What the IAEA Actually Does — And Its Role in the SMR Boom
The world's nuclear watchdog is also its busiest nuclear midwife, and right now it's helping birth a generation of reactors.
If you follow the SMR space even casually, you’ve seen the IAEA’s name pop up in press releases, policy briefs, and conference agendas. The International Atomic Energy Agency is everywhere. But most coverage treats it as backdrop — a logo stamped on a document, a quote from Director General Rafael Mariano Grossi, a mention of “international safety standards.” That’s a shame, because what the IAEA actually does is far more interesting than its reputation as a bureaucratic stamp of approval. It’s simultaneously the world’s nuclear cop, nuclear trainer, nuclear librarian, and — increasingly — nuclear matchmaker for the SMR industry. Understanding its real function is essential if you want to understand why the global SMR pipeline looks the way it does in 2025.
Start with the origin story, because it’s a good one. The IAEA was born in 1957, the direct offspring of President Eisenhower’s famous “Atoms for Peace” speech to the UN General Assembly in December 1953. The Cold War was at full boil, nuclear terror was very real, and Eisenhower’s gambit was to channel the atom toward electricity and medicine rather than warheads. The resulting agency was given a mandate that sounds almost paradoxical: promote atomic energy and control it at the same time. That tension is baked into the IAEA’s DNA, and it shapes everything the organization does today — including its work on SMRs.
What the IAEA actually is
People often assume the IAEA is a regulatory body that tells countries what they can and cannot build. It’s not. The IAEA has no direct enforcement power over member states. What it has is moral authority, technical expertise, and — when it comes to nuclear safeguards — a legally binding verification role under the Nuclear Non-Proliferation Treaty. That distinction matters a lot once you start looking at how the organization actually functions. 🌍
The IAEA does four things well. First, it runs the global nuclear safeguards system — the inspections regime that verifies countries are not diverting peaceful nuclear material toward weapons. Second, it sets safety standards that member states are expected to follow. Third, it provides technical assistance to countries developing or expanding nuclear programs. Fourth, it acts as an information hub, maintaining databases like ARIS (the Advanced Reactors Information System) that track every SMR design on the planet.
None of these are small jobs. Right now, more than 1,300 facilities and locations worldwide are under IAEA safeguards. The safeguards team — over 800 people at headquarters alone — uses tamper-proof seals, surveillance cameras, environmental sampling, and good old-fashioned physical inventory checks. Think of it as a global nuclear audit, every year, everywhere. One IAEA inspector described the work as similar to a bank audit: you compare what’s in the accounting records against what’s physically present. Except the stakes are considerably higher than an overdraft. 🔬
The safeguards challenge for SMRs
Here’s where it gets genuinely complicated. Traditional safeguards were designed around large light-water reactors at fixed sites with predictable fuel cycles. SMRs break most of those assumptions. 🚀
Consider the range of designs now in development. The Nuclear Energy Agency’s July 2025 SMR Dashboard identified 127 distinct SMR designs globally, of which 74 had enough public information to analyze. Of those 74, 51 are already in pre-licensing or licensing processes across 15 countries. The designs range from conventional pressurized water reactors scaled down to 50 MW, to molten salt reactors, high-temperature gas-cooled reactors, and fast-spectrum designs that use enrichment levels traditional plants never approached. Some are designed to be factory-built and shipped to site. Others are designed to float on barges in the Arctic.
Safeguarding a module-based reactor where fresh fuel arrives and spent fuel leaves on a different schedule than anything the existing inspection regime was built around — that’s a real problem. Some SMR designs use high-assay low-enriched uranium (HALEU), enriched to between 5% and 20%, which requires different accounting controls than the standard fuel the inspection system was optimized for. A floating reactor in international waters raises questions about jurisdiction that the existing frameworks were never designed to answer.
The IAEA is working on all of this, though “working on” is doing some heavy lifting in that sentence. The agency has published technical documents on “Safety, Security and Safeguards by Design in Small Modular Reactors”, which pushes the idea that safeguard considerations should be embedded at the design stage rather than bolted on afterward. That’s the right approach. Whether the inspection infrastructure can keep pace with 127 designs in 18 countries is a harder question.
Key safeguards challenges the IAEA is actively addressing:
Novel fuel types that require new accounting and verification procedures
Modular, factory-built designs where “the facility” isn’t a fixed location
Designs using HALEU, which carries higher proliferation sensitivity than standard fuel
Floating and offshore reactors with complex jurisdictional questions
New coolants (molten salt, gas, liquid metal) that change how material flows are tracked
The NHSI: the most important initiative nobody’s talking about
If the safeguards work is the IAEA’s defensive job, the Nuclear Harmonization and Standardization Initiative — NHSI, pronounced as an acronym that only nuclear insiders use — is its offensive play for the SMR era. 🌱
The core problem NHSI addresses is embarrassingly basic: if a company designs an SMR and wants to sell it in multiple countries, it currently has to get regulatory approval in each country essentially from scratch. Every regulator uses different codes, different safety analysis methods, different documentation formats. This isn’t just expensive and slow — it kills the economics of the factory-built, mass-produced model that makes SMRs theoretically cheaper than large reactors. You can’t serialize production if every unit requires a bespoke licensing campaign.
NHSI, launched in 2022, runs on two parallel tracks. The regulatory track is building frameworks for:
Sharing safety review information between national regulators so work doesn’t get duplicated
A multinational pre-licensing joint review process where regulators from different countries evaluate a design together
A collaborative review process that lets regulators work in parallel during ongoing national licensing
Mechanisms for leveraging one country’s completed review when another country starts its own
The industry track, meanwhile, has pulled in more than 200 contributors from over 30 countries and is working on harmonized user requirements, common approaches to codes and standards, and shared experimental data. As IAEA Director of Nuclear Power Aline des Cloizeaux put it at the 2025 binding.energy conference: “We don’t build reactors — but we help the world build them better.”
By mid-2025, NHSI had moved into Phase II, focused on implementing Phase I’s recommendations rather than just developing them. Three technical documents capturing Phase I’s regulatory work are expected to be published in 2025. That’s slow by startup standards. It’s remarkably fast by international treaty organization standards, and the work is genuinely complicated. Harmonizing national regulatory frameworks without compromising national sovereignty or safety standards is the kind of problem that looks simple until you actually try it. ♻️
The SMR School and the newcomer country challenge
Here’s a dynamic that gets undercovered in Western SMR coverage: the countries most interested in deploying SMRs in the next decade are not primarily the US, UK, or France. They’re countries like Kenya, Ghana, Jordan, Poland, Thailand, and Mongolia — nations with limited or no existing nuclear infrastructure, policy frameworks still being written, and regulatory bodies that may not yet have staff trained in nuclear licensing.
About 30 “newcomer” countries are actively considering or advancing toward nuclear power, and many of them have specifically identified SMRs as the path in — precisely because a 100 MW SMR is a more realistic starting point than a 1,200 MW gigawatt-class plant. This is actually one of the more compelling arguments for SMRs that often gets lost in debates about economics and timelines: smaller reactors lower the minimum viable grid size and the upfront capital requirement. A country with a peak electricity demand of 500 MW can’t meaningfully integrate a traditional large reactor, but it can integrate two or three SMR modules. 💡
The IAEA’s response to newcomer demand is the SMR School program, launched in 2025. The inaugural session was hosted by Kenya in May 2025, drawing 28 officials, policy makers and managers from Kenya, Ghana, Niger, Nigeria, Uganda and Zambia. A second SMR School followed in Thailand in September 2025, focused on Asia. Future schools are planned for Latin America.
The curriculum covers:
Energy planning and how SMRs fit into national grids alongside renewables
Technology development and the current state of available designs
Legal frameworks and the regulatory infrastructure a country needs before it can license a reactor
Safety, security and safeguards obligations under international law
Economics, financing models, and how to structure agreements with vendors
This matters enormously for the long-term SMR market. A country that goes through this process with IAEA guidance is more likely to end up with a functional regulatory framework — which means fewer surprises, fewer delays, and fewer political reversals once construction starts. And for vendors, a customer country that understands what it’s buying and has the institutional capacity to oversee it is a vastly better commercial partner than one that doesn’t.
Does the IAEA process move fast enough? Probably not, by the timelines SMR advocates tend to cite. The IAEA’s milestone approach — a structured framework guiding newcomer countries from initial exploration to actual operation — estimates 7 to 10 years for a first-of-a-kind SMR in a newcomer country, dropping to 4 to 5 years for subsequent units. That’s a long runway for technology companies accustomed to software development cycles. But it’s also probably realistic for an industry where the downside of rushing the regulatory process is a meltdown, not a bad product review.
What this means for the SMR industry
The IAEA is not an obstacle to SMR deployment. That’s worth stating plainly, because the agency sometimes gets cast in that role by people frustrated with the pace of nuclear regulation. The NHSI’s entire purpose is to speed up licensing. The SMR School’s purpose is to expand the addressable market. The safeguards-by-design work is trying to solve the inspection problem before hundreds of novel reactors get built, rather than after. 📈
The more honest tension is this: the IAEA is a consensus organization with 178 member states, and consensus organizations move slowly. NHSI has over 200 industry contributors and has been operating since 2022. Its Phase I results are being published in 2025. For comparison, a software API standard might get from proposal to adoption in six months. That speed mismatch is a structural feature of international governance, not a bug the IAEA can fix.
What the IAEA can do — and increasingly is doing — is create the connective tissue that the global SMR market needs. A database of every SMR design with standardized technical specifications. A forum where regulators in a dozen countries can share safety review findings. A curriculum that turns energy ministers in sub-Saharan Africa into informed buyers of nuclear technology. A safeguards architecture that can actually handle modular, factory-built, fuel-flexible reactors operating on four continents. None of that is glamorous work. None of it generates headlines like a reactor groundbreaking or a funding round. But without it, the SMR market would be a collection of isolated national experiments rather than a coherent global industry.
The IAEA’s 2024 SMR report puts the current situation accurately: the agency’s unique role is in “catalysing technology development and deployment in Member States” — not building reactors, not licensing them, but creating the conditions in which building and licensing becomes possible at scale. ⚡
The question worth sitting with is whether the pace of that institution-building is keeping up with the pace of design development. With 127 designs globally, 51 in active licensing processes, and an 81% rise in SMR financing announcements since 2024, the commercial side of the industry is accelerating fast. The governance infrastructure is accelerating too — just not at the same speed. That gap is probably the single biggest systemic risk in the SMR sector right now, more than any individual technology question. How do you think that gap gets closed — does the IAEA need to move faster, or does the industry need to slow down and wait for the rules to catch up?