Data centers are consuming power faster than the electrical grid can keep up. This strain on power infrastructure stems from digital infrastructure and artificial intelligence (AI) expanding more rapidly than they can be connected to the power grid. Technology companies are driving this AI boom through major capital investments, but even their expansive agreements with power providers may not be sufficient to meet the massive and increasing power needs. This surge in demand will force significant changes in how both the power and technology industries operate and evolve.

The rising demand shows no signs of slowing. The International Energy Agency (IEA) projects that global electricity demand from data centers will more than double by 2030, reaching approximately 945 terawatt-hours (TWh) – roughly equivalent to Japan’s current total electricity consumption. In the U.S., data centers are expected to consume 580 TWh annually by 2028, accounting for 12 percent of the nation’s total electricity use. AI will drive most of this increase, with electricity demand from AI-optimized data centers expected to more than quadruple by 2030. In the United States alone, data centers will account for nearly half of all electricity demand growth between now and 2030. By that time, powering data processing will consume more electricity than manufacturing all energy-intensive goods combined, including aluminum, steel, cement, and chemicals.

Supply is struggling to keep pace: nearly 2 TW of clean energy - 1.6 times current grid capacity – is delayed in interconnection queues. The urgency of this supply crunch is underscored by the global race for AI leadership, with technology companies investing billions into AI development and infrastructure. Yet this creates a striking paradox: the same companies making these ambitious AI investments have also committed to aggressive carbon-free goals, with net-zero targets as soon as 2030 to 2035.

A modern data center featuring sleek server racks illuminated by colorful LED lights in a futuristic, tech-inspired environment.

How are power and technology industry players responding?

Both the power and technology industries are advancing two waves of activity to meet this challenge. Wave 1 utilizes energy bridging solutions and leverages all options that are available. Wave 2 will be preparing for a quantum leap in clean energy generation capacity to be ready by 2030.

Wave 1: Building bridges between existing solutions

Utilities and data center companies are creatively leveraging options today to enable faster, more flexible grid connections. To fill immediate needs, technology and power companies are creatively tapping existing capacity resources that may be underutilized or undercompensated in existing markets. For example, Microsoft has agreed to purchase power from the Three Mile Island nuclear plant in Pennsylvania, while Amazon has struck deals with the Susquehanna plant in Pennsylvania and the Clinton plant in Illinois, demonstrating this approach of leveraging existing nuclear capacity. In various markets, the spare capacity of gas plants is also being tapped.

At the same time, many utilities are building new gas-fired generation capacity based on data center demand expectations. For example, to support its growing energy needs, Meta will source power from a new 200-megawatt (MW) gas-fired plant in the village of New Albany, Ohio, being developed by Advanced Power and expected to begin commercial operation in 2026. Relatedly, some data center developers are securing guaranteed gas supply connections to support on-site power generation; for example, CloudBurst has signed a long-term agreement with Energy Transfer to receive up to 450,000 Metric Million British Thermal Units (MMBtu) per day of natural gas via the Oasis Pipeline to its campus outside San Marcos, Texas—enough to generate nearly 1.2 gigawatt (GW) of power. These projects reinforce the point that existing dispatchable energy sources, such as nuclear and natural gas, remain a critical enabling technology for data center reliability and growth.

Utilities are also deploying new technologies and innovative strategies to improve the efficiency of existing facilities and more quickly engineer, site, permit, and construct new substations. For example, National Grid UK recently announced a $100 million commitment to invest in AI startups such as Amperon and Sensat, aimed at accelerating AI‑driven solutions for grid efficiency, resilience, and the quicker connection of data centers to the electricity system. In addition, utilities are exploring innovative contracts to compensate non-data center customers, such as other businesses and homeowners, who are willing to provide demand flexibility for a few hours (i.e., reducing their power consumption during peak demand hours), as a more flexible demand profile allows more data center demand onto the grid without significant upgrades.

When data centers are not connected to the power grid quickly enough, developers often look to behind-the-meter (BtM) solutions like solar, wind, fuel cells, and batteries. In some areas, utilities are requesting that BtM generation become a permanent grid asset. The 1-GW fuel cell deal between AEP and Bloom Energy to power data centers would set a new record, marking the largest utility-scale fuel cell procurement in U.S. history and enabling faster deployment compared to traditional grid upgrades. Similarly, PowerSecure is powering Edged’s new 168-MW data center campus in Atlanta with Tier 4 Final microgrid generators and on-site electric distribution, part of a broader rollout across multiple Edged sites.

These bridging strategies can help decrease lead times for getting new power plants and physical grid infrastructure in place, addressing what is currently the biggest bottleneck to powering AI with grid power — transmission interconnections — which, in many parts of the world, face a 2-to-3-year backlog.

Wave 2: Ensuring clean power for 2030 and beyond

Simultaneous to the Wave 1 bridging strategies, the industry is preparing for a quantum leap to ensure clean power for 2030 and beyond, while also exploring step change reductions in power consumption. On the supply side, efforts are focused on advancing nuclear energy, including the deployment of small modular reactors (SMRs), geothermal energy, and even fusion technology. For example, Fervo Energy recently raised $255 million to deploy carbon-free geothermal power, and Google signed an agreement with Commonwealth Fusion Systems for 200MW of power from a proposed nuclear fusion plant in Virginia by the early 2030s.

To expand the capacity of the grid and bring clean generation online more quickly, the power sector is deploying advanced technologies to modernize the grid. For example, National Grid is partnering with VEIR to increase the capacity of the grid through the use of high-temperature superconductors.

On the demand side, quantum leap innovation is also underway, with the industry working to dramatically reduce chip cooling loads. The ARPA-E “COOLERCHIPS” program supports more than 15 innovative startups focused on reducing cooling demands in data centers, which is a particularly energy-intensive aspect of the industry.

Abstract visualization of flowing blue light and particles, creating a dynamic wave-like pattern against a dark background.

Policy reform, commercialization, and collaboration – core pillars for enabling solutions

Policy reform and regulatory modernization are essential to unlock investment. There is a massive need for better alignment between government agencies and utility regulators, especially as data center developers face high upfront fees without clear connection timelines. These costs are prompting some to pursue behind-the-meter (BtM) solutions or relocate to more favorable regions (e.g., those with more favorable energy policies and tax incentives). If not carefully managed, such fees could deter important grid customers and limit grid integration, underscoring the need for thoughtful tariff and fee design by regulators.

Commercialization support is also crucial. As Wave 2 technologies move through the prototype and demonstration phases, the greatest coordination and policy challenges lie in deploying them at scale. These tech solutions have the potential to provide clean power for data centers, but all of them will require cross-industry collaboration to bring them through the R&D, demonstration, and deployment phases.

Conclusion

To meet the soaring power demands of AI and data centers, the power and technology sectors must continue forging bold, collaborative strategies across both short-term and long-term horizons. As we look ahead, several key priorities will be critical to ensuring a sustainable and scalable digital future:

1. Bridging strategies are critical but temporary: Short-term solutions like spare capacity, on-site gas, and behind-the-meter generation are helping to relieve immediate grid pressure—but must be deployed faster and more strategically as part of a broader transition plan.

2. Build a path to scale for long-term clean power and efficiency technologies: Geothermal, nuclear (including SMRs), and chip cooling innovations offer long-term solutions to AI’s rising energy demand. Scaling them requires coordinated investment, R&D support, and demonstration at pace.

3. Modernize regulatory frameworks to unlock investment: Reform permitting, streamline interconnection, and align agency oversight to accelerate grid upgrades. Fee and tariff design must encourage, not discourage, data center integration.

4. Forge cross-sector collaboration as a foundation: Gas, electric, and tech sectors must jointly plan infrastructure and coordinate demand. Proactive collaboration between utilities, tech companies, and regulators is now essential to meet the scale and urgency of AI-era energy needs.


ERM brings strategic and technical advisory, as well as procurement support, across the capital asset lifecycle, to help our clients access secure and sustainable energy when and where it is needed, scale energy-efficient solutions to reduce costs, and retain a license to operate through compliance and integration with local energy systems.

Please contact us for more information:

Mackay Miller, Partner
Sam Foster, Consulting Partner
Charlie Knaggs, Partner
Takayuki Shibata - Partner