Schneider Electric: Why Retrofitting is Key For Data Centres

The question of energy is fast becoming a hot potato for the companies behind AI data centres.

The issue for operators is not only how to scale fast enough, but how to power and cool this next generation of infrastructure sustainably.

For Steve Carlini, Chief Advocate for Data Centre and AI at Schneider Electric, the fastest way to meet the demand isn’t through new construction but by transforming the energy performance of existing sites.

Retrofitting older data centres for AI workloads, he says, is both a strategic and sustainable way to deliver capacity where it’s needed most.

Why brownfield sites are so important

While hyperscalers invest in large-scale new builds, smaller and mid-sized operators are turning to brownfield sites where power connections and cooling systems already exist. These projects can accelerate deployment while easing strain on local grids.

“Sites that exist will already be permitted as a data centre and can circumvent any lengthy permitting cycle,” Steve explains.

“Additionally, many existing data centres may be strategically located close to the data sources or applications, which can have significant advantages in faster speed, lower latency and lower data transfer costs.”

That proximity delivers both operational and energy gains. As AI shifts from centralised model training to distributed inference workloads, placing compute power close to demand reduces energy wasted on long-haul data transfer and helps balance overall grid usage.

Choosing power-positive sites

When selecting facilities for retrofit, energy capacity is the deciding factor. “Sites that have an abundance of utility power are pure gold as accelerated compute AI requires more power,” Steve notes.

“Second are sites where more utility capacity can be added quickly.”

Sites with strong utility connections or grid expansion potential can support denser AI workloads without major new infrastructure. But access to power isn’t the only concern.

“Additional considerations, picking sites that are isolated from neighbourhoods and highly populated areas,” Steve advises. “These areas can be difficult for the new breed of data centres with more generators and chillers that can make quite a bit of noise and can trigger complaints.”

This makes location strategy a crucial energy planning issue as operators weigh grid capacity, community impact and cooling requirements in tandem.

Unlocking the business and energy returns

The drive to retrofit isn’t only about grid access – it’s a business opportunity aligned with the broader energy transition.

“Many data centre operators would like to add accelerated compute AI and become and run applications for in-house application automation or offer pay for AI models,” Steve says. “The monetisation of AI working models or inference is the next big wave.”

As AI projects move from pilot to production, smaller data centres with efficient power systems are ideally positioned to capture the inference market.

Sites near end-users can operate with lower energy transmission losses, making them competitive on cost and sustainability.

The challenge of upgrading power and cooling systems

Schneider Electric solutions seen in the Start Campus data centre (Credit: Schneider Electric)

Retrofitting brings technical hurdles, particularly in modernising electrical distribution and cooling systems. “On the power side, the main issue is grid power,” Steve states. “The second issue will be the entire power distribution inside the data centre.

Traditional data centres were designed for lower power densities or distributed workloads. AI workloads demand concentrated power delivery, which may require upgrades to: PDUs, medium-voltage switchgear, low-voltage switchgear, transformers, circuit breakers and busways or cabling.”

Cooling poses just as big an energy challenge. “On the cooling side, most next-generation AI servers are natively liquid-cooled and come with integrated cold water inlet and hot water outlet connections,” Steve explains. “These are not optional – they are required for operation.”

Implementing liquid cooling requires:

• CDUs (Coolant Distribution Units) to manage flow between servers and the facility’s cooling loop

• Chillers or dry coolers designed for liquid operation

• Piping infrastructure for water circulation

• Heat exchangers connecting server loops to building systems

• Monitoring and control systems for temperature and flow management

Even with liquid systems, Steve notes, up to 30% of the load still needs air-based cooling, keeping hybrid approaches central to energy planning.

The efficiency dividend

The environmental benefits of retrofitting are equally significant. Steve points to the energy and water efficiencies that come with modern systems.

“Closed loop liquid-cooling systems use much less water than traditional cooling,” he says. “Additionally, a new generation of 800VDC electrical power distribution to the AI servers will use less current and produce less heat in the future.”

These innovations cut both operational emissions and resource intensity, supporting corporate ESG goals and national energy efficiency targets. For operators, the ability to scale AI without proportionally increasing energy consumption is a decisive advantage.

2026: The year energy and AI converge

Steve sees 2026 as the point where AI’s evolution and energy infrastructure intersect.

“As production-ready AI inference applications start gaining momentum, companies improve their business process efficiency and they start to automate their business processes with agentic AI and eventually start Artificial General Intelligence (AGI), each progression will require significantly more computing horsepower enabled by data centres,” he explains.

“These applications will benefit from being located in data centres that are located closest to the application and data generation and processing, which will be in smaller, retrofitted data centres.”

The shift from experimental AI to mainstream implementation will demand far more distributed energy availability. Retrofitting gives operators the ability to expand capacity quickly, sustainably and in alignment with grid realities – a crucial step in meeting the energy demands of the AI era.