Interested in more of our content? Sign up for our newsletter. Sign up here:
The digital age of AI and cloud computing and its rapidly growing demand for data centers and computing power is intersecting with industrial and consumer electrification trends including electric vehicles. The North American Electric Reliability Corporation (NERC) warns that this surge in consumption, coupled with the retirement of aging power plants, could result in power supply shortfalls across half of the United States within the next decade.
Rapidly growing electricity demand has caught many electric companies, municipalities and state regulators in the US by surprise and has led to a global hunt for locations with affordable and reliable access to electricity for data centers.
“Goldman Sachs Research estimates that data center power demand will grow 160% by 2030. Data centers will use between 8% - 12% of US power by 2030, compared with 3% in 2022. That kind of spike in power demand hasn’t been seen in the United States since the early years of this century and represents a seismic shift in composition of energy utilization in the United States. In Europe, power demand could grow up to 50% by 2033 and Europe may need to invest as much as $1 trillion-plus to prepare its power grid for AI driven consumption.
Utilities have struggled in key markets to keep up with new data center demand. “AI turned what was a backlog with annoyingly long wait times into an existential crisis. That’s because AI runs on high power GPUs, which, while technically more efficient than older chips also require more power. So, in addition to requiring more intensive cooling, the proliferation of these chips means they’ll also require more energy to be piped to data centers.” In addition, the average ChatGPT query takes 10 times more energy than a Google search query.
As the power ecosystem grapples with meeting data centers’ voracious need for power, it faces substantial constraints, including limitations on reliable power sources, sustainability of power, upstream infrastructure for power access, power equipment within data centers, and electrical trade workers to build out facilities and infrastructure. Currently, for example, the lead time to power new data centers in large markets such as Northern Virginia can be more than three years. And, in some cases, transformers and other electrical equipment are in short supply and may have lead times of two years or more as transformers and other parts.
Electric supply may be further strained by the shortage of skilled electricians and transformers and other parts, many being used in Ukraine and other war torn areas. McKinsey estimates anticipate a potential shortage of up to 400,000 trade workers in the United States based on projected data center build-out and comparable assets requiring similar skills, such as semiconductor fabrication and battery gigafactories.
Potential Risks and Implications
Data center and cloud computing costs could increase in the near term as the market adjusts and businesses should be realistic in their ability to manage or reduce these costs.
New technology infrastructure could be limited or companies may experience long lead times. Build time and cost contingencies for construction projects due to parts and labor shortages and potential inflationary impacts.
The power grid might become more fragile before the market catches up and stabilizes - ensure business contingency plans address potential of extended power outages.
Increased M&A activity and consolidation in the power market could affect current suppliers.
Commercial real estate suitable for data centers and similar projects is becoming scarce. Companies are looking globally for acceptable data center locations as intense competition by Google, Amazon, MSFT, OpenAI, Oracle, Meta and others will continue to drive scarcity.
Expect to see a resurgence in nuclear power as leading technology companies increasingly turn to nuclear power for data centers, companies may need to review and clarify any ESG clean energy related commitments regarding the use of nuclear power.
Request more information on emerging risks, DelCreo’s Risk Universe and enterprise risk services.
As a reminder, here are our Risk Universe categories that we leverage to tackle and understand risk which include:
External Risk
Governance Risk
Strategic Risk
Product Risk
Business Operations Risk
Legal & Compliance Risk
Financial Risk
Technology Risk
We leverage our understanding of risk maps and risk universes to better advise our clients in strategic business decisions and to optimize the management of risk throughout the enterprise.
More Details
AI Needs So Much Power, It’s Making Yours Worse
AI data centers in the US are consuming vast amounts of electricity, leading to distorted power flows that threaten billions in damages to home appliances and aging power equipment, particularly near major cities and rural areas with significant data center activity.
The phenomenon of "bad harmonics," caused by deviations in electrical wave patterns, is linked to data centers and can lead to overheating electronics, rattling motors, and increased risk of electrical fires, with more than three-quarters of distorted power readings occurring within 50 miles of data centers.
A Bloomberg analysis of Whisker Labs data shows a strong correlation between proximity to data centers and worsening power quality, with urban areas like Northern Virginia's "data center alley" showing higher levels of distortion, though rural areas are also affected.
As the demand for electricity is projected to surge by 16% over the next five years due to data center expansion, significant improvements to the power grid are necessary to prevent worsening harmonics issues, with solutions like dedicated substations and devices like filters and capacitors being considered.
Could data center expansions spark ‘power wars’?
Neil Sahota, UN AI advisor, warns that the increasing demand for electricity by data centers could lead to conflicts over power resources, emphasizing that energy consumption is outpacing new power generation capabilities.
Sahota suggests redesigning computing systems from the ground up to reduce power usage and heat generation, with AI potentially playing a role in this transformation; however, this requires significant investment and faces challenges due to existing infrastructure costs.
While some, like Rodolfo Rosini, advocate for new computing paradigms such as reversible computing, others like Wesley Cummins see nuclear power and tapping into stranded energy sources as more immediate solutions.
The debate continues over prioritizing sustainability versus maintaining technological leadership, with concerns about convincing governments to shift focus towards sustainable energy solutions amidst the push for AI infrastructure dominance.
Four key trends disrupting data centers in 2025
The data center industry is experiencing significant growth in electricity demand, driven by compute-intensive workloads and the expansion of AI data centers, with electricity consumption expected to exceed 1,000 TWh by 2026.
Key trends disrupting data centers include the adoption of liquid cooling technologies to manage higher power densities and thermal demands, as well as the rise of dedicated AI data centers optimized for efficiency and scalability.
The industry is focusing on sustainability and efficiency, with companies and governments promoting innovative technologies and practices to reduce energy consumption and greenhouse gas emissions.
AI plays a crucial role in optimizing data center operations and reducing environmental impact, with companies like Google leading efforts to enhance energy efficiency and achieve net-zero emissions.
Biden’s executive order on AI data centers could benefit Amazon and Microsoft
President Biden's executive order aims to boost the construction of domestic data centers and clean energy plants to maintain U.S. leadership in AI, benefiting companies like Amazon and Microsoft.
The order directs federal agencies to lease sites for AI infrastructure and emphasizes the importance of clean energy sources such as next-gen nuclear and geothermal.
Challenges include shortages of suitable land and clean energy, with permitting delays and electricity access impacting data center projects' timelines.
Despite these challenges, tech giants continue to invest heavily in AI and data centers, with significant capital expenditures reported by companies like Amazon and Microsoft.
U.S. electricity demand is set to explode. That will make it harder to cut climate pollution
U.S. electricity demand is projected to increase by nearly 16% over the next five years, driven by new data centers, manufacturing growth, and the shift to electric vehicles and appliances, leading to higher natural gas consumption and challenges in reducing emissions.
Despite significant investments in clean energy under the Inflation Reduction Act, U.S. greenhouse gas emissions have not significantly decreased, making it unlikely to meet the goal of cutting emissions in half by the end of the decade.
The reliance on natural gas is expected to grow, as it is currently viewed as a reliable energy source, but this poses risks due to methane emissions and the need for faster adoption of renewable energy and new technologies like nuclear power.
The continued use of fossil fuels to meet rising electricity demand could hinder global efforts to combat climate change, highlighting the urgency for streamlined regulations and infrastructure to support clean energy projects.
Energy is at the top of the data center agenda heading into 2025
In 2024, data center operators faced an existential crisis due to increased power demands from AI facilities, leading to a scramble for reliable energy sources.
Nuclear power emerged as a leading solution due to its reliability and low carbon emissions, with major companies like Amazon, Microsoft, Oracle, and Google securing nuclear power deals.
In the short term, data centers are also focusing on renewable energy sources like solar and wind to meet immediate power needs and carbon reduction goals.
Looking forward, the industry anticipates innovations in renewable energy sourcing and integration, along with the potential use of technologies like small modular reactors and enhanced geothermal options.
AI is poised to drive 160% increase in data center power demand
AI is expected to drive a 160% increase in data center power demand by 2030, with data centers potentially consuming 3-4% of global power, up from 1-2% today, leading to a significant rise in carbon dioxide emissions.
A ChatGPT query requires significantly more electricity than a Google search, contributing to the increased power consumption driven by AI technologies, which could add 200 terawatt-hours per year to data center power usage from 2023 to 2030.
The US is anticipated to see a surge in electricity demand, with data centers accounting for 8% of power consumption by 2030, necessitating $50 billion in new generation capacity and increased natural gas demand.
Europe will need over €1 trillion in investments to prepare its power grid for AI, with a potential 40-50% increase in power demand by 2033, driven by data center expansion and electrification, particularly in countries with abundant renewable energy or incentives for tech companies.
AI investment forecast to approach $200 billion globally by 2025
AI investment is projected to reach approximately $200 billion globally by 2025, with significant upfront investments needed in physical, digital, and human capital to drive large-scale transformation and productivity gains.
The U.S. is positioned as a market leader in AI technology, with early adoption expected from larger firms, particularly in information, professional, scientific, and technical services, potentially peaking at 2.5 to 4% of GDP.
Despite rapid market interest and increased mentions in company earnings calls, the near-term GDP impact of AI investment is expected to be modest, as AI-related investment currently represents a small share of U.S. and global GDP.
AI investment is concentrated in training and developing AI models, infrastructure, AI-enabled software, and enterprise end-users, with a larger hardware and software push required for generative AI to scale, suggesting broader macroeconomic effects will emerge in the latter half of the decade.
Nuclear Energy Stocks Enter 2025 On A High. AI Data Centers Need Them, A Lot
Nuclear energy stocks, particularly Constellation Energy and Vistra, experienced significant growth in 2024 due to increased electricity demand from AI and data centers, despite facing regulatory challenges and uncertainties from the Trump administration.
Analysts remain optimistic about nuclear energy stocks in 2025, citing continued high electricity demand driven by AI applications and data centers, which are expected to grow significantly in the coming years.
Small modular reactors (SMRs) are gaining traction as a promising solution for providing power to tech companies, with major investments from companies like Amazon, Alphabet, and Oracle, although they face regulatory and operational hurdles.
The nuclear production tax credit in the Inflation Reduction Act is seen as a stabilizing factor for the industry, and despite potential policy changes under the Trump administration, bipartisan support for nuclear energy remains strong.
Constellation Energy Agrees to Buy Calpine for $16.4 Billion
Constellation Energy has agreed to acquire Calpine for $16.4 billion, valuing the deal at $26.6 billion with debt, to combine two major U.S. power generators amid increasing electricity demand driven by tech companies and AI computing needs.
The acquisition expands Constellation's presence in Texas and California, key markets with growing electricity demand, and strengthens its position as the largest nuclear power producer in the U.S., while Calpine contributes significant natural gas and geothermal generation capacity.
Constellation's shares surged following the announcement, reflecting the high demand for power producers as tech-driven energy consumption rises; its market value has exceeded $95 billion, with a 163% increase in stock price over the past year.
The merger aligns with Constellation's strategy to provide reliable and low-carbon electricity, combining nuclear and natural gas resources, and is supported by recent long-term power agreements with entities like Microsoft and the General Services Administration.
Access our LinkedIn