In partnership with
EDF generated 373 TWh of nuclear electricity in 2025. Its corporate carbon intensity sits at 26.5 gCO2/kWh, among the lowest of any utility globally. France's grid is 95% carbon-free. Yet no standardised carbon credit mechanism exists for nuclear generation. The accounting frameworks that govern corporate emissions were designed for a world where energy was a commodity input. They recognise neither the infrastructure that produces it nor the energy independence it creates. This disconnect costs the nuclear industry billions in unrecognised climate contribution.
12 grams, zero credits, one structural gap
The IPCC median estimate for nuclear lifecycle emissions is 12 gCO2 equivalent per kilowatt-hour. The UNECE assessment places it lower still: 5.1 to 6.4 gCO2/kWh. Both figures are comparable to wind and lower than solar when lifecycle manufacturing is included. A 2025 study in the Journal of Industrial Ecology found nuclear fuel cycle emissions as low as 3.0 gCO2/kWh.
Wind and solar receive Renewable Energy Certificates (RECs). Nuclear receives nothing equivalent in most markets. The gap is not scientific. It is structural. Carbon accounting standards treat nuclear as a separate category from renewables. The climate does not.
34.7 GW contracted. The accounting framework shrugs.
The GHG Protocol's Scope 2 guidance allows two methods for reporting electricity emissions. The location-based method uses grid average factors. The market-based method uses contractual instruments such as Power Purchase Agreements (PPAs) and RECs.
Nuclear PPAs qualify as contractual instruments under Scope 2. Microsoft's deal with Constellation to restart the Crane Clean Energy Centre demonstrates this. The 835 MW facility will supply carbon-free power under a long-term agreement. Microsoft has now contracted 34.7 GW of clean power across 26 countries. Meta signed three nuclear deals totalling up to 6.6 GW in January 2026. The rationale is not just carbon. It is baseload resilience for AI infrastructure.
The complication is what happens next. The GHG Protocol is updating its Scope 2 guidance, with final standards expected in 2027. The proposed changes require hourly matching of energy purchases. They also limit contractual instruments to "deliverable" sources within the relevant grid region. Nuclear baseload, which runs continuously, actually aligns better with hourly matching than intermittent renewables. Yet the standards discussion focuses almost exclusively on wind and solar.
Behind the meter, beyond the protocol
Behind-the-meter nuclear co-location creates a carbon accounting blind spot. When a datacenter draws power directly from an adjacent reactor, it bypasses the grid entirely. No transmission. No distribution charges. No grid emission factor applies.
The GHG Protocol was not designed for this configuration. Scope 2 assumes grid-delivered electricity. Behind-the-meter generation sits outside the standard framework. The UAE's Barakah complex, powering 25% of national electricity at 5.6 GW, could support co-located compute. France's EDF has identified four nuclear-adjacent sites for datacenter development.
Neither configuration has a clean path through existing carbon accounting standards. The operator generates clean baseload. The datacenter consumes it directly. The accounting framework has no category for this relationship.
Should nuclear-powered datacenters get carbon credit?
France exports 101 TWh. The carbon standards see nothing.
Nations that invested in nuclear capacity decades ago now operate the cleanest and most sovereign grids in the world. France exported 101 TWh of electricity in 2024, a record. It became Germany's largest electricity supplier, delivering 15 TWh across the border. Finland's Olkiluoto 3 replaced Russian energy dependence. Domestic production now covers 96% of Finnish demand. Sweden combines nuclear and hydroelectric for grid intensity below 30 gCO2/kWh. Canada's Ontario runs on 60% nuclear power.
Japan's Kashiwazaki-Kariwa restart will displace 1.3 million tonnes of LNG imports annually. The rationale is energy security, not carbon credits. A country that imports 90% of its energy needs nuclear for sovereignty. The carbon standards ignore this entirely.
These countries receive no structural accounting credit for their baseload decisions. A hyperscaler in France already consumes some of the cleanest electricity on Earth. Buying additional RECs adds marginal climate value but costs real capital. The framework rewards the purchase, not the existing infrastructure.
Compare this to a hyperscaler in Poland or Australia. The grid is carbon-intensive. RECs create genuine additionality there. The same accounting mechanism produces different climate outcomes depending on location. The standards do not distinguish between the two.
Brussels says yes. The credits still say no.
The EU Taxonomy included nuclear under strict conditions from January 2023. The European General Court upheld this in September 2025. Nuclear is now classified as a sustainable economic activity for EU reporting purposes.
This is progress. It is not parity. Nuclear's taxonomy inclusion comes with waste management requirements that solar and wind do not face. The framework acknowledges nuclear's climate contribution. It does not equalise the accounting treatment.
New York State's Zero Emission Credits offer a closer model. ZECs compensate nuclear operators for their carbon-free generation. The mechanism exists. It works. It has not spread beyond a handful of jurisdictions. California's SB 253 will require phased Scope 3 reporting from 2026. As disclosure requirements expand, the gap between what nuclear delivers and what standards recognise will become more visible.
Three actions before the 2027 rewrite
Three actions bridge the gap between current standards and engineering reality. First, structure nuclear PPAs with hourly generation data. The coming GHG Protocol update favours continuous supply. Nuclear's 90%+ capacity factor, the percentage of maximum output actually delivered, is an advantage. The operator's data proves it.
Second, advocate for behind-the-meter accounting standards. The configuration exists. The economics work. The carbon methodology lags behind. Industry bodies, grid operators, and sustainability consultants need a shared framework.
Third, track jurisdictions creating nuclear-specific credit mechanisms. New York's ZEC model will spread as climate disclosure mandates expand. The European Energy Efficiency Directive requires datacenter energy reporting from May 2026. Nuclear-powered facilities will need clear attribution methods.
Accounting standards count megawatts. Not the infrastructure.
The IEA's 2024 report explicitly linked nuclear to energy security for the first time. Governments are treating reactors as sovereign infrastructure. The carbon accounting standards still treat them as incidental generation sources.
The disconnect is between how we measure value and where value is created. Carbon accounting frameworks evaluate outcomes. They do not credit the infrastructure that produces those outcomes, nor the energy independence it delivers.
Last week, we showed how China compresses the gap between generation and compute. It integrates energy, grid, and demand under one decision chain. This week, the pattern repeats. Carbon accounting standards measure the output without crediting the source. The disconnect exists because the frameworks do not recognise the infrastructure stack.
Nuclear carbon credits do not exist. What exists is a measurement system designed for a world where energy was interchangeable. In a world where energy source determines climate outcome, the standards must evolve. The operators, investors, and engineers who understand this will position first.
Next week: We examine the five layers between a power plant and GDP. Where does your organisation break?
Your AI tools are only as good as your prompts.
Most people type short, lazy prompts because writing detailed ones takes forever. The result? Generic outputs.
Wispr Flow lets you speak your prompts instead of typing them. Talk through your thinking naturally - include context, constraints, examples - and Flow gives you clean text ready to paste. No filler words. No cleanup.
Works inside ChatGPT, Claude, Cursor, Windsurf, and every other AI tool you use. System-level integration means zero setup.
Millions of users worldwide. Teams at OpenAI, Vercel, and Clay use Flow daily. Now available on Mac, Windows, iPhone, and Android - free and unlimited on Android during launch.
