Artificial intelligence is moving fast, much faster than the power systems built to support it. Across the world, data centres that feed AI models are multiplying, and so is their hunger for electricity. What was once a technical issue for a few tech companies has become an energy challenge with global implications.
Companies like Google, Amazon and Microsoft are racing to build larger data facilities, each packed with thousands of servers that run non-stop. These machines demand huge amounts of power and cooling, often more than the grid can comfortably provide. Keeping them online is no small task. It takes flexibility, backup capacity and smarter use of renewable energy, which is why energy storage is quickly stepping into the spotlight.
According to the International Energy Agency, global data centre power use could reach nearly 945 terawatt-hours by 2030, more than double today’s levels. In the United States, the sector might even surpass all traditional heavy industries combined. That scale is already reshaping how utilities plan for supply and how governments think about resilience.
As the data economy grows, one truth is clear: the connection between AI and energy storage is no longer technical or optional. It is the backbone of digital continuity.
Why AI Needs Energy Storage?
Data centres are designed for constant operation. Even a few seconds of downtime can disrupt millions of online interactions, financial transactions or AI-driven processes. But as their energy appetite grows, so does the strain on local grids, increasing the risk of interruptions.
At the same time, major operators are racing to decarbonise. “They don’t just need enormous amounts of power; they need renewable power,” says Michael Hunter, head of commercial at Apatura’s UK data centre portfolio.
Renewable generation, though expanding, is intermittent. In the UK, this mismatch between supply and demand could cost consumers around £3 billion by 2030, according to grid estimates.
Energy storage provides a practical solution. “It helps stabilise supply, reduce pressure on the grid and make full use of renewable generation,” explains Rehaan Aleem Shiledar, senior energy analyst at GlobalData. Batteries can also lower costs through smart energy management, charging during off-peak hours and discharging when electricity prices rise.
Equally important, storage strengthens cybersecurity by enabling more decentralised, self-sufficient power systems. For hyperscalers running critical AI workloads, this kind of resilience is not optional — it is fundamental.
Batteries Take the Lead
Among storage technologies, batteries have become the undisputed favourite. Hydrogen and pumped hydro solutions still face high costs or geographic limitations, while Battery Energy Storage Systems (BESS) are modular, scalable and cost-effective.
“BESS delivers flexibility and operational savings, and it can even serve as part of the uninterruptible power supply,” says Hunter. “That combination of benefits makes it the most practical choice today.”
In recent years, battery technology has improved dramatically. “Costs have fallen and technical performance has improved to levels we could not have imagined before,” adds Andrés Acosta, Director of Innovation at LevelTen Energy.
Lithium-ion remains the leading chemistry, but long-duration batteries capable of storing energy for many hours are rapidly emerging. These innovations could redefine how data centres and power systems interact in the coming decade.
Scaling Up for the AI Era
AI’s energy needs will only grow. Meeting them sustainably requires larger and smarter storage capacity. “Improving energy density and extending battery life will be key,” says Shiledar. “Government support through incentives and clear regulations can make that possible.”
The UK’s 2025 AI Opportunities Action Plan is an example of such alignment. It created dedicated “AI Growth Zones” where data centre development is paired with land and grid capacity for battery storage.
But hyperscalers are not waiting for public policy. Many are already investing directly in storage partnerships. These arrangements allow data centre operators to secure stable, renewable power while giving developers the financial backing needed to expand.
“Policymakers can connect the dots between tech, energy and infrastructure,” Hunter notes. “But private collaboration is already driving results.”
Building Smarter Business Models
The commercial side of this collaboration is evolving just as quickly. Traditional renewable power purchase agreements (PPAs) are becoming less effective as market volatility increases and renewable generation saturates some grids.
“Adding storage changes the equation,” says Acosta. “It allows buyers to secure energy when they need it most and helps sellers maintain predictable revenue streams.”
Some developers rely on tolling agreements that offer stable returns. Others use mixed models that combine fixed prices with merchant exposure to capture more market value. The Storage-as-a-Service (STaaS) approach is also gaining traction, allowing companies to access battery capacity without large upfront investments.
Energy-integrated service level agreements (SLAs) now link performance guarantees with renewable sourcing, ensuring that sustainability targets are met in measurable ways. These evolving business models are helping storage mature from an emerging technology into a mainstream infrastructure asset.
Power Loop Insight
For Power Loop readers, one thing is clear: energy storage is no longer a secondary feature of data centres. It has become a defining part of their design and strategy.
By incorporating battery storage directly into their operations, hyperscalers are stabilising the grid, improving efficiency and accelerating the transition to renewable power. The result is not only cleaner AI operations but also a stronger, more flexible electricity system.
As Hunter concludes, “We can’t rely on policy alone. Innovation and collaboration in the private sector are essential to achieve scale.”
Energy storage now sits at the crossroads of digital progress and sustainable power. As AI continues to grow, the battery systems supporting it will play a central role in shaping a smarter and more resilient energy future.
Sources:
Future Power Technology, August 2025 – “How Energy Storage and Battery Deals Enable AI Data Centres and Vice Versa.”
International Energy Agency – Energy demand from AI Global Data Centre Outlook 2025.
