As data centres become more important to the digital economy, the definition of efficiency is changing.
For years, discourse has centred around how efficiently a facility uses the power it receives. Metrics such as Power Usage Effectiveness (PUE) have helped operators identify opportunities to reduce energy overheads, particularly across cooling and other facility systems.
But the next phase of data centre growth brings a different challenge. AI workloads are increasing power requirements, while grid constraints, energy price volatility and sustainability expectations are making access to reliable, lower-carbon electricity just as important as how efficiently that electricity is consumed.
And this challenge extends beyond the data centre itself. As large facilities place greater demands on local power networks, the way they generate, store and manage energy can have implications for the wider grid and the communities that depend on it.
So, what if we can address data centre efficiency not just through what happens inside the facility, but through the energy infrastructure around it?
Efficiency starts before power reaches the rack
The rapid expansion of AI is putting pressure on every layer of data centre infrastructure. Higher-density computing requires more power and generates more heat, but the challenge starts well before the rack, with how electricity is generated, supplied, converted, stored and managed.
S&P Global’s 451 Research estimates that Asia-Pacific data centre demand could reach 493 TWh by 2030, up from 267 TWh in 2025, even in its low-end case. At the same time, grid constraints and volatile power markets are increasing the cost of inefficiency.
For operators, this means looking beyond consuming less electricity to having greater control over how energy is sourced, distributed and managed as grid conditions change.
This is where microgrids can play a role.
From backup power to an active energy system
We know that a microgrid is more than just a collection of backup generators and batteries.
When properly designed, it can bring together distributed energy resources such as solar, energy storage and other forms of generation with power conversion, monitoring and intelligent controls.
The objective is to give data centres greater visibility and control over their energy ecosystem. Beyond improving operational resilience, more flexible local energy systems can help ease pressure on the wider grid by reducing reliance on grid power during periods of peak demand. When integrated with renewable energy sources such as solar and battery storage, these systems can also provide cleaner power and contribute to greater energy resilience for surrounding communities.
For example, energy storage can help manage peak demand, while Distributed Energy Resources (DER) can reduce reliance on grid electricity when conditions allow. Intelligent controls can coordinate these resources according to factors such as demand, electricity prices, renewable availability and grid conditions.
Rather than relying solely on the wider grid to meet their energy needs, data centres can take greater ownership of their energy ecosystem, with more control over how power is sourced, stored and managed to support their operational requirements.
That distinction will become increasingly important as data centre loads become more dynamic. AI workloads can create significant changes in power demand, while high-density computing and advanced cooling systems introduce new operational requirements. A more flexible energy system can help the facility respond to these changes rather than simply building more capacity around them.
Resilience and efficiency are becoming connected
There is also a case for looking at efficiency and resilience together.
A data centre cannot afford to optimise energy consumption at the expense of reliability. Equally, maintaining resilience through excessive redundancy or inefficient operation comes with an energy and cost penalty.
The opportunity is to design energy infrastructure that supports both.
On-site generation and energy storage, for instance, can provide additional flexibility during periods of grid constraint. Demand response can help manage peak loads. Intelligent energy management can coordinate different energy sources and loads, allowing operators to respond to changing conditions while maintaining the reliability required by mission-critical infrastructure.
Industry research points the same way: decarbonisation efforts are increasingly expected to improve operational efficiency and resilience, rather than serve solely as sustainability initiatives. On-site generation, energy storage are widely cited as ways organisations can reduce exposure to grid instability.
IDC reports that 68% of Southeast Asian data centre operators plan energy management and monitoring retrofits in the next 24 months, uniquely high compared with other regions. This reflects the need to optimize efficiency rapidly in this region as scale grows, particularly across markets with variable grid reliability and rising energy costs.
The next step is connecting the pieces
However, installing more energy assets alone will not solve the problem.
A microgrid needs to operate as part of a wider energy management system. Energy, carbon and asset data need to be brought together so operators have a common view of what is happening across the facility.
This is where digitalisation becomes important.
The progression should start with visibility: metering and sensors establish a trusted baseline of energy use. From there, automated controls can respond to predefined operating conditions. More advanced systems can then use analytics and AI to optimise energy, carbon and operational requirements in real time.
For data centres, this could mean coordinating grid power, DER, energy storage, cooling systems and IT loads rather than optimising each component independently.
The result is a move away from isolated efficiency improvements towards an integrated energy system.
Building data centres for the energy reality ahead
The data centre industry is entering a period where adding more capacity alone will not be enough. Operators must consider how that capacity interacts with the energy systems supporting it.
This is particularly relevant in markets across Asia where data centre investment is growing alongside constraints on power availability, land and infrastructure. As more facilities compete for electricity, the ability to use energy intelligently could become as important as the amount of energy available.
Microgrids should therefore be considered not simply as an alternative source of power, but as part of the infrastructure strategy for the next generation of data centres.
This approach is already moving into the mainstream in the US, where long grid connection delays have pushed developers to treat on-site generation, storage and microgrids as a primary strategy. Many of these projects rely on gas-fired generation paired with batteries, so the emissions benefit depends on how each system is designed.
These systems also typically carry a cost premium, so operators need to weigh the gains in speed and resilience against the added expense.