Phillip Bruner, Professor & Founder of Climate Risk Lab, Foster School of Business, University of Washington
Phillip Bruner
In the race to dominate artificial intelligence (AI), U.S. tech giants are rapidly expanding data centers and computing infrastructure. Yet they now face a fundamental bottleneck: energy. AI’s rapid growth is straining our outdated power grids beyond anything the tech sector has seen before.
The solution is clear: companies must urgently adopt “net-zero AI” and integrate renewable energy microgrids directly into their infrastructure planning. Doing so isn’t just about sustainability, it’s about reliability, defending against growing threats, and ensuring companies and our nation lead in the next era of innovation.
The status quo can’t keep up.
A single data center can consume as much power as a small city. In fact, U.S. data center
energy use has tripled in the past decade. Left unchecked, data centers could soon demand 10% of national electricity, stressing a grid designed for the mid-20th century.
That grid is also increasingly vulnerable. In 2022,
armed extremists attacked substations in Moore County, North Carolina, plunging tens of thousands into darkness. It took days to restore service. Such incidents are on the rise, alongside more frequent extreme weather events that expose the fragility of centralized systems. Relying solely on traditional grid power is no longer viable for AI-scale operations.
Federal incentives are shifting and states must step up.
Recent changes to the Investment Tax Credit (ITC) mean the long-term federal backstop for renewables is narrowing.
States now hold the keys to accelerating deployment through such actions as fast-tracking permitting and prioritizing projects that can be built within the ITC window.
This creates an opening not just for coastal tech hubs but also for energy-rich states like Texas, Oklahoma, and Georgia. States that move decisively can attract AI infrastructure investment, lower power costs, and strengthen grid security, all while tapping near-term federal incentives before they expire.
Microgrids are the solution within reach.
Renewable microgrids that combine onsite solar, wind, and battery storage give data centers the ability to run through outages, price shocks, and grid delays. For AI infrastructure, that means three things: uninterrupted service during grid failures, faster deployment by bypassing congested interconnection queues, and long-term cost certainty in an energy market defined by volatility.
Bottom-line, this approach is already anchoring America’s most advanced digital infrastructure and should be widely adopted.
A unifying strategy is necessary.
Other solutions lag or overpromise. Traditional nuclear plants take decades to build and aren’t popular with citizens. Modular reactors remain untested at scale and face similar pushback. Commercial fusion is yet unproven.
We need a hybrid approach. First, renewable microgrids are essential, augmented with controlled gas support to firm supply during transitions. Natural gas has a role as a bridge especially where batteries alone can’t yet meet demand. This isn’t about ideology. It’s about ensuring that America’s most critical digital infrastructure has continuous, affordable power that we can depend on.
Leadership is needed to set our path forward.
Tech executives must adopt and consider this strategy as urgent as they do cybersecurity or supply chain integrity. To stay competitive in the marketplace, state leaders should move quickly to prepare their communities for both renewable energy microgrids and AI adoption.
Moving quickly will secure America’s AI leadership, harden communities against growing threats, and set the gold standard for resilient digital infrastructure.
