Google Uses AI to Fix Grid Backlog as AI Power Demand Set To Quadruple by 2030

• Google is set to deploy artificial intelligence technology to speed up wait times for connecting the U.S. grid to new power connections.
• The partnership with electrical grid operator PJM Interconnection marks the first time AI has been used this way.
• By 2030, global AI is predicted to use the same amount of energy as Japan uses today.

Alphabet’s Google is partnering with PJM Interconnection, a leading electrical grid operator, to use artificial intelligence (AI) to accelerate the connection of new electricity supplies to the grid.

The move comes amid soaring electricity demand driven by the rapid expansion of AI technologies.

According to the International Energy Agency (IEA), U.S. AI-related power demand is projected to quadruple by 2030, surpassing the total energy usage of all other energy-intensive sectors combined.

Google To Utilize AI for Power

On Thursday, April 10, Google and PJM announced a partnership to reduce the waiting times for connecting the country’s grid to new electricity supplies.

The partnership initiative marks the first use of AI to manage an interconnection queue.

Over the past few years, waiting times for connecting grids have grown to record lengths throughout the U.S.

“The industry has been talking about building smarter grids for well over a decade, and now with AI, we have a real opportunity to turn discussion into action,” Amanda Peterson Corio, Google’s data center energy lead, said at a press conference.

The project is being developed in partnership with Tapestry, an Alphabet-backed startup focused on modernizing the electrical grid to make it more “reliable, affordable, and sustainable.”

“This is really about automating a lot of the things that are being laboriously reviewed,” said Page Crahan, General Manager of Tapestry.

Electric Grid Automation

Google’s partnership will first use AI to automate grid work that is usually completed by hand.

Eventually, the companies will be able to develop an interactive, detailed model of the PJM grid.

“Because the grid wasn’t built to sustain today’s demands, there aren’t adequate tools to see, manage, or plan it,” Tapestry said on its website . “Information is siloed between dozens of different organizations, and no one has a complete picture of the grid.”

Crahan added that the eventual map will bring in “different layers that planners might need to see in a single toggle on-and-off view to guide faster decisions, introduce new insights, hopefully find efficiencies in those ways.”

PJM is North America’s largest electrical grid operator, covering 67 million people.

AI Power Demand Set to Quadruple by 2030

According to a report from the IEA, AI technologies will require as much power as Japan uses today by 2030, with only half of it likely to be powered by renewable energies.

Despite advances in energy-efficient hardware and cooling systems, the sheer scale of AI deployment across industries is outpacing efficiency gains.

However, the report claims that fears of AI destroying climate goals have been “overstated.”

The IEA reported that this is because AI will be able to utilize it to reduce other activities and boost the efficiency of systems, lowering the world’s carbon footprint.

“With the rise of AI, the energy sector is at the forefront of one of the most important technological revolutions of our time,” Fatih Birol, the executive director of the IEA, said. “AI is a tool, potentially an incredibly powerful one, but it is up to us – our societies, governments, and companies – how we use it.”

In October, ex-Google CEO Eric Schmidt claimed industries should scrap their climate goals as AI will outpace any measures put in place, and likely be able to solve the problem once it matures.

Speaking at a Washington AI summit, Schmidt said that companies will not meet their climate goals “because we’re not organized to do it.”

“Yes, the needs in this area will be a problem, but I’d rather bet on AI solving the problem than constraining it and having the problem,” Schmidt said.