Download your complimentary copy to read the top eight most pertinent questions and answers

.
Opinion

Gridlock: Q&A on how the US power industry can cope with bumper demand

The decoupling of US electricity demand from economic growth over the last seven decades is set to reverse amid burgeoning data centre development, a resurgence in energy-intensive US manufacturing, and more transport and heating electrification - resulting in electricity demand growth not seen since the 1990s

5 minute read

Wood Mackenzie recently held a live Q&A session with a panel of experts, based on its recent Horizons thought leadership article, entitled Gridlock: the demand dilemma facing the US power industry 

Fill out the form at the top of the page to read the top eight most pertinent questions and answers discussed with our panel during the event, or read on for a sneak peek at just three of the questions asked:

Q: What is the Horizons article about? 

A: The first takeaway from our October edition of Horizons – Gridlock: the demand dilemma facing the US power industry - is that US energy demand growth is not just about data centres and artificial intelligence (AI). Data centres are probably the most important part of the story, but the resurgence of energy-intensive manufacturing is another key element. The electrification of transport is yet another, and it's the cumulative impact of these sources that make this particularly challenging. 

The second takeaway is that the electric utility industry is not really used to dealing with this type of load. The average proposed data-centre size has gone from 150 MW to 300 MW. By way of example, one chip-manufacturing facility currently operating in Arizona could ultimately have more than 1,200 MW of load. 

This raises a lot of questions. If a utility has multiple interconnection requests this size, all requiring major infrastructure, who do they prioritise? How do they protect shareholder interests? How do they protect existing customers? How do they best optimise the interconnection requests they're considering for generation resources with what's happening on the demand side? This is all new territory. 

It's clear that the system is constrained right now in its ability to meet this demand growth. Ultimately, the question of what electricity demand growth is going to be over the next five years is partly a question of what utilities will allow it to be and how fast can we scale the infrastructure needed to accommodate it. 

Most significantly, the utility industry is sitting at the intersection of protecting national security interests, promoting economic growth and helping deal with the challenge of sustainability and climate change. It needs to step up and innovate and find a way to accommodate all this growth and meet this combination of goals. In other words, it needs to show us what the utility of the 21st century is going to look like and be a role model for others. 

The tools are there; harnessing them is the challenge. There's a very big opportunity for renewable energy companies to grab ground here if they can figure out a way to provide firm power supply. 

Q: A lot of the companies that use data centres have zero-carbon goals for 2030 or 2040. What impact is this having in terms of power demand? 

A: There are two main challenges. The first is round-the-clock power supply. The reliability of that supply is highly important, and intermittent renewable resources alone aren't particularly well placed to serve a new load of that kind. The second challenge is that data centres and manufacturing facilities are not typically located next to the significant acreage where you could put a wind or solar farm.

Supplying these data centres using such renewable resources is often going to require a lot of transmission investment, which takes time. Grid-enhancing technologies such as dynamic line rating and advanced power flow optimization could be low-cost accelerators, but utility and grid operator adoption has been slow thus far. 

Because of these challenges, it is very likely that we're going to see increasing amounts of gas-fired generation to solve some of this, given the scale of the loads trying to connect to the system, along with some deferment of coal-fired power plant retirements. This is going to come into conflict with a lot of sustainability goals. 

Q:  What innovative strategies are emerging to address the growing electricity demand of data centres? 

A: No creative models are being deployed widescale yet but there are some interesting things happening. One model is to put the resource behind the meter. The problem there is that if you want to have sustainable resources, you either have a cost problem, as with batteries, or you have a land problem, as with solar, and that ends up taking you towards a gas solution. This may be how things get done for colocation providers, though less so for the hyperscalers with ambitious climate goals. 

The second model is utilities having different technologies they can deploy, such as storage as a transmission asset, grid-enhancing technologies that let them eke out more capacity from their transmission system, and even potentially distributed solar. The benefit here is that they can site assets that serve multiple data centres in a forward-looking way that is going to solve multiple problems at once. 

The third option, which is pretty compelling, is virtual power plants. The key issue here is that the supply side of the grid is suffering from heavy interconnection constraints due to transmission capacity, but the demand side has much more flexibility and can react much faster. So, there are opportunities for other customers that a utility serves to provide support and flexibility to help accommodate data centres. 

There’s an important equity angle here. Rather than focusing on the extent to which rate payers will bear the cost of data centres and the infrastructure that gets built, customers can be compensated through grid services that accommodate the data centre in a sort of public-private partnership (PPP) model. Google, for example, has a programme to subsidise energy efficiency for customers in the vicinity of its data centres, which makes it easier for it to get served. 

One big challenge with regard to PPPs is that data centres are almost always transmission-sited, so you may not have customer DERs that are local to it on the network. They may, therefore, be providing system-level capacity from a utility perspective, but not the kind of local transmission capacity needed. That is where the pinch point is when it comes to serving the customer. 

Read the rest of our top Q&As

To read the top eight most pertinent questions and answers from the session, please fill out the form at the top of the page to download your complimentary copy of the Q&A list.