The UK’s Streamlined Energy and Carbon Reporting (SECR) framework was designed to help large companies disclose their energy use and emissions transparently. However, as the use of artificial intelligence (AI) grows, the environmental impact may no longer be fully captured by conventional reporting techniques. The time has come for businesses adopting AI to reconsider their approach to SECR compliance and make sure the actual carbon cost of digital innovation is taken into consideration.
Understanding SECR
In order to improve corporate transparency regarding carbon and energy usage, the UK government implemented the Streamlined Energy and Carbon Reporting (SECR) framework in April 2019. It is intended to extend reporting requirements to a wider range of organisations, replacing the previous Carbon Reduction Commitment (CRC) programme.
The scheme is administered by the Department for Energy Security and Net Zero (DESNZ) and applies to financial years starting on or after 1 April 2019. It aims to:
- Improve the quality and consistency of carbon reporting
- Encourage energy efficiency initiatives
- Reduce the UK's overall carbon footprint
- Support businesses in transitioning to net-zero emissions
Who Must Comply with SECR
Quoted companies
All UK-incorporated companies listed on a stock exchange are required to comply, including those listed on the London Stock Exchange, European Economic Area markets and the New York Stock Exchange or NASDAQ.
Large unquoted companies
UK companies that do not meet the "quoted" definition but satisfy at least two of the following:
- 250+ employees
- Annual turnover ≥ £36 million
- Balance sheet total ≥ £18 million
Large LLPs (Limited Liability Partnerships)
LLPs that meet the same size criteria as unquoted companies must also comply, including energy and carbon disclosures within their annual reports.
SECR also applies to UK-registered subsidiaries of overseas parent companies if they meet the size thresholds.
What Must Be Reported
To achieve SECR compliance, organisations are required to report:
Total energy consumption (UK and global):
Must include energy from purchased electricity, gas, transport fuel and other sources (e.g., steam or heat networks). This must be reported in kilowatt-hours (kWh) and contextualised against the organisation’s operations.
Scope 1 and Scope 2 GHG emissions:
Scope 1 includes direct emissions from owned or controlled sources (e.g., company vehicles or on-site gas boilers). Scope 2 accounts for indirect emissions from the generation of purchased electricity or steam.
Energy efficiency actions:
Companies must outline all practical measures taken within the reporting year to reduce energy usage such as LED lighting retrofits, server optimisation or moving to renewable energy contracts.
Methodologies:
A transparent explanation of how figures were calculated (e.g., emission factors used, data sources, estimation techniques). This ensures comparability and traceability for stakeholders.
Learn More: SECR Explained Guide
AI Use Amongst Organisations
Over the last decade, artificial intelligence has come to be widely adopted on a global scale. This isn’t limited to individual usage alone but also encompasses AI-driven enterprise solutions created for organisations. AI is being used in business strategy, product design and service delivery in a variety of ways, from automating back-office tasks to creating advanced chatbots that customers can talk to. This is most prevalent in tech, computer software and information technology (IT) sectors. In 2023, Google’s GHG emissions were almost 50% higher than they were in 2019 largely due to energy demand from expansion of their data centres.
Features like code generation (e.g., GitHub Copilot), sentiment analysis and language processing are now standard in leading platforms. These features often rely on continuous, high-frequency AI inferences hosted on energy-intensive GPU servers. Over time, this contributes to an exponential increase in baseline electricity consumption.
The Rising Energy Footprint of AI
A report from the International Energy Agency (IEA) indicates that the largest data centres currently under construction would use 20 times as much electricity as the average AI-focused data centre, which uses as much as 100,000 households. With global investment in data centres nearly doubling since 2022, there is a need for organisations to proactively monitor and manage energy consumption from their use of artificial intelligence and how it factors into their energy and carbon reporting.
Credit: World Economic Forum.
- Computer software: Software companies integrating AI into their platforms, such as code copilots, automated testing suites and recommendation engines, will result in increased server usage and, consequently, electricity bills and emissions.
- Machine learning research: Research companies experimenting with large language models (LLMs), computer vision or autonomous systems often incur steep energy costs during model training and hyperparameter tuning.
- Cloud platforms: Major cloud platform providers have been rapidly expanding their server capabilities and physical data centres, leading to higher energy costs.
Factoring AI Energy Usage into SECR Compliance
These developments carry significant implications for SECR compliance. Because of their size and financial footprint, many tech companies are already covered by SECR. However, what’s often missing from their streamlined energy and carbon reporting is a breakdown of how much energy is being consumed by AI-specific workloads, e.g., Scope 3 if purchased as a service, or Scope 2 if hosted on site. AI computing is often hidden behind APIs, Software as a Service (SaaS) platforms or cloud billing dashboards, which makes it easy to miss in energy audits and annual reports. Failing to include these emissions, particularly those within Scope 3, could lead to a carbon footprint which is not representative due to energy use outsourcing. Here are steps organisations can take to integrate AI energy usage in their SECR compliance approach:
Step 1: Update Energy Accounting Protocols
Ensure that all AI-associated electricity (on-premises or cloud-hosted) is captured in your total energy consumption figures. This includes:
Step 2: Provide Transparent Methodologies
Specify whether AI-related energy and emissions are included in your SECR framework. Always describe how data was collected and calculated, including emission factors. If data was estimated, explain the assumptions clearly in your SECR filing.
Step 3: Disaggregate AI Emissions Where Possible
For example, segment AI model training from other IT operations and present inference emissions separately if operating at scale. If AI workloads are a major business activity, include a standalone narrative in the energy efficiency section.
Step 4: Engage Cloud Providers
Cloud Proactive organisations can take it a step further by asking their cloud or AI service vendors if they can provide essential information like:
- Carbon intensity of AI compute services
- Location-based energy usage data
- Renewable energy matching or offset policies
Include this data in your supply chain due diligence and voluntary Scope 3 disclosures.
The Bottom Line
The rise of new-age digital innovations for organisations has all but permanently altered the energy landscape. Regulatory frameworks that require the tracking, monitoring and managing of energy usage, like the SECR, will continue to evolve to reflect this changing dynamic. In the age of AI, SECR compliance offers a vital lens through which companies can measure and manage the true cost of digital progress. Businesses that proactively incorporate AI impacts into their disclosures will be better prepared for future ESG expectations. Learn more about how you can better align your SECR filing with emerging trends here: Streamlined Energy Carbon Reporting (SECR).