Scaling AI Computing and Infrastructure within Net Zero Goals

Meeting the Energy and Security Challenges of Scaling Public AI Computing in the UK

The UK must now confront the interconnected "3Cs" challenge in managing its digital future: the Cash, Carbon, and Compliance costs associated with data growth. These three interlinked pressures, economic, environmental, and regulatory, are converging as artificial intelligence, data infrastructure, and encryption technologies scale rapidly.

The UK government's ambition to increase public-controlled AI computing capacity twentyfold by 2030 presents significant infrastructural, energy, and security challenges. While this scale-up supports digital transformation and economic innovation, it will substantially heighten electricity demand at a time when the UK must also electrify heating and transport sectors. In parallel, emerging global risks in quantum computing threaten existing data infrastructure security. Strategic policy interventions are required to ensure AI’s growth aligns with national decarbonisation and security objectives.

A further dimension is the urgent need for standardised methods to quantify the Cash, Carbon, and Compliance (3Cs) costs of data growth. Without robust metrics, the UK lacks the tools to govern data sustainability, making it difficult to enforce low-carbon digital practices. Insights from the World Economic Forum Global Technology Retreat emphasised that most internet traffic will soon be generated by AI agents, not humans, underlining the need for clarity on AI’s systemic energy footprint. In response, the UK government should invest in data carbon accounting research, a vital move that would catalyse innovation and global standard-setting in digital decarbonisation.

Context and Strategic Importance

The rapid expansion of AI computing poses a dual challenge: balancing exponential energy demand with the UK’s legally binding net zero commitments, and safeguarding public and private digital infrastructure from quantum-enabled security threats.

• AI-driven Energy Demand: Training and inference of AI models are electricity-intensive. A twentyfold increase in public AI compute could raise consumption from an estimated 5 TWh to 100 TWh annually, 25% of the UK's total 2021 electricity use.
• Decarbonisation Tension: Electrification of transport and heating already strains renewable generation capacity. AI growth risks intensifying this pressure unless matched by aggressive renewable deployment and energy efficiency gains.
• Security Vulnerability: Quantum computing breakthroughs by Google (June 2025) indicate encryption methods protecting financial and personal data could be broken within five years, necessitating urgent shifts to post-quantum cryptography.
• Data Carbon Accounting Gap: The UK currently lacks standardised methods to quantify the Cash, Carbon, and Compliance (3Cs) costs of data growth. This absence of clarity hinders informed regulation and market innovation in green digital practices

Energy Infrastructure and Capacity Risks

a. Renewable Supply Constraints

• While over 40% of UK electricity now comes from renewables, growth has plateaued at 5–7% annually.
• Under high-consumption scenarios, renewable generation could lag behind cumulative demand from AI, heating, and electric vehicles.

b. Demand Competition

• Electrification of homes (heat pumps) and vehicles must be prioritised for decarbonisation.
• AI computing could displace renewable energy availability for these sectors during peak demand, triggering fallback to fossil generation.

c. Mitigation Strategies

1. Efficiency: Support development of energy-efficient AI algorithms, low-power chips, and waste heat reuse in data centres.
2. Data Minimisation: Implement national policies to reduce dark data and enforce data lifecycle management.
3. Grid Modernisation: Invest in smart grid technologies to allocate renewable electricity efficiently.
4. Supply Acceleration: Expedite offshore wind licensing, storage integration, and microgrid deployment.

3. Security and Governance Readiness

a. Quantum Risk Timeline

• Previous 2040+ timeline has collapsed to within five years.
• A new generation of cryptographic standards is required across public and private sector infrastructure.

b. Data Governance Gaps

• Poor interoperability and slow data collection reduce readiness for AI integration.
• Governance frameworks must address dark data and improve cross-border data sharing protocols.

c. Standardised Data Carbon Metrics

• A government-backed programme should be established to define and promote consistent methodologies for assessing the energy and carbon costs of digital infrastructure.
• As highlighted at the WEF retreat, clarity in data emissions measurement is essential for guiding AI infrastructure investment, compliance, and innovation.
• This would position the UK as a leader in setting global standards, attracting investment in green tech and sustainable digital services.

d. Educational and Workforce Reform

• AI expansion will require specialists not just in data science, but in infrastructure interoperability, data governance, and security.
• Education policy should promote advanced training (PhDs and apprenticeships) in digital infrastructure, smart energy systems, and AI governance.

Policy Recommendations

1. AI Carbon Governance: Introduce national guidelines to assess and report carbon impact of public AI compute projects.
2. Energy Cap for AI Compute: Consider capping public-sector AI computing energy use unless linked to renewable sources.
3. Green Data Standards: Establish regulatory and fiscal incentives for data centres powered by 100% renewables.
4. Post-Quantum Readiness: Fund R&D for post-quantum cryptography and mandate early adoption in public procurement.
5. Digital Skills Strategy: Launch national programmes to train digital infrastructure engineers, energy tech specialists, and AI governance experts.
6. Investment in Carbon Accounting R&D: Fund research and methodologies that standardise how data energy and emissions are measured across sectors.
7. International Standard Leadership: Use UK-based evidence to influence global benchmarks for sustainable data use and digital decarbonisation.

Conclusion

Investing in standardised methods to quantify the cash and carbon costs of data growth is a strategic imperative. This, alongside growing AI energy demand and emerging quantum risks, requires decisive and coordinated action. A twentyfold expansion of AI computing offers strategic advantage, but only if managed in harmony with net zero goals and digital security imperatives. Proactive government leadership, through regulation, funding, and infrastructure planning, will be essential to ensuring AI growth strengthens, rather than undermines, the UK’s long-term economic and environmental resilience.

Note: This briefing paper has been partially informed by insights from the World Economic Forum Global Technology Retreat (June 2025), to which the authors were invited, which brought together global experts to discuss infrastructure, AI, energy, and quantum security.