Slow Down On Building Power Plants For All Those New AI Datacenters, Report Warns

Datacenters are slurping ever more energy to meet the growing demands of AI, but some estimates of future demand imply an increase in hardware that would be beyond the capacity of global chipmakers to supply, according to an environmental nonprofit.

Warnings about the amount of energy that AI datacenters will consume have been getting more strident. A recent report by Deloitte Insights estimated that the total power required by bit barns in the US will increase by a factor of five by 2035, and consultants Bain & Company issued advice to utility companies to revamp the way they operate to support a rapid scale-up of energy resources.

But what happens if those estimates are overinflated? If power companies invest heavily in additional power generation and transmission infrastructure, but datacenter growth does not come near the forecast level, the cost of that expansion would have to be borne by other customers.

For this reason, US nonprofit environmental group the Southern Environmental Law Center (SELC) commissioned an analysis by London Economics International (LEI), a professional services firm specializing in energy and infrastructure. The report can be found here [PDF].

LEI found a lot of uncertainties in forecasting future electricity demand from datacenters, and a resulting bias toward overstating that demand. For example, it says that datacenter builders often duplicate requests to connect a particular facility to the electrical grid across different legal jurisdictions, which can lead to overcounting.

But the real clincher? There may not be enough chips available to support the massive AI buildout forecasts in the first place. If the datacenters can't be built, it stands to reason that excess energy capacity won't be needed.

The authors claim that if all the datacenter projects forecast for the US between 2025 and 2030 were to go ahead, it would require 90 percent of all the growth in the global AI chip supply for the same period to be funnelled to the US market.

Such a scenario is unrealistic, it says, as the US currently accounts for less than 50 percent of global semiconductor chip demand. Meanwhile, other regions in the world are also projecting strong growth to supply their own datacenters.

The LEI's approach was to sum up as many bit barn demand forecasts published by US regional transmission organizations, independent system operators, and balancing authorities as it could find, taking the most conservative estimate from each.

It then worked backwards from these datacenter demand figures to estimate the implied need for semiconductors. These semiconductor figures are not meant to be definitive, the report said, but are more of a "sanity check" on the scale of growth projections.

The report concludes that projections of datacenter electric load growth in the US are "beset by uncertainty," and that there is a greater risk of over-forecasting demand than under-forecasting – and getting it wrong may prove costly for existing customers of the energy firms.

Meanwhile, some US power companies are already set to impose price hikes on consumers because of those pesky bit barns, according to various reports.

The Financial Times said that National Grid, with users in New York and Massachusetts, is to raise rates by $50 a month, while Northern Indiana Public Service Company is upping monthly rates by $23 a customer.

Reuters reports that PJM Interconnection, which serves a number of states clustered near the east coast, is set to increase its energy bills by more than 20 percent this summer. Its area of coverage includes Virginia, home to the largest concentration of datacenter capacity in the world. ®

RECENT NEWS

From Chip War To Cloud War: The Next Frontier In Global Tech Competition

The global chip war, characterized by intense competition among nations and corporations for supremacy in semiconductor ... Read more

The High Stakes Of Tech Regulation: Security Risks And Market Dynamics

The influence of tech giants in the global economy continues to grow, raising crucial questions about how to balance sec... Read more

The Tyranny Of Instagram Interiors: Why It's Time To Break Free From Algorithm-Driven Aesthetics

Instagram has become a dominant force in shaping interior design trends, offering a seemingly endless stream of inspirat... Read more

The Data Crunch In AI: Strategies For Sustainability

Exploring solutions to the imminent exhaustion of internet data for AI training.As the artificial intelligence (AI) indu... Read more

Google Abandons Four-Year Effort To Remove Cookies From Chrome Browser

After four years of dedicated effort, Google has decided to abandon its plan to remove third-party cookies from its Chro... Read more

LinkedIn Embraces AI And Gamification To Drive User Engagement And Revenue

In an effort to tackle slowing revenue growth and enhance user engagement, LinkedIn is turning to artificial intelligenc... Read more