Big Tech's AI ambitions are remaking the US power grid. Consumers are paying the price.

The AI boom is driving unprecedented electricity demand, with Goldman Sachs projecting 50 gigawatts of new U.S. capacity needed, presenting both significant opportunities and risks for utilities and the broader economy. While utilities like Duke and Southern Co. report massive data center load pipelines and equipment manufacturers such as GE Vernova see surging orders, concerns are rising over potential over-forecasting and stranded assets, highlighted by Microsoft canceling projects and regulatory challenges to Amazon's deals. Utilities are implementing new tariffs, exemplified by AEP Ohio's requirement for data centers to pay for committed capacity, to mitigate risk, a strategy that has already led to reduced demand forecasts, alongside federal support for grid modernization.

Analysis

The AI boom is driving unprecedented electricity demand, with Goldman Sachs projecting 50 gigawatts of new power capacity required in the US, a significant increase after a decade of flat electricity needs. This surge is creating substantial opportunities for utilities, as evidenced by Duke Energy (DUK) and Southern Company (SO) reporting massive data center load pipelines and equipment manufacturers like GE Vernova (GEV) experiencing surging orders, with GEV's electrification division seeing nearly $500 million in H1 2025 orders compared to $600 million for all of 2024. This demand has contributed to GEV's stock gaining over 80% this year.

However, the rapid expansion presents considerable risks, drawing parallels to the early 2000s natural gas boom where overcapacity led to bankruptcies. Microsoft (MSFT) recently canceled 2-gigawatt data center projects, and regulatory bodies, such as Monitoring Analytics, are challenging deals like Amazon's (AMZN) transmission agreement with PECO Energy over reliability and cost concerns for ratepayers. This indicates a potential for over-forecasting and stranded assets if demand does not materialize as expected.

In response, utilities are implementing proactive risk-mitigation strategies; AEP Ohio, for example, secured approval for a tariff requiring data centers to pay for at least 85% of committed energy and exit fees, which subsequently reduced its data center load demand forecast by over half. This approach shifts financial risk to developers, ensuring greater commitment and protecting utilities from potential stranded assets. Federal support, including a $1.6 billion loan to AEP for transmission upgrades, also aims to bolster grid infrastructure.

The increased energy load is already impacting US electricity costs, with Bank of America (BAC) projecting further upside for consumer bills due to grid enhancement spending. The scarcity and expense of new electrical equipment, as highlighted by Grid Strategies, could lead to utilities regretting premiums paid if the long-term demand picture becomes murky, potentially passing these costs to ratepayers or absorbing losses.

Key Decisions for Investors

• Investors should closely monitor utility regulatory filings for new tariffs and risk-sharing mechanisms, as these directly impact revenue stability and asset utilization for power providers.
• Evaluate equipment manufacturers like GE Vernova (GEV) for sustained order growth, but be mindful of potential demand volatility and pricing pressures if AI data center buildouts slow or projects are canceled.
• Assess utilities with significant data center exposure, differentiating those with robust risk-mitigation strategies (e.g., AEP's tariffs) from those potentially exposed to stranded asset risk.
• Consider the broader macroeconomic impact of rising electricity costs on consumer spending and industrial operations, as this could create headwinds for other sectors.