Draft SHANTI Rules 2026: The Liability Stack an Indian Nuclear Supplier Needs Before It Bids

In mid-August 2026 the government published draft rules operationalising the SHANTI Act, and the reporting around them settled on one structural point. The Economic Times on 17 August 2026 described the draft as proposing a licensing and liability framework for nuclear operators. Business Standard on 16 August 2026 put the operative obligation more plainly: the new rules mandate financial security for plants. The Indian Express on 21 August 2026 added the feature that changes how any long-dated contract should be priced, reporting that liability caps are to be reviewed on a five-year cycle under the draft rules.

Read together, three things are being proposed at once. An operator must be licensed. An operator must hold insurance or equivalent financial security against its liability. And the number that security is sized against is not fixed for the life of a plant, because it comes up for periodic review.

For an operator this is a treasury and compliance question. For an EPC contractor, a reactor-component fabricator or a control-systems vendor bidding into the build-out that the December 2025 nuclear energy Bill opened to private capital, it is a pricing question, and it arrives before the contract is signed rather than after. The Hindu reported on 17 December 2025 that the Lok Sabha passed the Bill allowing privatisation, and the draft rules are the first document that tells a private consortium what its financial obligations will actually look like.

The gap this post fills is the flow-down. Almost everything published so far describes what the operator must do. Very little describes what lands on the supplier's balance sheet as a result, and which of those exposures the Indian market will actually write.

Under the Civil Liability for Nuclear Damage Act, 2010, supplier exposure ran through Section 17. The operator carried strict, no-fault liability for nuclear damage, capped under Section 6(2) at Rs 1,500 crore for reactors above 10 MW thermal, and then held a right of recourse against the supplier. Section 17(a) covered recourse expressly provided for in the contract. Section 17(b), the clause that made global vendors refuse to bid in India for a decade, gave the operator statutory recourse where the incident resulted from an act of the supplier or its employee, including supply of equipment or material with patent or latent defects or sub-standard services.

The SHANTI Act reworked that channel, and the draft rules are where the practical form of it becomes visible. A supplier reading the final rules should be looking for answers to four specific things, not for a general reassurance:

1. Whether recourse exists only where a written contract provides for it, or whether a statutory route survives alongside it.
2. Whether recourse is time-limited, and if so whether the clock runs from the incident, from the operator's payment of compensation, or from commissioning.
3. Whether the recourse amount is capped, and whether that cap attaches to the supply contract value or to the operator's own liability limit.
4. Whether the licensing conditions imposed on the operator require it to pass specified terms down the contract chain, which would make the flow-down non-negotiable rather than a matter of commercial bargaining.

That fourth point is the one bid teams underestimate. If a licence condition requires the operator to preserve recourse against suppliers as a condition of holding its licence, the operator cannot waive recourse in your contract even if it wants to. Every liability cap you negotiate then has to be a cap on quantum, not an exclusion of the right itself.

Our earlier note on nuclear supplier and vendor liability under the CLNDA and the SHANTI Act walks the recourse mechanics in more detail. The draft rules do not change the underwriting logic there; they change the inputs.

The phrase carrying the most weight in the draft is insurance or equivalent financial security. It is a familiar formula in nuclear liability regimes, and it usually means the regulator will accept something other than a policy provided the money is genuinely available when a claim crystallises. The instruments that typically compete for that role are not equally good, and they are not equally good for the supplier either.

• An India Nuclear Insurance Pool policy. The cleanest option, because it is an insurance contract issued by a licensed insurer, the regulator understands it, and the claims-handling machinery already exists. The India Nuclear Insurance Pool (INIP) was launched in June 2015 by GIC Re with eleven domestic non-life insurers, at an initial capacity of Rs 1,500 crore, with GIC Re as pool manager and New India Assurance as the policy-issuing member. Capacity above that level depends on reinsurance support.
• A bank guarantee. Certain in payment, expensive in balance-sheet terms because it consumes the operator's credit lines, and awkward on duration because guarantees are renewed while nuclear liability tails run for years.
• A parent-company indemnity. Cheap, and only as good as the parent's covenant. For a special-purpose vehicle formed by a private consortium, the regulator has to be willing to look through to a parent that may sit outside Indian jurisdiction.
• A captive. Viable for a large industrial group, but a captive fronting a statutory liability needs reinsurance behind it, and the reinsurance market for Indian nuclear liability is narrow.

For a supplier, the important question is not which instrument the operator picks. It is which one the operator's recourse claim would be funded from, and whether that instrument's terms give the funder subrogation rights against you. An insurer that pays an operator's liability claim will look at recourse as a recovery route. So will a bank that has honoured a guarantee. A parent indemnity is more likely to be settled commercially. The choice of instrument therefore changes the probability that a claim actually reaches your defence, independent of the legal position on recourse.

Ask for a copy of the operator's financial-security instrument at the bid stage, or at minimum a description of its type and limit. It is not a confidential number in most tenders, and it tells you more about your realistic exposure than the indemnity clause does.

Capacity is the constraint that a licensing framework cannot legislate away. INIP was built at Rs 1,500 crore, matched to the CLNDA Section 6(2) operator cap and to a single state-owned operator, NPCIL, running a defined fleet. A privatised build-out changes three variables at once: more operators, more sites, and more simultaneous construction risk sitting alongside operating risk.

Three practical consequences follow.

Aggregation. A pool's capacity is not per-policy in the way a buyer intuitively assumes. Members allocate capital to the pool and that allocation is finite across all risks written. Adding private operators to the same pool does not add capacity unless members add capital or the pool buys retrocession.

The five-year review. If liability caps are reviewed every five years, as The Indian Express reported on 21 August 2026, the required financial security can be revised upward mid-life on a plant that will run for sixty years. An operator whose security is a pool policy needs the pool to have grown by then.

Foreign participation. The New Indian Express reported on 21 August 2026 that the US nuclear industry is preparing detailed submissions to India on the SHANTI Act. Those submissions will be about supplier liability, because that is what has kept US vendors out. If the final rules satisfy them, capacity demand rises faster than it otherwise would.

None of this is a reason for a supplier to wait. It is a reason to assume that the operator's security is a shared resource under pressure, and that recourse against suppliers becomes more attractive to a funder, not less, as the pool tightens. Our nuclear power plant operator risk profile covers the operator side of the same capacity question.

On 17 August 2026 The Economic Times reported that the Supreme Court has asked the Centre to respond on whether courts are precluded from granting fair and just compensation in the event of a nuclear accident. The question goes to the heart of a capped-liability regime. A statutory cap works only if it is the ceiling on what a claimant can actually recover. If a court retains the power to award compensation beyond the cap, the cap becomes a floor for the insurance programme rather than a limit on exposure.

That is not an abstract point for a supplier. Three consequences follow directly if the cap is judicially reopenable:

1. The operator's exposure above the cap has to be funded from somewhere, and recourse against suppliers is one of the few routes available.
2. A contractual liability cap negotiated against the statutory number stops being a reliable proxy for worst case.
3. Any excess layer a supplier buys should be sized against a scenario, not against the statutory figure.

Nothing about the outcome is settled, and a supplier should not price for the worst reading of it. What a supplier should do is make sure its contractual cap, its policy limits and its indemnity language are all drafted so they respond to a revised statutory number rather than a fixed one.

Most vendors bidding into this sector already hold a liability programme built for conventional industrial supply. It will not respond. Nuclear exclusions sit in almost every standard market wording, and they are broad, catching not just nuclear damage but loss arising from radioactive contamination however caused. The stack below is what has to be built deliberately.

Product liability with a nuclear carve-back

Standard product liability cover is the base layer for defective-component exposure, and its nuclear exclusion has to be carved back to the extent the market will allow. Full carve-back is rare. What is achievable is usually a defined write-back for the supplier's recourse exposure, placed through the pool rather than the open market, with the limit matched to the contract value rather than to the supplier's usual turnover-linked limit.

Erection all risks with delay in start-up

Long-lead nuclear equipment sits in fabrication for years. An erection all risks policy covers the physical damage during transit, storage and installation. The layer that suppliers routinely underbuy is delay in start-up, which responds to the financial consequence of that damage pushing commissioning out. On a nuclear programme the liquidated-damages exposure attached to a schedule slip can exceed the value of the component itself. Our note on contractors and erection all risks with ALOP for infrastructure covers how the DSU and ALOP triggers are drafted.

Professional indemnity on design

Where the supplier's scope includes design, analysis, qualification or safety-case work, the exposure is professional rather than product. Professional indemnity responds to negligent design that causes loss without any physical defect in a delivered item. Design scope is frequently bundled into an EPC package without anyone separating the PI trigger from the product trigger, which leaves a gap exactly where a nuclear claim is most likely to be argued.

Contractual caps aligned to the recourse period

The commonest structural error is a liability cap and an insurance period that track the warranty period when the exposure tracks the recourse period. A two-year defects-liability warranty tells you nothing about how long an operator can come back against you after paying a compensation claim. Align the cap duration, the policy period and any extended reporting period to the recourse window in the contract, and buy run-off cover for the tail beyond it.

The draft is a draft. Submissions are open, and the reported US industry submissions suggest the supplier-liability provisions are the part most likely to move. That argues for preparing the position now rather than waiting for the gazette.

1. Map your scope into liability buckets. Split the contract into supply of goods, installation and construction, design and analysis, and post-commissioning services. Each maps to a different policy trigger, and the split determines which limits you need.
2. Read the flow-down clause against the licence conditions, not just against the head contract. If the operator's licence obliges it to preserve recourse, an indemnity clause promising a waiver is unenforceable.
3. Fix the recourse period in writing. A defined window, expressed in years from a defined event, is insurable. An open-ended recourse right is not.
4. Ask the operator which financial-security instrument it will hold. Pool policy, bank guarantee, parent indemnity or captive each imply a different recovery posture against you.
5. Get a written confirmation from your existing liability insurer that its nuclear exclusion applies. It almost certainly does. Having it in writing is what unlocks the internal budget for the specialist placement.
6. Price the tail, not the term. Build the cost of run-off cover for the recourse period into the bid, because it is not recoverable after award.
7. Stress-test the cap. Run the numbers on a statutory limit revised upward at the first five-year review and on a judicially reopened cap, and see which contractual positions break.

A supplier that walks into a nuclear tender with these seven answers can price the liability line. One that does not is either quoting a number it cannot support or loading a contingency large enough to lose the bid.