As artificial intelligence drives a surge in data centre construction, the challenge of powering these energy-intensive sites is becoming critical. By 2035, data centres in the United States could consume up to 9 percent of the country’s electricity—underscoring urgent demand for scalable, sustainable power.

Technology firms are exploring new energy strategies, with small modular nuclear reactors (SMRs) gaining traction. Unlike traditional nuclear plants, SMRs are smaller, faster to build and can be located closer to data centres. Google has partnered with Kairos Power to develop 500 megawatts of SMR capacity, while Equinix is working with Oklo—backed by OpenAI’s Sam Altman—on 20 micro reactors.

However, SMRs face long regulatory lead times. Nuscale’s 77-megawatt design remains the only model to gain partial US approval, but was shelved in 2023 over rising costs.

The UK government is backing SMRs as part of its energy strategy, pledging £2.5 billion over the next decade, with Rolls-Royce at the helm. France is also investing in SMRs through EDF’s Nuward project. But even as interest grows, these technologies remain a medium-term solution.

In the short term, AI firms are turning to gas and hybrid models. xAI’s Colossus supercomputer in Memphis now generates 500 megawatts on-site using mobile gas turbines. Utilities are also scaling up—CenterPoint Energy is planning $65 billion in grid investment through 2035, while PG&E is committing $73 billion to meet new demand and improve fire resilience.

The US Energy Secretary has encouraged a “Bring Your Own Generation” model, urging data centres to supplement grid supply with on-site generation. While this approach reduces strain on public networks, it has drawn scrutiny over fairness, transparency and consumer impact.

AI power demands are vast. Deloitte projects that US AI data centre demand could hit 123 gigawatts by 2035—a thirtyfold increase. Some planned campuses will require electricity comparable to major gas or nuclear plants. Meta’s $10 billion data centre in Louisiana offers a glimpse of the future. Expected to draw more power than New Orleans during peak summer, it is prompting major new energy infrastructure builds by Entergy, raising questions over cost, tax relief and public accountability.

For the UK and other AI-focused nations, this energy race brings both risk and opportunity. Balancing innovation with environmental and social responsibility is now essential. SMRs may hold long-term promise, but in the near term, traditional energy, on-site generation and grid upgrades will be critical to fuelling the AI revolution sustainably.

Created by Amplify: AI-augmented, human-curated content.