Simon Flowers and Chris Seiple of Wood Mackenzie report that a shake-up of regulation and processes is essential to meet soaring demand. They write:
How much is power demand going to grow?
No one knows for sure, not even the technology companies closest to understanding their own electricity needs. But when we look at the range of possible outcomes and the drivers, US demand growth is likely to be at least 2% to 3% a year through 2030. The other critical uncertainties are how much demand utilities will allow to connect to their system and how much growth we’ll see in behind-the-meter generation that’s unable to connect to the grid. All in all, it’s a huge change after virtually no growth since the turn of the century.What’s driving the growth?
Three main factors. First, explosive growth from data centres to power the adoption of AI and the complex large language models used by bots like ChatGPT. Second, the industrial renaissance in energy-intensive cleantech manufacturing – batteries, solar equipment and semiconductors. Third, the wider electrification of the economy, with EVs and electric heating making ground in some parts of the country. […]Will there be enough power supply?
That’s one of the industry’s biggest challenges. Supply has already been reduced in some systems, with grid operators cutting ‘accreditation’ – effectively downgrading the capacity they regard as reliable. Planned coal retirements will exacerbate the reduction in supply. Projected lead times of three years or more for new transformers and breakers will delay the grid-infrastructure upgrades needed to accommodate the new supply.And new supply is looking to come onto the system. We forecast renewables capacity additions will increase from 29 GW a year currently to 40 GW, barely enough to keep up with demand growth, let alone enough to decarbonise the power sector. New capacity beyond this level will run into permitting and interconnection constraints.
Importantly, intermittent renewables paired with storage won’t satisfy the 24/7 high-load needs of new manufacturing and data centre demand. There will likely be a diverse mix of alternatives with new gas capacity and delayed coal retirements near-term solutions providing flexibility – albeit with climate implications. Nuclear SMRs are among the options further out. […]
Will power prices go up?
That’s highly likely if demand comes through as we expect, and supply can’t respond quickly enough – particularly in systems where demand growth is strongest. The major uncertainty is how this plays out in different markets. In those with vertically integrated monopoly utilities, it will be easier for utilities only to add load when they have supply to accommodate it. In more deregulated markets, where retail customers have choice, there will be more volatility and upward pressure on prices, which could lead to regulatory intervention.[…]The US advantage is that, outside of China, all the major tech companies investing in AI and large language models are based there. The US administration, like any government, will be at pains to ensure as much processing and IP as possible resides in the US.
Read more here.
