The rapid growth of data centers is placing unprecedented demands on the grid. Facilities that would once have been considered exceptionally large are now being joined by campuses measured in hundreds of megawatts or even gigawatts, driven by growing demand for AI and digital services.

As these facilities expand, their relationship with the grid is also changing. Operators are being asked not only to consume power reliably, but also to help maintain stability when disturbances occur.

In this DCD>Talks episode, Flemming Johansen, senior principal architect, research and technology at Schneider Electric, explains why fault ride through requirements are becoming increasingly important and how data centers can become better neighbors to the grid.

When data centers become part of the grid

“Data centers are growing not just in numbers, but also in size,” says Johansen.

A 50MW facility would once have been considered a large data center. Today, operators are planning campuses measured in gigawatts. At the same time, increasing adoption of AI, electric vehicles and renewable energy is placing additional pressure on power infrastructure.

The result is greater competition for available grid capacity and longer queues for new connections.

Data centers also tend to cluster in specific locations, creating another challenge for grid operators. If multiple large facilities disconnect simultaneously following a disturbance, the sudden loss of demand can create instability across the wider network.

“A disconnection will create an imbalance between the production and the demand side, causing frequency problems and destabilization” Johansen explains.

To address this risk, grid operators are introducing new requirements designed to ensure large electrical loads remain connected during smaller disturbances rather than disconnecting immediately.

Understanding fault ride through

These requirements are commonly known as fault ride through (FRT).

Historically, data center UPS systems were designed to protect critical loads by disconnecting from the grid whenever a disturbance occurred. Even a minor voltage dip could trigger a transfer to battery power.

According to Johansen, that approach is becoming harder to justify as facilities grow larger. To explain the concept, he uses a simple analogy.

Imagine two people holding opposite ends of a rope. One represents the data center, while the other represents the grid. Both are pulling with equal force, maintaining a stable balance between electricity supply and demand.

If the person representing the grid briefly loses their footing, the traditional response would be for the data center to let go of the rope entirely.

“The grid person, he would fall over, because now he's totally out of balance,” says Johansen.

Under an FRT approach, the data center instead adjusts its own behavior, reducing its demand temporarily while helping the grid regain stability. Once balance is restored, normal operation resumes.

“The balance is now kept, even though there was a small glitch on the grid,” he explains.

Rethinking the role of the UPS

For data center operators, FRT requirements are changing how UPS systems are expected to behave during grid disturbances.

Modern UPS systems operating in double conversion mode already isolate critical IT loads from events occurring on the grid. However, FRT introduces new requirements governing how these systems interact with the wider power network.

Johansen explains that the UPS response depends on the severity of the disturbance.

During smaller voltage dips, the UPS can continue drawing some power from the grid while supplementing the remainder from battery storage. If voltage falls further, the UPS may temporarily rely entirely on battery power to avoid placing additional stress on an already weakened grid.

Once grid conditions recover, power consumption must return quickly. This process, known as post-fault active power recovery, is becoming part of emerging grid requirements.

“In a UPS that we have designed, we can do that within 0.3 seconds,” says Johansen.

The objective is not simply to protect the data center, but to restore balance across the wider power system as quickly as possible.

Regulation drives a new approach

One of the challenges facing operators is that FRT requirements vary considerably between regions.

Unlike some aspects of data center design, there is currently no harmonized global standard. Instead, individual transmission system operators (TSOs) are developing their own requirements based on local grid conditions.

“If you're building a data center, you need to understand which FRT requirements apply to your area,” says Johansen.

Compliance also extends beyond the UPS itself. Requirements typically apply to the data center as a whole, meaning operators must consider the behavior of mechanical systems alongside critical IT loads.

Johansen also expects simulation models to play a larger role. Grid operators are beginning to request detailed models that allow them to evaluate how facilities will behave under different fault scenarios before they are connected.

Complicating matters further, many regulations remain under development. Requirements can continue to evolve even before they become mandatory, making ongoing engagement with local TSOs essential.

Being a good neighbor

For Johansen, the wider message is straightforward. As data centers continue to grow, operators must think beyond their own facilities and consider their role within the broader power system.

“The takeaway is that if you build a data center, you need to be a good neighbor,” he says.

That means understanding local requirements, considering the behavior of the entire facility during disturbances and working closely with grid operators to maintain stability.

As power demand continues to grow, the relationship between data centers and the grid is becoming more closely intertwined. The facilities that succeed will not simply be those that secure enough power, but those that help keep the system around them in balance when the grid stumbles.

Watch the full DCD>Talks episode with Schneider Electric’s Flemming Johansen here.