AI Is Stressing the Grid. Pervez Siddique Says Batteries Are the Missing Piece
Data centers push grid power demand to new highs as experts warn policy and storage must catch up.
Artificial intelligence is fueling a surge in electricity demand that the world’s grids are struggling to absorb. The International Energy Agency estimates that data centers tied to AI could quadruple their power use by 2030. In the United States, where extreme weather and electrification are already straining supply, policymakers are scrambling.
Texas lawmakers recently gave utilities authority to disconnect large data centers in emergencies. Google (News – Alert) agreed to scale back AI workloads when demand peaks. And the Department of Energy has launched a “Speed to Power” initiative aimed at accelerating grid projects.
The rush reflects a growing reality: AI’s growth is outpacing the infrastructure built to support it.
Batteries as Shock Absorbers
Grid-scale batteries are emerging as one of the most effective tools for managing volatility. They can charge when renewable output is abundant, discharge during peaks, and respond within seconds to sudden swings. Texas has already deployed projects in the hundreds of megawatts, a scale once considered improbable.
Yet markets often undervalue the services batteries provide, such as frequency regulation or rapid ramping. That gap could slow investment just as demand is spiking. “Storage is not treated as core infrastructure in many regions, even though it delivers reliability the grid now depends on,” says Pervez Siddique, a clean-energy strategist with experience developing multi-gigawatt portfolios.
Moving Beyond Paper Renewables
Technology companies have long relied on contracts with remote wind or solar farms to claim “100 percent renewable” status. Analysts say that model is showing its limits.
“Buying clean megawatts on paper does not guarantee that power is available in real time,” Siddique notes. Co-locating data centers with renewables and storage, he argues, provides greater reliability, cuts transmission losses, and reduces congestion on the grid.
Several developers are now pursuing hybrid campuses that integrate computing facilities with on-site clean generation and battery storage. The approach is still new, but it is quickly moving from concept to practice.
Policy Playing Catch-Up
AI workloads are reshaping demand curves in ways that legacy rules do not anticipate. Experts point to three priorities for regulators:
– Incentives for data centers to participate in demand-response programs.
– Market structures that properly compensate storage for grid services.
– Streamlined permitting and interconnection to prevent multi-year delays.
The Department of Energy’s recent initiative signals recognition at the federal level, but the scale of demand suggests that state and regional reforms will be equally important.
Lessons From the Field
Industry practitioners emphasize that technology alone does not guarantee success. Financing, permitting, and community engagement often decide whether projects move forward.
“Large-scale projects succeed when they diversify revenue streams and build local support from the outset,” Siddique says. Modular deployments, he adds, can provide quick capacity while larger projects navigate approvals.
A Narrow Window
Analysts see the next five years as critical. AI demand is climbing sharply, but grid infrastructure moves at a slower pace. Without targeted investment and reform, reliability risks will rise.
Siddique points to integrated planning as the way forward: “We need frameworks that recognize AI demand explicitly, accelerate interconnection, and align financing with projects that combine clean energy, storage, and computing.”
Transparency will also matter. As pressure builds from regulators and investors, data centers may face requirements to disclose energy use, efficiency, and emissions more rigorously.
The Test Ahead
The surge of AI represents more than a technical challenge. It is a test of whether innovation can scale without overwhelming the systems it depends on.
“The transition is not just about adding more renewables,” Siddique says. “It is about redesigning the grid so generation, storage, and demand are aligned. If we get that right, AI can accelerate the build-out of smarter infrastructure. If we get it wrong, it becomes a liability.”
