The nexus between energy and AI has implications for energy security. There are at least two broad dimensions to this relationship. The first arises from the impact of AI on energy security. AI can be – and indeed already is being – applied to address specific challenges relating to energy security concerns. At the same time, greater digitalisation and connectivity in the energy sector – which enable the use of AI – can create new energy security challenges. The second dimension arises from the need to mitigate energy sector-related supply chain risks, which have implications for the scaling up of data centres to meet the growing demand for AI.

Energy security is characterised by several elements that include, but are not limited to, first, reliable access to energy to meet an economy’s needs; second, the affordability of this energy with limited volatility in prices; and third, resilience against energy market shocks – or the ability of the energy system to quickly recover from them.

A notable use case of AI towards the security of critical energy infrastructure is in places that are typically hard for humans to access. For example, the use of automated surveillance using unmanned maritime systems enabled by AI to secure critical undersea energy infrastructure. Similarly, AI-based predictive maintenance is revolutionising energy infrastructure management by ensuring reduced downtime and improved operational efficiency. 

As the energy sector has become more electrified, digitalised and connected, it has also grown increasingly vulnerable to cybersecurity threats. This vulnerability is compounded by the presence of legacy information technology infrastructure, automation, cloud computing and reliance on third-party vendors that might not have secure systems.

AI acts as a force multiplier in both directions, enhancing threat detection and enabling more responsive protection on the one hand while simultaneously empowering adversaries with tools for sophisticated attacks on the other. AI applications can enable real-time threat detection, automated responses to incidents and enhanced phishing defences. On the flip side, AI-based tools can also be exploited to automate attacks and evade detection. Generative AI tools have been documented as being used by malicious actors for reconnaissance to target organisations, obtain deeper access to target networks, and for malicious scripting and evasion techniques. In view of these evolving threats, the deployment of more proactive AI-enabled cybersecurity systems that are quick to respond to threats is critical for ensuring the resilience of the energy sector. Upskilling, threat mapping and expertise sharing will be essential for keeping the energy sector ahead of the curve

Securing the supply of affordable and reliable power for data centres is at the heart of the challenge of energy for AI.  In particular, the growing expansion of AI data centres has amplified the urgency of addressing power equipment supply chain constraints. Infrastructure expansion across multiple regions has placed considerable pressure on the supply chain for key grid components. The heightened demand extends beyond equipment for high-voltage transmission to include low-voltage solutions, the integration of variable energy resources and new consumer demand, making supply chain resilience more critical than ever.

Besides the additional electricity demand, a major consideration related to the rapid growth of AI and data centres is the demand for critical minerals. Apart from bulk materials like steel and concrete, the construction of data centres requires sizeable amounts of several minerals and metals, such as copper, aluminium, silicon, gallium, rare earth elements and battery minerals. There is a significant overlap between the minerals needed for building new data centres and those that are critical to energy technologies.

The geographical concentration of the supply of most critical minerals is a key concern. In 2030, for example, data centre demand for gallium could equal up to 10% of today’s supply, and China accounts for 95% of gallium refining. The high market concentration for critical minerals highlights significant vulnerabilities to supply shocks if, for any reason, supply from large producers were to be disrupted, whether from extreme weather events, industrial accidents, trade disruptions or geopolitics. 

The global expansion of data centre capacity faces risks from grid connection delays, particularly in regions experiencing high concentrations of demand growth. Connection queues for new data centres can already be long in many key regions. In recent years, several jurisdictions have placed moratoriums on new data centres, as system operators process a backlog of connection requests and assess the capacity of the grid to meet additional connections.

To understand the extent to which data centres might face connection delays, we examined current congestion levels, grid policies and connection timelines. Based on a location-specific analysis of upcoming data centres, we developed different scenarios for the possible number of data centres that may be delayed in connecting to the grid. Our analysis reveals that grid constraints could delay around 20% of global data centre capacity planned for construction by 2030.

Avoiding this risk will require a range of actions from both the energy and technology sectors. Permitting times for new projects need to be cut. Grid operators should rationalise the confusing tangle of data centre connection applications. The technology sector should maximise the buildout of data centres in areas of high power and grid availability and explore strategies to incentivise their operational flexibility. Better management of growing data centre load would be facilitated by better data on both grid constraints and the data centre demand outlook.