Learners Live

Scaling Intelligence: The Exponential Growth of AI’s Power Needs

The rapid advancement of artificial intelligence (AI)—particularly the training of large-scale “frontier models”—is driving renewed growth in electricity demand. This report analyzes the technical drivers of AI power consumption, projects future demand trajectories for individual training sites and broader AI needs, and highlights energy sector implications. Their analysis found not only that the power demands of AI have increased steadily, but also that they will keep increasing. While training large, advanced AI models currently requires between 100 and 150 megawatts each, they are projected to require more than four gigawatts apiece by 2030.This Product is publicly available at Electric Power Research Institute (EPRI): Scaling Intelligence: The Exponential Growth of AI’s Power Needs

Westinghouse Plans to Build 10 Large Nuclear Reactors in U.S., Interim CEO Tells Trump

Westinghouse plans to build 10 large nuclear reactors in the U.S. with construction to begin by 2030.  Westinghouse’s big AP1000 reactor generates enough electricity to power more than 750,000 homes, according to the company. Building 10 of these reactors would drive $75 billion of economic value across the U.S. and $6 billion in Pennsylvania. Trump issued four executive orders in May that aim to quadruple nuclear power in the U.S. by 2050. The U.S. has built only two new nuclear reactors over the past 30 years, both of which were Westinghouse AP1000s at Plant Vogtle in Waynesboro, Georgia.  Westinghouse emerged from bankruptcy in 2018 and is now owned by Canadian uranium miner Cameco and Brookfield Asset Management. Westinghouse plans to build 10 large nuclear reactors in U.S., interim CEO tells Trump

U.S. and Israel Pledge to Work Together to Unleash AI Innovation

The Memorandum of Understanding (MOU)  highlights the transformative potential of AI to improve the security and resilience of America and Israel’s energy systems. The two countries further announced their intent to pursue cooperation in areas including analyzing the impact of rising energy demand due to data centers, opportunities for grid optimization, enhanced AI-enabled cybersecurity tools, sharing best practices on the use of AI in energy infrastructure, and the launching of bilateral pilot projects. U.S. and Israel Pledge to Work Together to Unleash AI Innovation – electrifiED

Data Center Construction Continues to Grow by Tom Zind

How do you paint a rosy picture for a construction market sector that’s been in the dumps and shows scant signs of a quick change of fortune? There’s one way: Stretch the definition, and throw in a category that’s come so far so fast that standard market indices haven’t been able to catch up. That’s what FMI Corp., does in the second-quarter installment of its 2025 North American Engineering and Construction Industry Outlook Reasonably, it lumps data centers into the office component of its non-residential construction analysis, yielding a bullish four-year forecast for the sector that beats nine of 10 others. Citing forecasts of data center demand growing 34% in 2025 and 18% next year, FMI sees office construction growing every year between 2026 and 2029 at an average of about 7% annually. Data centers may be the office of the future, housing the infrastructure enabling higher productivity that could be a driving force for the economy for decades to come. Data Center Construction Continues to Grow | EC&M

Trump’s Permitting Boss Aims to Deliver on AI Data Center Plans by Stephen Lee

The White House’s new permitting director wants to vastly expand her office’s work in getting projects built—including, for the first time, data centers and artificial intelligence facilities. That would be a boon to President Donald Trump, whose economic policy advocates for the construction of projects like OpenAI Inc.‘s $500 billion Stargate AI venture, Taiwan Semiconductor Manufacturing Co.‘s $100 billion investment in new chip factories, and as many new hardrock minesas possible. Many of those projects will have to secure federal permits under the National Environmental Policy Act before they can break ground, a daunting prospect that can take years. Enter the Federal Permitting Improvement Steering Council, which shepherds complex projects through the process by harmonizing different agencies’ environmental reviews, anticipating hurdles before they create delays, and holding agencies accountable to deadlines. On average, the 25-person council says it’s able to shave 18 months off a project’s review time. Trump’s Permitting Boss Aims to Deliver on AI Data Center Plans

DOE Announces Accelerated Siting of Data Centers

Oak Ridge National Laboratory (ORNL) and Elementl Power have teamed up to advance a data-driven siting approach for advanced nuclear projects. The work started through a U.S. Department of Energy (DOE) GAIN voucher and contributed to Elementl’s recent agreement with Google to prepare three U.S. sites for advanced nuclear deployment. Each site would have at least 600 megawatts of capacity and would likely be used to power large-scale data centers. Nuclear energy can provide around-the-clock abundant and reliable electricity, making it an attractive solution to meet rising energy demand from AI and data centers. Google will provide early-stage capital to develop three sites for advanced reactor projects with final technology selection and site confirmation pending further development milestones. DOE Announces Accelerated Siting of Data Centers – electrifiED

N.J. Explores Nuclear Strategy for Energy Goals

 New Jersey is projected to need as much as 10 gigawatts of clean, firm energy capacity by 2035 to maintain system reliability and meet the state’s clean electricity goals. The New Jersey Board of Public Utilities has initiated a request for information to evaluate the possible role new nuclear power plants could have in meeting the state’s energy objectives. The RFI seeks insights into how nuclear can support growing energy demands, particularly from large consumers like AI data centers.  New Jersey BPU eyes potential for new nuclear | Utility Dive

Can Solar Help Offset the AI’s Energy Demand? by Greg Zimmerman

Demand from data centers is expected to more than triple by 2028, according to the Department of Energy. This is largely due to the explosion of Artificial Intelligence, and its intense demand for electricity.  Meta, Amazon, Google, and Microsoft are just a few of the big tech companies investing billions in solar technology. Two of the main reasons for the investment in solar is that the solar development timeline is much shorter than for other sources electricity. At 1.4 years, solar is the shortest timeline to develop an 800-megawatt solar plant. That’s compared to 15 years for a nuclear power and 6.7 years for a coal-fired power plant.  Also, solar is much less expensive, is easily scalable, and pays back much quicker than other sources of electricity.   Can Solar Help Offset the AI’s Energy Demand?  – Facility Management Green Quick Read

Coal Plant to be Transformed into Gas-Powered Data Center Campus

 Homer City Redevelopment (HCR) and Kiewit Power Constructors Co. (Kiewit) announced the future of the former Homer City Generating Station. Homer City – previously the largest coal-burning power plant in Pennsylvania – will be transformed into a more than 3,200-acre natural gas-powered data center campus, designed to meet the growing artificial intelligence (AI) and high-performance computing (HPC) needs of the innovative technology companies shaping America’s digital future. GE Vernovawill provide seven high-efficiency 7HA.02 hydrogen-enabled, gas-fired turbines, with the first deliveries expected to begin in 2026.  Coal Plant to be Transformed into Gas-Powered Data Center Campus – electrifiED

NEMA Releases Grid Reliability Study

The National Electrical Manufacturers Association (NEMA) has released its new study on grid reliability, “A Reliable Grid for an Electric Future”, which focuses on challenges we face with electric demand, consumption and production over the next 25 years. The study highlights U.S. electricity demand outlook to 2050 and outlines technology and policy solutions to meet demand growth. Two key drivers behind the demand will be data centers and EV charging. To prioritize grid reliability and ensure our electrical infrastructure can adapt to surging demand and the integration of renewable energy, we can’t build capacity fast enough. We need to invest in a suite of innovative near-term technologies to meet demand growth and prioritize policy and regulatory certainty to deliver a reliable and affordable energy system. makeitelectric.org Study at: pa-nema-grid-flexibility-study.pdf

Why Use AI – 5 Reasons besides making us smarter:

  1. Efficiency and Automation: AI can automate repetitive tasks… saves time and reduces human effort
  2. Data Processing: AI excels at analyzing large datasets quickly… uncovers patterns and insights that humans might miss
  3. Scalability: AI systems can handle growing workloads… without a proportional increase in costs or resources
  4. Accuracy: AI can perform tasks with high precision… reducing errors in areas like diagnostics or forecasting
  5. 24/7 Availability: AI tools can operate continuously… improving productivity and customer support