warranty curve, or the renewable power feeding it is curtailed. None of that is a property loss. It is a shortfall against a number in a contract.

For brokers, the practical takeaway is that you are insuring three different animals under one project name:

• A construction asset (physical damage, EAR territory)
• A commissioning gauntlet (testing and prototype-plant cover, often the gap)
• A contractual promise (performance warranty, surety and delay cover)

Most declined or mispriced hydrogen placements fail because the broker priced the first animal and ignored the other two.

The construction-phase placement is the foundation, and on a FOAK plant the standard template needs deliberate surgery.

Start with the sum insured logic. The contract value of an electrolyser plant is dominated by imported, high-value, long-lead equipment. Reinstatement on these items means re-import, which means foreign exchange exposure, fresh customs duty, and lead times that can run many months. If the sum insured reflects only the landed cost at order date, a mid-project loss leaves the startup catastrophically underinsured. Insist on an escalation provision and a realistic view of replacement at today's prices, not the purchase order.

Delay in Start-Up (DSU), also written as Advance Loss of Profits (ALOP), is the piece that actually protects the offtake. If a covered EAR loss pushes commercial operation past the SECI-contracted supply date, the startup faces liquidated damages and lost SIGHT production incentives, which step down from around Rs 8.82 per kg in year one to Rs 7.06 and then Rs 5.30 over the first three years. DSU is what converts a physical loss into protection for the revenue and incentive timeline.

Getting DSU right on a hydrogen plant requires care on three points:

1. The indemnity period must extend beyond the planned commercial operation date with genuine buffer, because re-importing a failed stack is slow.
2. The insured gross profit or standing-charges basis must capture the SIGHT incentive, not only the merchant ammonia price, or the policy under-indemnifies the real loss.
3. The DSU trigger must align with the EAR material-damage proviso; a gap between the two is the most common reason a delay claim fails.

A broker who builds a clean EAR plus DSU spine, with the indemnity period and incentive basis correctly set, has solved roughly half the problem. The harder half is the performance promise itself.

The handover from construction cover to operational cover is where hydrogen startups quietly lose protection, and where a sharp broker earns the fee.

Once the plant is commissioned, EAR falls away and operational covers take over: property or fire on the material damage, and machinery breakdown (MB) on the rotating and electrical plant. The problem is the seam between them. Electrolyser stacks, rectifiers, compressors and the ammonia synthesis train are exactly the items most likely to fail in the first operating years, and they sit at the boundary where insurers are most cautious.

Three underwriting frictions show up repeatedly:

Proto-typical plant restrictions. Many MB wordings carve out or sub-limit equipment that is prototypical or lacks an established loss record. An electrolyser at Indian commercial scale frequently meets that description, so the broker must negotiate the definition rather than accept the printed exclusion.

The warranty-versus-insurance overlap. The OEM gives a performance and degradation warranty on the stack. Insurers will not pay for what the warranty should cover, and will push gradual deterioration, wear and tear, and design defect into exclusion territory. The broker has to map precisely where the OEM warranty ends and where insurable sudden-and-accidental damage begins, then place cover for the genuine gap, not the whole stack.

Consequential loss from a breakdown. An MB event that takes the stack offline does not only cost the repair. It interrupts hydrogen and ammonia supply, which interrupts the offtake. Machinery Loss of Profit (MLOP) cover, the operational cousin of DSU, is what bridges a breakdown into the revenue and incentive shortfall. Few first-time founders ask for it, and many brokers forget to offer it.

The practical move is to run an explicit boundary exercise with the founder and the OEM in the room: what does the stack warranty pay, for how long, against what degradation curve, and what is left over for insurance to carry. Done properly, this exercise also surfaces counterparty risk on the OEM warranty itself, which feeds directly into the performance-guarantee structure discussed next.

This is the part that decides whether the company is financeable, and it is the part Indian brokers are least practised at structuring.

When the startup signs the offtake, it makes two kinds of binding promise to SECI or the downstream buyer. The first is delivery: the plant will be built and supplying by a contracted date. The second is performance: the plant will deliver the contracted volume at the contracted quality. Each promise is usually secured by a bank guarantee or a bond, and tying up that guarantee capacity strains a young company's balance sheet at exactly the moment it needs liquidity for the build.

This is where surety bonds become genuinely relevant. The IRDAI surety bonds framework, refreshed for the Indian market, lets insurers issue surety in place of bank guarantees for performance and contract obligations. For a hydrogen startup, a surety solution can in principle free up banking lines that would otherwise be locked behind the offtake guarantee. The catch is that surety underwriting looks hard at the obligor's ability to actually perform, so a FOAK plant with no operating track record is a tough risk. The realistic 2026 position is that surety is available for the more bankable promises (timely construction, milestone delivery) and far harder for the pure technical-performance guarantee.

The technical-performance guarantee itself sits awkwardly between products. A true efficiency or output shortfall, where the plant is undamaged but consumes too much power per kg, is closer to a warranty or financial guarantee than to insurance, and most Indian property and engineering insurers will not write it. The credible structure layers the promise:

• OEM performance and degradation warranties carry the technology risk, backed by the OEM's balance sheet.
• Surety or contract bonds carry the delivery and milestone risk where the obligor is bankable.
• EAR, DSU, MB and MLOP carry the damage-driven causes of shortfall.
• A negotiated efficiency or serial-defect endorsement, where an insurer is willing, carries a defined slice of the residual technical risk.

The broker's job is not to find one magic policy. It is to assemble this stack so that, read together, the lender's risk committee sees the offtake guarantee as credibly supported. That assembly, more than any single wording, is the value a specialist intermediary brings to a hydrogen account.

A founder who walks into the market expecting hydrogen cover to be priced like a textile shed will be unpleasantly surprised, and a broker who does not set that expectation early will lose credibility fast.

Capacity is the first constraint. FOAK green hydrogen and green ammonia plants are large, technically novel, and have almost no Indian loss history, so domestic insurers lean heavily on reinsurance, with GIC Re and international treaty markets effectively setting the terms. Lead lines on a several-thousand-crore plant typically come together through a co-insurance arrangement, and the placement can take real time to build. Founders working to a tender timeline should start the insurance conversation months before financial close, not weeks.

Pricing reflects the uncertainty. Rates on FOAK energy-transition assets sit well above conventional process plant, and underwriters defend that with deductibles that bite. Expect meaningful self-insured retentions on the electrolyser and stack items, time-based excesses on DSU and MLOP measured in weeks rather than days, and sub-limits on prototype or testing-phase exposures. The deductible structure is itself a negotiation, and a broker who arrives with strong risk-engineering documentation can move it.

Three levers genuinely move price and terms in this market:

• Quality of the engineering and HAZOP documentation. Hydrogen is a flammable, leak-prone, embrittling gas, and underwriters reward a serious safety case. A weak fire and explosion study invites loadings or declinature.
• Credibility of the OEM warranty chain. If the stack warranty is strong and the OEM is a substantial counterparty, insurers carry less residual risk and price accordingly.
• A staged, layered submission. Presenting EAR, DSU, MB, MLOP and the surety story as one coherent structure, rather than a pile of disconnected requests, lets the lead underwriter see a managed risk rather than a science experiment.

The honest message to the founder is that 2026 capacity exists but is selective and disciplined. The plants that place well are the ones whose promoters treated insurability as an engineering deliverable from the design stage, not a procurement afterthought once the tender was already won.

The test of all this structuring is whether it pays when something goes wrong, and on hydrogen plants the failure modes are unusually contentious.

The most disputed claims will turn on causation. Was the supply shortfall caused by an insured physical loss to the stack (DSU or MLOP territory), by gradual degradation the OEM warranty should carry, by a design defect that insurers exclude, or by grid curtailment that no property policy touches. On a complex electrochemical plant these causes overlap, and the proximate cause analysis decides who pays. This is why the boundary exercise between warranty and insurance has to happen at placement, in writing, not be argued after a loss.

The second realistic battleground is the testing and commissioning window. If a stack fails during extended hot commissioning and the EAR wording restricts prototype testing cover, the founder can find the single most likely loss event sitting in a coverage gap. Brokers should stress-test the testing clause specifically, and where possible negotiate explicit cover for the commissioning of first-of-a-kind plant.

For the broker advising a green hydrogen or electrolyser startup in 2026, the action list is concrete:

• Read the offtake contract before you design the programme. The liquidated-damages clause, the performance test protocol and the guaranteed delivery date define the insurance you actually need.
• Run the warranty-versus-insurance boundary exercise with the OEM and the founder, and document where each carries the risk.
• Build EAR and DSU as the spine, with the indemnity period and the SIGHT incentive correctly reflected in the loss basis.
• Bridge construction to operations cleanly, closing the seam into MB and MLOP so the most failure-prone equipment is never uninsured.
• Test whether surety can free up banking lines on the bankable promises, and be honest that the pure technical-performance guarantee will mostly sit with OEM warranties.
• Start months ahead of financial close, because reinsurance-led capacity is slow to assemble.

The firms that cross the valley of scale will be the ones whose brokers treated the offtake guarantee, not the prototype, as the thing to insure.