Last summer, Wells Fargo projected that the demand for power to support artificial intelligence (AI) will surge by 8,050% by 2030, increasing from 8 terawatt-hours (TWh) in 2024 to 652 TWh.
Most of this demand is expected to come from AI training, which is projected to reach 40 TWh by 2026 and 402 TWh by 2030. The power demand for AI inference is also anticipated to accelerate toward the end of the decade.
In August 2024, Amanda Peterson Corio, the global head of data center energy at Google, was interviewed on an episode of the U.S. Department of Energy‘s Grid Talk.
During the interview, she discussed how Google’s data centers consume an enormous amount of electricity, which led the tech giant to participate in the accelerating energy revolution.
In the first week of August 2025, Indiana Michigan Power, part of American Electric Power, announced it had filed a special joint contract with Google to support the utility’s capacity needs through clean energy generation.
This special contract also aims to provide reliable and affordable service for all customers.
The press release indicated that by participating in this program, Google would leverage new capabilities to reduce or shift electricity demand, performing non-urgent tasks during hours when the electric grid is under less stress.
This initiative is designed to minimize peak load during periods of high energy demand, similar to programs currently available for residential and commercial/industrial customers.
As a large load customer, Google’s involvement boosts Indiana Michigan Power’s ability to manage electricity demand during peak times, helping to lower overall energy costs.
Additionally, the Tennessee Valley Authority announced a similar partnership, expanding on a 2019 agreement to provide carbon-free renewable power for Google’s data centers in Tennessee and Alabama.
Tennessee Valley Authority is the third-largest electricity generator in the country, with over 16,000 miles of transmission lines. They provide power for the 10 million people they serve across a seven-state region.
Earlier this year, the American Public Power Association reported that Goldman Sachs Research forecasts global power demand from data centers will increase by 50% by 2027 and could rise by as much as 165% by the end of the decade compared to 2023.
Goldman Sachs also noted that large hyperscale cloud providers and other corporations building extensive language models capable of natural language processing require vast amounts of information and power-intensive processors for training.
Currently, it is estimated that the global data center market consumes around 55 gigawatts of power, with cloud computing workloads accounting for 54%, traditional business functions (such as email and storage) at 32%, and AI workloads making up 14%.
As data centers contribute to the growing demand for power, the electric grid will require approximately $720 billion in spending through 2030.
In June 2025, AutoGPT.net, a pioneering website providing breaking news and expert insights on artificial intelligence, published a list of the top 20 AI energy companies that are transforming the industry.
According to AutoGPT, these energy companies that have integrated AI into their operations are reshaping the energy sector by lowering carbon emissions and improving the reliability of clean energy sources like wind and solar.
The energy sector is now better able to predict energy demand, making renewable energy more accessible by balancing supply and demand, and enhancing grid reliability to reduce power outages. The AutoGPT blog describes AI as the “brain” of the energy industry.
The top AI energy companies include Siemens Energy General Electric (GE), Shell AI, BP AI, ExxonMobil AI solutions, Tesla Energy, Iberdrola, Schneider Electric, Enel, Repsol, Duke Energy, E.ON, NextEra Energy, National Grid, Engie, Envision Energy, Xcel Energy, Dominion Energy, Ørsted, and the EDF Group.
Seeking Alpha, a financial services company, has noted that increasing electricity demand is driving utilities to invest heavily in grid modernization initiatives, which involve substantial capital costs partly funded by significant rate hikes.
These infrastructure investments are essential for replacing aging systems and meeting the skyrocketing demand from data centers, highlighting the partnership between major tech companies and energy providers.
Furthermore, Nvidia and AMD have seen their semiconductor technologies increasingly adopted in nuclear applications, energy management, and radiation detection systems, demonstrating the deep relationship between AI and the utilities sector.
Notable AI-utilities deals include a landmark agreement between Meta and Constellation Energy, signed in early June, which guarantees nuclear energy for Meta’s data centers. Amazon has also secured a contract for 1.9 GW from Talen Energy’s Susquehanna atomic station to support AWS facilities in Pennsylvania.
Matt Sage recently shared on social media that reducing power usage in AI data centers during times of grid strain is a crucial wake-up call for the energy grid.
He emphasized that Google’s decision is significant because data centers are rapidly increasing in size and power demand. The industry is currently valued at $242 billion and is projected to reach $584 billion by 2032.
Utilities are struggling to keep pace with this growth and are unable to pass on billions in additional costs to consumers. For example, I&M’s $1.3 billion cost recovery plan was rejected in court.
Google’s “back-down” system temporarily halts non-urgent computing tasks, such as processing a YouTube upload, and sometimes redirects them to other data centers.
According to Sage, this approach is more than just minor operational adjustments—it’s a serious warning. If the growth of AI and cloud computing remains unchecked, it could create bottlenecks in the grid.
Curtailment of power usage can lead to lost productivity and increased long-term costs.
Sage said that the clear solution is to expand energy storage significantly. Energy storage serves as a form of insurance for uptime; with sufficient battery capacity, data centers could operate seamlessly during peak demand times without needing to throttle services.
This is why energy storage is not only a vital component of clean energy initiatives but also an essential aspect of investing in AI and digital infrastructure.
Wells Fargo is projecting AI power demand to surge 8,050% by 2030, from 8 TWh this year to 652 TWh.
AI training is expected to drive the bulk of this demand, at 40 TWh in 2026 and 402 TWh by 2030, with inference’s power demand accelerating at the end of the decade.$NVDA $AMD… pic.twitter.com/3b4eVN8LVq
— Beth Kindig (@Beth_Kindig) July 8, 2024
Indiana Michigan Power and Google have filed a contract to participate in a Demand Response program aimed at managing peak energy demand effectively.
Learn more about this impactful collaboration here: https://t.co/QBreSKu5N3 pic.twitter.com/EFCcOEJdZt
— I&M (@IN_MI_Power) August 7, 2025
As the Southeast grows and summer temps climb, reliable power matters now more than ever!☀️⚡
Our diverse energy portfolio, which features nuclear, hydro, natural gas, coal & renewables, is helping us meet a nearly 3% rise in demand since 2019.
We're building tomorrow together… pic.twitter.com/YeWwfsLzyP— Tennessee Valley Authority (@TVAnews) July 17, 2025
Google forced to cut AI data centre power use — a wake-up call for the grid.$GOOGL just signed deals with two U.S. utilities (Indiana Michigan Power & Tennessee Valley Authority) agreeing to reduce its AI data centre electricity demand whenever the grid is under strain.
Here’s…
— Matt Sage (@MattSageInc) August 8, 2025
Building on our effort to procure renewable energy worldwide, here's why we’re putting 1.6 million solar panels in Tennessee and Alabama – our biggest solar project yet → https://t.co/dyes53nb5d pic.twitter.com/hlK7uql3RJ
— Google (@Google) January 16, 2019
The U.S. Army Corps of Engineers has been tasked with…
Brown and Caldwell, a leading environmental engineering and construction firm,…
Humboldt State University, one of four campuses within the California…
