Grid StrAIn Part 2 — How Utilities Can Get Ahead of AI’s Energy Strain — Before It’s Too Late
Artificial intelligence is already reshaping the energy landscape. Now, utilities must move beyond recognizing the threat and start preparing for it.
The rise of data centers is creating a new kind of load that’s fast, unpredictable, and intense. But the conversation is no longer about whether AI will strain the grid. It’s about how utilities can stay ahead before it’s too late.
The Grid Isn’t Built for AI Volatility
AI energy demand doesn’t follow the predictable cycles. It spikes with new models, facilities, applications, or even small shifts in usage patterns that strain local infrastructure almost instantly. Yet, U.S. data centers alone could account for nearly half of all electricity demand growth by the end of the decade.
Many utilities are already operating at the edge. Much of the U.S. grid was designed for a different era. Transmission capacity is limited. Grid infrastructure is aging. Permitting delays slow down upgrades that are already overdue.
Without faster investment and smarter planning, even minor disruptions can cascade into major failures. In 2024, a lightning arrestor failure in Virginia caused 60 data centers to abruptly disconnect from the grid. While a blackout was narrowly avoided, the event revealed just how little margin for error remains as AI energy density grows.
Four Strategic Moves Utilities Can Make to Strengthen the Grid
Here are four strategic moves utilities can make to prepare for the complexity ahead:
1. Redesign Infrastructure Plans for Volatility
Utilities must adopt scenario-based planning that accounts for demand volatility, regulatory delays, and supply chain constraints. That means rethinking transmission and substation upgrades to accommodate decentralized, high-density computing loads, not just traditional industrial or residential growth curves.
Utilities should also transition from time-based maintenance to condition-based strategies using real-time asset diagnostics. Tools like dissolved gas analysis (DGA) monitors and partial discharge testing systems enable operators to detect stress before failure, protecting uptime while extending asset life.
2. Fortify Supply Chains Against Global Risks
Even the best plans can be upended by global disruption. Tariffs, trade disputes, and inflation are already impacting procurement timelines and costs. Tariffs might not hit chips directly, but they hit everything else that keeps AI infrastructure running—from transformers and bushings to the substations powering data centers. That’s where the real cost risk lies.
According to the U.S. Department of Energy, lead times for large power transformers can now stretch up to 60 months. Utilities can reduce exposure by building domestic supplier partnerships, diversifying sourcing strategies, and integrating global risk scenarios into procurement planning. Policy collaboration with regulators can also help accelerate infrastructure approvals when global headwinds threaten timelines.
3. Build a Future-Ready Workforce
Resilient infrastructure requires a resilient workforce. According to McKinsey, up to 400,000 U.S. energy sector employees are expected to retire within the next 10 years. That knowledge drain is a direct threat to system reliability and modernization.
Utilities must prioritize knowledge transfer, workforce development, and diagnostic interpretation. Condition monitoring, protection testing, and real-time asset diagnostics are only effective when field teams can analyze and act on that data under pressure.
To build that bench strength, utilities can:
- Partner with technical schools and universities to develop next-gen training programs
- Launch mentorship programs to transfer institutional knowledge
- Equip field teams to use diagnostic tools like DGA and partial discharge testing systems
4. Moving from Planning to Action
Strategic planning is essential, but only if it turns into execution. Many utilities have long-term goals but lack short-term mechanisms to course-correct or measure systemic risks.
Implementing regular failure analysis reviews is one way to close that gap. These reviews help utilities identify recurring issues, understand root causes, and refine both design and maintenance strategies based on real-world events.
Grid resilience isn’t just about long-range forecasting; it’s about incorporating agility, accountability, and institutional learning into every layer of the operation.
Shaping the Grid of the Future Starts Now
The AI revolution is exposing critical gaps in our grid, but it also presents a powerful opportunity. The utilities that act decisively, modernizing infrastructure, fortifying supply chains, and developing talent, will define the next era of reliable energy.
This isn’t just a technical challenge. It’s an organizational and geopolitical one that requires long-term thinking, collaboration across sectors and departments, and a commitment to smarter grid modernization.
At Doble, we believe resilience isn’t automatic. It’s built through experience, planning, and action. The time to start is now.
More Information:
- Grid StrAIn: AI & Grid Reliability Part 1—The AI Energy Crunch: How Data Centers Are Reshaping Grid Reliability
- Time is Everything to Power System Reliability
- Living with and by Condition Monitoring
