The rapid growth of artificial intelligence (AI) is putting immense pressure on our energy infrastructure. It’s like trying to renovate an old house while still living in it—we must keep the power flowing today while overhauling the system for tomorrow.

As AI-driven data centers multiply, their insatiable demand for electricity is straining an already overburdened grid system. This presents a complex challenge: how do we simultaneously maintain economic competitiveness, ensure grid reliability, and meet clean energy goals—all while keeping our current system operational? We can’t simply move out while we renovate. Solutions must work in the short term while laying the foundation for long-term transformation.

The energy dilemma and business as usual

Data centers have become essential infrastructure, as critical to modern economies as highways and ports. Their role in economic growth and national security makes meeting their energy needs a top priority. Meanwhile, many tech companies have committed to sourcing clean energy and scaling emissions-free power solutions.

However, our aging energy infrastructure wasn’t built for this surge in demand. Our regulatory framework, energy markets, and permitting processes operate at a sluggish pace, ill-suited to the rapid expansion and innovation required to support the AI revolution.

The default response has been to squeeze more demand onto the existing grid. Tech companies search for pockets of available capacity in primary and secondary markets, while utilities rush to expand wind, solar, natural gas generation, and battery storage. But this approach is hitting a breaking point. In many regions, it’s a major factor driving up electricity prices, threatening grid reliability, and increasing emissions as older, more polluting power plants are forced to run more often to keep up with demand.

Exploring solutions

As the challenge gains attention, a range of solutions has emerged—from optimizing existing infrastructure to expanding grid capacities and innovating new energy technologies.

Near-term strategies that focus on bypassing grid bottlenecks:

While these options may offer speedier adoption, they come with limitations, including uncertain scalability and reliance on natural gas in some cases. This dependence puts the U.S. at a disadvantage compared to countries with abundant clean energy—China, for instance, has added 34 gigawatts of clean firm nuclear in the past decade, more than ten times the U.S. nuclear additions over that period.

Streamlining permitting and deployment processes—particularly for transmission infrastructure—could unlock hundreds of gigawatts of power capacity to support AI needs. However, comprehensive reform is politically complex and slow-moving, meaning its full impact will take years to materialize. In the meantime, targeted “band-aid” solutions could help alleviate specific bottlenecks while broader reforms take shape.

Innovative energy technologies like advanced nuclear reactors, next-generation geothermal systems, and carbon capture and storage could provide round-the-clock, carbon-free power with intentional innovation and commercialization support. But even under the most favorable conditions, these solutions won’t be available at scale in the U.S. until well into the 2030s—like comprehensive permitting reform, new technologies are essential for the long-term energy transition but won’t provide immediate relief in the fast-moving race to meet AI-driven demand today. 

Grid enhancing technologies (GETs) and reconductoring existing transmission lines could help maximize the capacity of our current infrastructure. While these solutions can be implemented relatively quickly, their impact varies by location and they are unlikely to fully address the scale of demand. 

Stronger coordination among stakeholders—tech companies, utilities, grid operators, and regulatory agencies—will also be crucial. Improved load forecasting, streamlining interconnection processes for large loads, and better collaborative planning can lead to smarter, more efficient infrastructure deployment. This requires strong leadership and compromise to align competing interests toward broader, long-term benefits.

Each of these solutions can play a role in addressing the energy demands of data centers. Policymakers must carefully evaluate their potential impact, feasibility, and tradeoffs to develop a balanced approach that ensures reliability, economic competitiveness, and clean energy progress.

A two-track strategy: rapid response and system reform

The best path forward requires balancing quick-response solutions with systemic reforms that will reshape the long-term energy landscape.

Potential rapid-response solutions (next few years):

  • Creating energy zones with pre-approved infrastructure and permitting, such as the Special Compute Zones concept proposed by Institute for Progress experts
  • Improving line-of-sight for timing, size, and locations of data center demand growth 
  • Adopting GETs and reconductoring to maximize existing grid capacity
  • Supporting quick-to-deploy behind-the-meter solutions
  • Using creative commercial and regulatory mechanisms to catalyze first-generation volumes of emerging technologies such as geothermal

Longer-term systemic changes (2030s and beyond):

  • Comprehensive permitting reform for energy infrastructure
  • Modernization of grid operations, planning, and market rules
  • Innovation and deployment acceleration for emerging clean energy technologies
  • Creation of sustained coordination mechanisms among private and public sector actors

These tracks may reinforce one another. Quick wins build momentum and stakeholder confidence for deeper reforms, while long-term changes prevent reliance on short-term patches. The goal is to enable sustainable data center growth without jeopardizing grid stability, energy affordability, or emissions reductions.

Moving forward

Like a home renovation, modernizing our energy system requires both immediate fixes to keep things running and fundamental improvements for the future. We can’t afford to shut down the system while we work, nor can we delay action until every detail is perfected.

The good news is that we know many of the tools we need. Policy and regulatory reforms can unlock both rapid-response solutions and systemic changes. With proper coordination and commitment, we can ensure that energy infrastructure doesn’t become a bottleneck for technological progress while maintaining reliability and advancing clean energy goals.

By acting now on both immediate and long-term strategies, we can maintain U.S. leadership in AI and clean technology while ensuring a reliable and sustainable energy future.