AI's Power Trip Will Leave Energy Grids Begging For Mercy By 2027

AI-driven datacenter energy demand could expand 160 percent over the next two years, leaving 40 percent of existing facilities operationally constrained by power availability from 2027.

This is according to Gartner, which estimates the energy required for bit barns to run additional AI-optimized servers is forecast to hit 500 terawatt-hours (TWh) per year in 2027, which it says is 2.6 times the level seen in 2023.

Microsoft accused of 'greenwashing' as AI used in fossil fuel exploration

READ MORE

"The explosive growth of new hyperscale datacenters to implement GenAI is creating an insatiable demand for power that will exceed the ability of utility providers to expand their capacity fast enough," claimed Gartner VP Analyst Bob Johnson.

"In turn, this threatens to disrupt energy availability and lead to shortages, which will limit the growth of new datacenters for GenAI and other uses from 2026," he said.

Gartner is just the latest industry watcher to warn about the impact on energy supplies caused by massive additional investment in high-performance compute triggered by the Gen AI craze over the past 18 months or so.

In June, Omdia reported that capital expenditure on datacenters is likely to be up nearly 30 percent in 2024 alone, with AI on track to become the top server workload within a few years.

Likewise, analyst IDC expects the surging demand for AI workloads to lead directly to a substantial increase in capacity, energy consumption, and greenhouse gas emissions. Bit barn capacity is projected to experience a compound annual growth rate (CAGR) of 40.5 percent through 2027, it claims.

IDC modelled three future scenarios using energy pricing and growth rates for the United States, Germany, and Japan. In all three, an increase in electricity costs was forecast to exceed a CAGR of 15 percent.

Karim Amin, EVP for the Siemens Energy Generation division, said this month the proportion of global energy consumption accounted for by data facilities is set to more than double before 2030. Writing in a LinkedIn post, Amin warned the world currently does not have enough generation or transmission capacity to power the datacenter projects that are in the pipeline.

This won't come as a great shock to Reg readers: we've been reporting for some time that datacenter power demands are outpacing the ability of energy grids to add capacity, causing major concern across the industry.

Last month, management consultancy Bain & Company issued a report indicating that rising energy use in America could outstrip supply within just a couple of years. US energy companies, it warned, need to adapt quickly from previous strategies that prioritized efficiency in the face of flat or shrinking demand.

While the energy grid operators grapple with how to meet soaring demand, Schneider Electric said recently that bit barn operators should consider generating their own energy on-site, or at least some of it.

• UK government plays power broker with small modular reactor suitors
• Meta's plan for nuclear datacenter reportedly undone by bees
• Amazon's nuclear datacenter dreams stall as watchdog rejects power deal
• Datacenter developer says power issues holding up new builds

VP of Innovation and Datacenter Steven Carlini said in the short term this would likely mean deploying gas-powered turbines to generate electricity, but favors the nuclear option for the medium to long term.

Gartner's Johnson claimed power shortages are likely to continue for some time as "new power transmission, distribution, and generation capacity could take years to come online and won't alleviate current problems."

So much for green Google ... Emissions up 48% since 2019

READ MORE

Some advice Gartner offers organizations is to anticipate higher power costs when evaluating plans, and negotiate long-term contracts for datacenter services at reasonable rates for power. Businesses should also consider alternative approaches that require less power.

Sustainability goals may be negatively affected by short-term solutions to provide more power, Gartner says. Suppliers may decide to keep fossil fuel plants in operation beyond their scheduled retirement date, for example.

A similar warning was already sounded by financial services biz Morgan Stanley, which forecast that generative AI will cause three times more greenhouse gas (GHG) emissions by 2030 than if it had not been developed. ®