Why New York Has Paused New 50MW+ Data Centre Projects

The AI boom has run headlong into the limits of the power system. By introducing a one-year freeze on new hyperscale data facilities, New York has become the first US state to formally halt development of data centres using 50MW of power or more for up to one year.

The order was signed by New York Governor Kathy Hochul on 14 July, with the aim of creating a more comprehensive environmental regulatory framework.

Although the moratorium applies across all commercial data storage facilities of this scale, it was driven largely by the extraordinary electricity demand created by the generative AI surge.

“We’re in the midst of one of the most significant economic upheavals in generations… perhaps ever,” Kathy says.

“These hyperscale AI data centres consume enormous amounts of power, truly threatening to outpace our grid’s capacity. They drive up costs for local ratepayers and I refuse to let those costs get passed down to New Yorkers.”

During the temporary pause, state officials will draft a Generic Environmental Impact Statement to assess how these facilities affect regional power grids, water reserves and air quality.

While the policy is local, its implications reach well beyond New York. It signals to the tech sector that unrestricted digital infrastructure expansion is now running into hard energy constraints.

Why New York has paused 50MW data centres

The main reason behind New York’s intervention is the volatile and unusually heavy power demand of AI data centres. Traditional data facilities tend to draw a relatively steady and predictable level of electricity from local grids.

AI infrastructure behaves very differently. The intensive training of foundation models can require enormous amounts of power over multi-week periods, then fall into much more variable usage once systems are serving live queries.

That unpredictability is making life difficult for utilities. More than two thirds of utilities, or 77%, are struggling to forecast the energy demand linked to the rapid growth of AI-driven data centres, according to Capgemini’s report, AI meets the grid: Shaping the data center power play.

The scale of demand is also increasing rapidly. Capgemini projects that electricity consumption from AI training and inference will rise from 25% to 60% of total data centre power demand within the next three to five years, crowding out more conventional digital workloads.

The physical load is simply too large for many existing regional grids to absorb without consequences. While a standard data centre consumes roughly the same amount of electricity as 100,000 homes, the International Energy Agency estimates that newer AI campuses now under development will require up to 20 times that amount.

If left unchecked, this growth could push up electricity bills for households, strain water supplies used for cooling and place added pressure on local infrastructure.

Regulatory controls and energy maps

New York’s pause is not an isolated move but part of a broader shift towards tighter oversight. Governments are increasingly moving away from simply encouraging data centre investment and are instead demanding more accountability around energy use and system impacts.

In Europe, regulation is already changing how operators design and run sites. Germany, for example, has introduced energy efficiency rules requiring data centres to reuse at least 20% of their waste heat by 2028, often by feeding it into district heating networks. However, requiring these changes before the supply chain is fully ready has created friction.

“It was very, very optimistic for Germany to put that law in place before the industry was actually ready for it,” said Mandar Pandit, Chief Strategy and Growth Officer for Data Centres and Electrification Systems at GE Vernova, during a Schneider Electric press briefing on data centres in May.

“While there are indeed technologies to take that waste heat and turn it back into electricity, we are currently seeing waste heat deployed more successfully in things like greenhouses and industrial processes.”

Beyond temporary bans, governments are also using financial tools to manage grid risk. New York is exploring the creation of a Grid Acceleration Fund that would require developers to help pay for public grid upgrades and clean energy generation, alongside specialist insurance mechanisms for speculative power loads and the possible removal of sales tax exemptions.

As a result, these regulatory pressures are redrawing the global map for technology investment. Rather than clustering in traditional tech hubs, developers are looking for places with stronger grid stability and more abundant power.

India, for instance, is sidestepping legacy grid constraints by combining state-backed solar programmes with High-Voltage Direct Current transmission lines to move clean electricity over long distances to high-demand hubs, putting the country on a path towards becoming a global 'electrostate'.

“The Indian government and Indian companies are planning a lot of new data centres,” says Gerhard Salge, CTO at Hitachi Energy, who spoke with us on how AI data centres can become “good citizens”.

Engineering the good grid citizen

Faced with the threat of lengthy development freezes, data centre operators and hardware makers are redesigning their technical systems.

The industry’s focus is shifting away from simply maximising processing power and towards ensuring grid compatibility and greater self-sufficiency.

To maintain public confidence and reduce pressure on local power networks, technology companies are changing how they work with utilities.

Gerhard emphasises: “To be a ‘good citizen’ and a good partner, data centres need to work closely with power companies from day one.”

That means developers must coordinate demand profiles with utilities well before construction begins. From an engineering standpoint, major hardware firms are also working on more compact, energy-dense systems. Hitachi Energy is currently working with NVIDIA on a new 800-volt architecture designed to significantly reduce infrastructure size.

While current transformers, converters and uninterruptible power supply systems can technically support 800-volt operations, innovation is still needed to overcome space limitations.

“One of the very popular discussions these days is what you can achieve with solid-state transformer concepts, and that is also something we are looking at together with NVIDIA and others,” adds Gerhard.

Operators are also deploying advanced power electronics such as Static Synchronous Compensators to buffer operations from the public grid. By converting power from AC to DC and back again, these systems fully isolate the internal environment of a data centre.

When AI graphics processing units create sudden power spikes, on-site battery storage can absorb or release energy instantly. That means the public grid sees a smoother, more stable load rather than a sharp disturbance.