India has 8.78 GW of installed nuclear capacity. The 2047 target is 100 GW. The public-sector operators do not have the balance sheet, the construction bandwidth, or the procurement architecture to build the gap. SHANTI, enacted December 2025, opens the sector to private and foreign participation; PFBR achieved first criticality four months later. Westinghouse and EDF were locked out for fifteen years by a supplier-liability provision SHANTI has now recalibrated. Space opened the same way after PSLV; defence opened after fielded platforms. What is the state actually opening, and what is it holding back?
The institutional fact underneath SHANTI is that India cannot afford to miss the nuclear window. India's installed nuclear capacity stood at 8.78 GW in October 2025; the Viksit Bharat target is 100 GW by 2047, a twelve-fold expansion over roughly two decades against cumulative investment requirements of approximately USD 180 to 214 billion. The current public sector operators, NPCIL and BHAVINI, do not have the balance sheet, the construction bandwidth, or the procurement architecture to deliver this build-out. By comparison, China's installed nuclear capacity is already 125 GW today, projected to reach 200 GW by 2040, and its atomic authority has reported the construction capability to build fifty reactors simultaneously. India's 100 GW by 2047 target is, in absolute terms, where China stands now. The interim target within this trajectory, 22.38 GW by 2031-32 composed of indigenous 700 MW units and imported 1,000 MW reactors, requires the installed base to nearly triple in six years; the public-private capital split the architecture assumes is calibrated around roughly USD 26 billion (INR 2.16 lakh crore) of private investment underwriting 11,000 MW of capacity by 2040, with the balance of the 100 GW trajectory following through the 2040 to 2047 window. The global nuclear industry is now in its Small Modular Reactor window, where advanced designs are being commercialised in the United States, the United Kingdom, France, Russia, China, and South Korea during the 2025 to 2040 period; countries that do not build domestic SMR capability in this decade will either import reactor technology at scale from those jurisdictions or fall behind on clean firm power entirely. India's net-zero commitment by 2070 cannot be met without nuclear at scale, and the baseload arithmetic against the country's 2047 energy-demand trajectory does not leave nuclear as one option among several; it makes nuclear the only low-carbon technology that can substitute coal for industrial and commercial baseload. SHANTI is not the state revising its view of strategic autonomy in nuclear. It is the state acknowledging that the nuclear capacity India needs in the 2025 to 2047 window cannot be built through public capital and public construction alone, and that the window to build it is now.
The Prime Minister's own framing at the SHANTI passage on 17 December 2025 made this explicit: the Act was described as delivering "a decisive boost to a clean-energy future" with specific reference to "safely powering AI" and "enabling green manufacturing." The nuclear expansion is institutionally tied to the same industrial-strategic logic that drives the PLI architecture, the Semicon mission, and the green hydrogen mission. The state is not treating nuclear as a standalone clean-energy programme; it is treating nuclear as infrastructure for the larger industrial ambition. Missing this window would mean meeting the 2047 electricity demand through continued coal dependence, importing reactor technology at the scale the domestic fleet implies, or scaling back the Viksit Bharat industrial targets themselves.
The industrial-strategic framing has now produced its own institutional window. Industry readings from the reactor-manufacturing side, L&T's heavy engineering leadership; from the advisory side, PwC's infrastructure practice; and from the operator side, ESDS and the larger Indian data-centre industry, converge on the same observation: the five years from 2026 to 2030 are the policy and partnership window in which nuclear-data-centre co-location frameworks will be composed, and the compute infrastructure buildout will follow the frameworks set in this window. The first wave of nuclear-linked data centre parks in India is expected to begin taking shape between 2030 and 2035, with thorium-based systems moving closer to commercial deployment as Stage 3 of the three-stage programme matures. Nuclear-powered AI compute is emerging as the category in which India's uranium reserves, the world's largest thorium deposits, the indigenous reactor programme, and the SHANTI reform architecture are being read as a single geopolitical proposition, and the institutional question is whether the state composes the co-location framework inside the 2026 to 2030 window or after it has already closed.
This is why the sequencing pattern matters. The Indian state has consistently opened strategic sectors only after the sovereign capability inside them has reached institutional confirmation under public monopoly. In 2020, IN-SPACe was established and space opened to private participation after ISRO had built the commercial launch capability the private sector could now inherit; private launch providers work with PSLV and SSLV-class derivatives, not with the cryogenic upper-stage capability that took two decades of sovereign development. In 2021, the Ordnance Factory Board was corporatised into seven defence PSUs and the Strategic Partnership Model opened major platform acquisition to private firms; this followed, rather than preceded, the indigenous platform maturity that made the model viable. In December 2025, SHANTI was enacted once BHAVINI had demonstrated that the Prototype Fast Breeder Reactor was ready for criticality, the regulator had cleared the first approach, and the plutonium-breeding capability that unlocks Stage 3 of Bhabha's three-stage programme was at the threshold of operational demonstration. First criticality was achieved at Kalpakkam on 6 April 2026, four months after SHANTI came into effect. The legislative moment did not wait for operational completion; it waited for institutional confirmation that operational completion was imminent. The state holds the sector closed while the capability forms, and opens it once the capability is stable enough, or credibly close enough to stability, that private participation inherits rather than replaces the sovereign layer.
What the state actually opens, and what it retains, is visible in SHANTI's reservation architecture. Uranium enrichment beyond a notified threshold, spent fuel management, reprocessing, and heavy water production remain reserved for the Union Government or its wholly owned entities. Private licensees may construct, own, operate, and decommission nuclear power plants; they may not touch the fuel cycle. They will run reactors on state-supplied fuel and return spent fuel to state-run facilities. Private participation in nuclear means operating the plant; it does not mean operating the fuel cycle. This is the same architecture that governs space and defence liberalisation. IN-SPACe authorises private launches; ISRO retains cryogenic propulsion and strategic payload certification. The Strategic Partnership Model permits private platform manufacture; DRDO retains core weapon system development, and strategic weapons manufacture sits with government entities. The downstream is opened; the upstream chokepoint is held.
The liability recalibration inside SHANTI is the legal change that actually unlocks foreign vendor entry. The prior regime had allowed plant operators to seek statutory recourse against equipment suppliers for damages arising from a nuclear incident, even where intent to cause damage could not be established. This exposed suppliers to open-ended liability indefinitely after a plant became operational, and it kept foreign reactor vendors out of India for fifteen years; the EDF project at Jaitapur and the Westinghouse project at Kovvada were the most visible casualties of this architecture. SHANTI converts supplier recourse into a contractual right only, activated where intent to cause damage is proven. Operator liability is graded by reactor size, rising to INR 3,000 crore for large reactors, with an overall cap calibrated for project financeability and insurability. This is the legal architecture global reactor vendors had been waiting for; its presence, not its precise numbers, is what now brings Jaitapur, Kovvada, and the SMR bid pipeline into commercial motion. The Indian nuclear insurance pool will have to expand to support the expanded fleet, and new reinsurance arrangements with European and Lloyd's markets will need to form; these are institutional consequences of the reform, but they follow from the central liability recalibration rather than being the principal purpose of it.
A fiscal-treatment mismatch sits alongside the legal recalibration. Nuclear power is currently taxed at 18 percent GST while renewables attract 5 percent, and nuclear does not yet carry green-energy classification for the financing-cost concession solar and wind receive. Industry representations from L&T and the PwC infrastructure desk have framed this as a structural disadvantage in the cost-of-capital arithmetic, particularly for the first wave of private licensees whose projects will be financed before fleet scale is achieved. The fiscal architecture has not yet been reconciled with the policy direction SHANTI institutionalises, and the GST Council and Ministry of Finance pathways that would close this mismatch operate on a tempo independent of the nuclear reform itself.
Three structural features of SHANTI carry longer consequence than the current commentary captures.
The first is the foreign-incorporated company exclusion. SHANTI does not permit companies incorporated outside India to hold licences. The Act itself does not address foreign direct investment; the Department of Atomic Energy clarified in Parliament on 10 December 2025 that FDI in atomic energy remains prohibited under existing law, and that the foreign equity architecture will be set through subsequent rules issued by the Department for Promotion of Industry and Internal Trade. Every international reactor vendor, Westinghouse, EDF, Rosatom, KEPCO, GE-Hitachi, and the emerging SMR vendors, must therefore route through an Indian-incorporated joint venture; the Indian JV partner holds the licence, the regulatory interface, and the long-term revenue claim regardless of where the DPIIT cap eventually settles. The Adani Group's reported discussions with Uttar Pradesh on eight 200 MW SMRs, and the indicated intent from Reliance Industries, Tata Power, JSW Energy, Hindalco Industries, the Naveen Jindal Group, L&T, and Indian Railways, are the visible opening moves of this formation. L&T's position inside the cohort is distinctive; it is both a potential licensee and the country's most institutionally capable reactor-equipment manufacturer, which gives it dual-role leverage that none of the other aspirants hold. The Indian private nuclear champions of the next two decades are being selected now through JV formation, not through licensing rounds that have not yet opened. A company whose institutional posture is built on reading formal licensing notifications will encounter the selection after it has already happened.
The second is the Atomic Energy Regulatory Board's statutory independence. Legal independence was a foreign vendor and insurer precondition for entry, and SHANTI delivers it; AERB now has statutory backing rather than the executive-order status it held since 1983. Institutional independence, however, is a different matter. AERB has historically been staffed through deputations from BARC and NPCIL, funded through Department of Atomic Energy channels, and embedded in the promotional apparatus it is meant to regulate. Statutory status is the container; the cultural and budgetary separation between regulator and promoter will take a full generation to achieve, and it will be stress-tested the first time a licence decision goes against a state-linked operator. Foreign insurers pricing Indian nuclear risk, and foreign vendors underwriting SMR designs against long tail liabilities, will watch this separation closely because their risk models depend on a regulator that can act against the promoter when required.
The third is the procurement architecture of the first unit. The BSMR-200 tender, currently in inter-ministerial consultation ahead of CCEA approval, will do far more institutional work than the 220 MWe unit itself generates. The clearance architecture assembled for the pilot, CCEA note, DAE concurrence, AERB clearance, environmental clearance, site safety assessment, MoF fiscal vetting, MEA concurrence on foreign participation, is carried out once at the first unit and inherited by every successor project as precedent. The approved per-megawatt cost, roughly ₹27 crore at the ₹5,960 crore project cost, becomes the fleet's pricing anchor; every subsequent SMR project will be benchmarked against it, and the cost contestation that appears continuous from outside the architecture effectively closes at the moment of first-unit CCEA approval. The foreign participation terms, the local-partner requirements, the technology transfer clauses, and the performance liability provisions are all being composed in the pre-tender drafting window now underway between DAE, MEA, MoF, AERB, and NITI Aayog. The first unit of a strategic-sector fleet in India compresses the clearance, pricing, and foreign-participation architecture of every subsequent project; the pre-tender drafting window is the only stage at which that architecture remains institutionally negotiable.