As artificial intelligence drives a new wave of digital infrastructure investment, data centres are entering a period where energy efficiency is becoming as strategically important as computational performance. The European Union’s Energy Efficiency Directive is pushing the sector towards greater transparency and operational discipline, forcing operators to measure, report and optimise energy consumption in ways that were previously voluntary.
The directive, first introduced in 2012 and significantly updated in 2023, sets a binding target for an 11.7 per cent reduction in final energy consumption across the EU by 2030. For data centres, particularly those exceeding 500 kW of IT power demand, the rules introduce mandatory reporting requirements that bring energy performance into sharper focus at a time when AI workloads are increasing electricity demand across the industry.
Data centre electricity usage is projected to exceed 1,000 TWh by 2026, a figure that underscores why policymakers are tightening scrutiny. According to Ian Tanti, segment and specification manager at Socomec Group, the regulatory framework is reshaping operational strategies across the sector, with compliance now tied closely to measurement accuracy and ongoing optimisation rather than periodic efficiency initiatives.
Reporting requirements reshape operations
Under Article 12 of the directive, data centres above the 500 kW threshold must report key performance indicators into a European database. Reporting deadlines began in September 2024, followed by annual submissions each May from 2025 onwards. Required data includes energy consumption, temperature set points, water use, waste heat utilisation and renewable energy share.
“This shift toward continuous reporting reflects a broader change in how data centres are assessed,” Tanti said. “Rather than focusing solely on capacity or uptime, regulators increasingly expect operators to demonstrate quantifiable improvements in efficiency over time. Article 7 reinforces this by requiring annual energy savings across member states, effectively translating policy targets into operational expectations for facilities managing large AI and cloud workloads.”
Power Usage Effectiveness, or PUE, remains the primary metric for measuring data centre efficiency. Defined in EN 50600-4-2, PUE compares total facility energy consumption with energy used directly by IT equipment. Accurate, granular monitoring of power flows is essential for calculating this value and identifying inefficiencies.
Tanti argues that permanent monitoring solutions are now fundamental to compliance, noting that temporary measurement campaigns are insufficient for meeting regulatory expectations. “Reliable and repeatable measurement, even at low load levels, is critical for building trustworthy reporting frameworks,” he added.
AI workloads accelerate infrastructure rethink
The growing adoption of AI is amplifying the importance of energy efficiency. AI-driven workloads often require higher power densities and more intensive cooling, creating pressure on both energy supply and infrastructure design. As a result, efficiency measures once treated as best practice are increasingly becoming baseline requirements.
The EU Code of Conduct on Data Centre Energy Efficiency calls for modular and scalable uninterruptible power supply systems, efficient part-load operation and smart conversion modes. The directive also mandates “elite” efficiency UPS systems, requiring 93.5 per cent efficiency in VFI mode and 98 per cent in VFD mode, alongside compliance with IEC 60240-3 standards.
“Rightsizing infrastructure is another key strategy,” Tanti explained. “Instead of deploying full capacity upfront, modular systems allow facilities to grow incrementally, reducing both capital expenditure and energy losses caused by underused equipment. This approach aligns with broader sustainability goals by limiting wasted capacity and extending equipment lifecycles.”
From compliance to competitive advantage
While the directive is regulatory in nature, its implications extend beyond compliance. Continuous measurement, scalable infrastructure and efficient power management are becoming central to how data centres maintain competitiveness in an AI-driven market. As workloads increase and expectations around sustainability intensify, operators that can demonstrate efficient operations may gain strategic advantages in attracting investment and customers.
The directive’s long-term goal of reducing energy consumption by 2030 positions data centres within a broader industrial transition. The challenge for operators is to balance the rising computational demands of AI with increasingly strict energy accountability.
For Europe’s data centre industry, the message is clear. The era of growth without scrutiny is ending. AI may be driving demand for infrastructure, but regulation is defining how that infrastructure must evolve.
