ENERGY & POWER
Ensuring continuous power supply
Electricity demand is set to rise sharply, driven by the computing capacity needed to deploy artificial intelligence (AI). Continuous supply is an even more decisive factor – and where grid access is constrained, many will need to buy power behind the meter (BTM), often from separate, private-equity-backed providers.
As capacity evolves, large-scale facilities, some of the largest planned data centre campuses are now approaching or exceeding 1 GW of potential capacity – comparable to the consumption of a major city like Berlin.
This is reshaping how projects are designed and delivered. Developers are increasingly required to secure reliable, scalable power solutions, often incorporating BTM generation, renewable energy integration, and energy storage to mitigate grid constraints. But this heightened demand places more pressure still on infrastructure. Longer connection timelines, increased costs, and, in some regions, restrictions on new development are in place.
The scale and complexity of these power solutions require a closer look at technology performance, supply chain dependencies and the resilience of critical systems. Early engagement with stakeholders, namely utilities, regulators, and insurers, is crucial to ensuring that power strategies are insurable, timelines achievable and operation reliable.
The availability of energy is also critical to a location’s success. Balanced the availability of tax incentives, proximity to demand centres, availability of water and the exposure to natural catastrophe risks gives a complex set of conditions to work in.
To secure sustainable, reliable power while ensuring commercial viability, regulatory support, and long-term operational resilience, operators must strike a balance.
Power supply options
No single source delivers the resilience a data centre needs. Leading users are turning to renewables like wind and solar to cut carbon, but they can't match gas turbines for continuity and flexibility alone. Nuclear is arguably becoming the favoured option, with a renewed reputation and ability to provide stable, continuous supply – and a combination of sources tends to create the greatest resilience.
As servers must run continuously, many rely on backup systems – uninterruptible power supply (UPS), generator sets, backup battery units (BBU) and redundant power supply (RPS) – to keep operating through outages.
Power is often taken directly from the grid via power purchase agreements (PPAs): fixed-price contracts that give buyers and sellers certainty and avoid price volatility. If a provider under-delivers it may face penalties, unless a 'force majeure' clause applies and the risk shifts to the user. But in some countries the grid simply isn't resilient enough to meet demands.
It's crucial to establish who covers replacement power if contracted supply fails. A PPA draws from the closest source regardless of distance, so a data centre can sap a community's power – an issue if availability drops. Securing local authority consent on allocation before choosing a location is well worth the effort.
Grid blackouts are rare. Insurers will cover a power plant for a property-damage event that forces an outage (subject to deductibles) but have little appetite to cover outages more broadly for energy clients; specialist and finite products exist, though availability is limited.
Alternatively, power can be generated on-site behind the meter – typically gas reciprocating engines or turbines, sometimes on-site wind or solar. Looking ahead, small modular reactors (SMRs) are being actively explored as nuclear advances, and pumped-storage hydropower offers a low-carbon option, though drought and shifting weather may affect long-term reliability.
Supply can be structured as self-supply or contracted from third parties, with BTM power generated on-campus or sourced 'over the fence' from an adjacent facility. Across every model, the interdependency of risks is key – particularly where a loss at the data centre stems from damage to a separate but connected power source. If that asset becomes unavailable, understanding the system's resilience is critical: is there enough redundancy or alternative supply, and what's the financial and operational impact of an outage?
Even with significant built-in redundancy, interruption can't be eliminated entirely. To soften the financial impact, developers can explore alternative risk transfer – including cover for revenue lost to short-duration outages, within the first 48 hours. Traditional power market policies often carry 30–60 day waiting periods that sit awkwardly with risk profiles; the parametric market can offer more responsive, tailored solutions.
Ultimately, energy availability remains a key factor in siting a data centre – but it must be weighed alongside the resilience and continuity of supply.
Ultimately, energy availability remains a key factor in siting a data centre, but it must be weighed alongside the resilience and continuity of supply.
Recommendations:
When assessing a backup power strategy, the following questions are vital:
- Technology selection: What technology is being deployed, and are spare parts readily available? How quickly can key components be repaired or replaced?
- UPS performance: How long can the uninterruptible power supply operate before generators are required to take over?
- System endurance: What is the maximum runtime of the UPS under full load conditions?
- Testing and maintenance: How frequently is the power system tested? Backup batteries and UPS systems can degrade or discharge over time if not properly maintained.
- Generator specification: Does the facility require emergency standby, prime, or continuous power generation capability?
- Load prioritisation: Which systems can be safely shut down to preserve power during an outage?
- Documentation and governance: Is the backup power strategy clearly documented, and how regularly is it reviewed and updated?
- Contracted supply risk: Where power is externally sourced, what are the cost and availability implications of switching to alternative supply in the event of disruption?
A structured and well-tested approach to backup power is essential to maintaining operational resilience and protecting revenue streams in the event of supply interruption.
