AI Data Centres
UPS stands for uninterruptible power supply, and depending on how you are reading this article, your device might already be using one. A UPS keeps electronic devices running when there’s a power outage. In simple terms, it’s a backup. At home, they are often used to protect desktop computers or gaming setups, giving you a few minutes to save your work, or your level, before everything shuts down. But the UPS under your desk is small compared with the systems used inside data centres.And demand for those systems is rising fast.
As AI, cloud computing, and digital services expand, the infrastructure supporting them has to grow as well. In 2024, the UK designated data centres as critical national infrastructure, recognising the role they play in keeping modern society running. Healthcare records, banking systems, emergency services, transport networks, and countless other services all depend on data centres operating continuously.
Data centres are essentially hotels for computers. They are large buildings packed with racks of servers designed to run 24 hours a day, seven days a week. If an online shop goes down, it’s frustrating. But if banking systems fail, hospitals lose access to patient records, or emergency services lose access to critical databases, the consequences can quickly become far more serious.
That’s why the data centre UPS market is booming, expected to grow from around $8.76 billion in 2025 to roughly $12.47 billion by 2030, as operators invest in the infrastructure needed to guarantee uptime.
The first line of defence
Think of a UPS as the first line of defence when something goes wrong with the power supply. When the grid experiences an outage, voltage dip, or sudden surge, UPS systems instantly step in to protect the IT load. They ensure servers, storage systems, and network equipment continue operating without interruption. But it’s important to note that UPS systems are not designed to power a data centre for hours.
Instead, they bridge the gap between a grid failure and backup generation starting up. In most facilities, that backup comes from diesel generators, which typically take a few seconds to reach full output.
Those few seconds might not sound like much, but for sensitive IT equipment they are critical. Without a UPS, even a momentary interruption could shut down servers, damage hardware, or corrupt valuable data.
In practice, UPS systems serve several key roles:
- Acting as the first line of defence when the primary power supply fails
- Protecting equipment from voltage drops, spikes, and surges
- Preventing sudden shutdowns that could damage servers or corrupt data
- Maintaining continuous operation while backup generation starts up
Lead-acid: the original UPS battery
For decades, the most common UPS technology in data centres has been lead-acid batteries.
The reason is simple. Lead-acid batteries are a mature and well-understood technology. They are relatively inexpensive, reliable, and capable of providing short bursts of backup power when needed. However, they come with several drawbacks.
Lead-acid batteries have a relatively short cycle life compared with modern alternatives. They also degrade more quickly when repeatedly charged and discharged, which can lead to higher maintenance and replacement costs over time.
As data centres have grown larger and power demands have increased, operators have begun looking for alternatives that offer better performance and durability. And that’s where flywheels entered the conversation.
Flywheels: perfect for rapid response
Flywheel energy storage takes a completely different approach to backup power. Rather than storing energy chemically like a battery, flywheels store energy as kinetic energy in a rapidly spinning rotor. These rotors can spin at extremely high speeds, often between 20,000 and 50,000 revolutions per minute, inside low-friction vacuum chambers. When electricity is needed, the spinning rotor drives a generator, converting that stored kinetic energy back into electrical power.
Flywheels offer several advantages over traditional lead-acid batteries.They can deliver power almost instantly, recharge extremely quickly once the grid returns, and tolerate a huge number of charge and discharge cycles because there is no chemical degradation involved. As with everything, they also come with limitations. Their energy storage capacity is relatively small, meaning they typically provide power for seconds rather than minutes. That makes them ideal for bridging very short interruptions, but less suitable for longer outages.
Flywheels also need to remain spinning at high speed so they can respond instantly when power fails. Modern systems minimise friction using magnetic bearings and vacuum chambers, meaning they require only a small amount of energy to maintain that speed, but once the rotor slows down, the stored energy is gone. So there’s a tradeoff. Flywheels offer speed and durability, while batteries offer longer runtime. But what if there didn’t have to be?
The rise of modern lithium batteries
Advances in lithium battery technology have transformed what UPS systems can do. Compared with traditional lead-acid systems, lithium batteries offer significantly higher energy density, longer lifespans, and far better cycle performance. They also take up less space, which is a valuable advantage inside large data centre facilities.
But the real breakthrough is happening in next-generation battery chemistries. Novel battery chemistries including new elements such as niobium, allow batteries to charge and discharge extremely quickly while dramatically extending their operational life. Some of these systems can retain around 80% capacity even after 25,000 cycles, compared with roughly 1,000 cycles for conventional lithium-ion batteries.
They also support extremely high C-rates, allowing them to respond to power fluctuations in microseconds and recharge far more quickly than traditional battery technologies. The combination of rapid response and long life makes lithium-ion UPS an increasingly attractive option for modern UPS systems.
Modular UPS systems
Alongside improvements in battery chemistry, the architecture of UPS systems is also evolving. Many operators are now moving toward modular UPS designs rather than relying on the traditional monolithic systems.
Modular UPS systems consist of multiple smaller units working together, allowing capacity to be expanded as facilities grow. This approach provides several advantages:
- Scalability, allowing operators to add capacity as demand increases
- Redundancy, meaning individual modules can fail without affecting the entire system
- Simpler maintenance, since modules can be replaced individually
For rapidly growing data centres, this flexibility is becoming increasingly valuable.
The infrastructure keeping the internet alive
UPS systems rarely attract the same attention as AI chips or high-performance servers, but they play a crucial role in keeping the digital world running. They ensure that even the briefest power interruption does not cascade into system failures, data loss, or equipment damage. As data centres expand to support the growing demands of AI, cloud computing, and digital services, the technologies protecting their power supply are evolving just as quickly. Because in a world that runs on data, even a few seconds without power can be too long.
