Bending the Cost Curve: The Agility Imperative

Michael Chua
PowerShift Contributor

First Published Q3 2025

Picture this: A utility executive receives a call about a proposed AI data center that will consume as much electricity as a quarter of their entire service territory. The facility is the size of a Walmart, the timeline is aggressive, and traditional grid expansion would take years and cost hundreds of millions of dollars. This isn’t a hypothetical scenario; it’s happening right now across America.

“[The demand] from electric vehicles was large, yet predictable; we knew how this huge new source of need for the grid would unfold for the utilities, bought one at a time and over a long period of time,” explains John Reed, Solution Architect at GridX, a utility rate strategy and technology firm. “But right now, we’re at an inflection point in the grid’s history. The electricity demand is growing massively, and new AI data center load is enormous and sudden.” These facilities appear almost overnight, creating massive loads that utilities must respond to with unprecedented speed.

This reality is driving utilities toward what Reed calls “bending the cost curve,” using rates, consistently sampled data, dynamic pricing, and strategic agility to drive efficiency while avoiding costly infrastructure buildouts. It’s a fundamental shift from the traditional utility playbook of building more capacity to meet growing demand.

The Multi-Front Challenge

Today’s utilities are navigating three converging pressures simultaneously. Electric vehicle adoption continues its steady climb as AI and cryptocurrency mining operations are creating unprecedented instantaneous demand spikes. Meanwhile, distributed energy resources (DERs) like rooftop solar and battery storage are fundamentally altering traditional load patterns and grid management requirements.

“The grid is not as predictable as it once was, and it’s becoming more complex,” Reed notes. “You have intermittent power that’s sometimes sunny when electricity isn’t needed, and sometimes cloudy when it is needed. Sometimes wind is blowing in one area and not in another. When you add massive new loads and EV charging, it creates a far more real-time and dynamic set of needs,” spawning new challenges for grid operators who must balance supply and demand in real-time across increasingly complex systems.

The Speed Imperative

Increasing supply by building more plants can be costly, so smart utilities that seek to improve efficiency strategically must look to influence demand. One of the most powerful tools a utility possesses to influence demand is rates, but traditional rate design and implementation cycles, which often take years from conception to customer billing, are fundamentally incompatible with today’s rapidly evolving grid conditions. Reed points to this timing mismatch as a critical weakness in current utility operations.

“Historically, the way rates were designed, analyzed, put into Customer Information System (CIS), and billed… that end-to-end process takes a long, long time,” he explains. “From the date that a rate is envisioned, all the way to the time that it’s actually billed to a customer, can be many years sometimes. With GridX, essentially, we cut all the way through that. When you enable a rate in our Design product, it’s in our rate engine, and it’s now immediately available for billing and customer display. We offer the ability to get rates set up in a faster way.”

This is where integrated approaches, like those at GridX, differ from legacy systems; rather than designing rates in one system, analyzing them in another, and implementing them through a third, this tactic of using a dedicated platform enables end-to-end rate management in a unified environment. In this way, utilities may leverage technology to speed up rate creation, which in turn allows them to more rapidly influence demand and adapt to a quickening world more swiftly.

“There’s a set of principles that have been around for a long time, called the Bonbright Principles,” Reed explains, “which essentially are how public utility commissions analyze rates. There are ten principles, like rates need to be fair among customer classes, they need to be economically efficient, and they need to be simple, etc. And,” he adds emphatically, “I would say that the Bonbright Principles need to add another principle,” Reed proclaims, “which is speed.”

“We’re the first end-to-end rate engine,” Reed highlights excitedly, “and that’s how we enable speed. That never existed previously in the world of rates for utilities.”

Dynamic Pricing: Beyond Time-of-Use

Speed isn’t the only consideration when driving demand efficiency. While time-of-use (TOU) rates have been around for decades, modern dynamic pricing encompasses a much broader spectrum of strategies to influence usage and further bend the cost curve.

Critical peak pricing, for example, which Reed likens to Surge Pricing on Uber, triggers during specific high-demand periods that are unpredictable but are statistically certain to occur. Customers must actually reduce usage at particular times or be penalized with a higher price.

Peak time rebates, on the other hand, offer “all carrot, no stick” approaches where customers are automatically enrolled and receive benefits for reducing usage during peak periods.

All these strategies and more Reed places on the spectrum towards true dynamic pricing. The most sophisticated approach, which Reed acknowledges utilities must still overcome certain technology hurdles to achieve, involves real-time pricing that exposes customers to actual market mechanisms reflecting current system conditions. “Whether it’s routing a ChatGPT request to the data center location with the cheapest power on the grid instead of to one experiencing peak demand, these optimizations are easy to describe but extremely difficult to build,” Reed says.

The Data Science Advantage

Success in dynamic pricing hinges on comprehensive analytics capabilities. “The big question with time-of-use rates and rollouts is: how effective are they?” Reed explains. “If they don’t benefit customers, customers resist. If they don’t actually reduce cost to serve by meaningfully shifting usage away from expensive times, they’re not really effective.”

One impactful approach involves crunching the numbers and doing full, data sciencetype analytics using Advanced Metering Infrastructure (AMI) data. “We can do full population recalculations of any kind based on that AMI data,” Reed says. The platform ingests historical AMI data across entire customer bases, along with customer information systems (CIS) data, creating what Reed describes as “a highly optimized architecture” capable of comparing any customer’s performance on any rate structure.

“We’re increasingly applying data science to our own data,” he notes. “We can study load patterns before a customer was on a TOU program and after, showing their propensity to shift or reduce usage with hard numbers.”

This data-driven approach proved crucial in recent analysis showing measurable load reduction from TOU programs, providing objective evidence to support continued investment in these initiatives.

Real-World Success: The BGE Model

One of Reed’s favorite implementation stories involves Baltimore Gas and Electric’s innovative rebate program: a pure “carrot, no stick” approach that demonstrated dynamic pricing’s potential.

“It was a truly innovative program that allowed BG&E to avoid building an entire power plant,” Reed recalls. “It saved the utility quite a lot of money, saved the customer base quite a lot of money, and cut through problems people have with rates; they didn’t want to penalize people.”

The program’s success lay in its design: customers could only benefit, never be penalized, while still providing meaningful load reduction during critical peak periods. This approach addressed a fundamental tension in rate design between creating sufficient price signals to drive behavior change while maintaining customer acceptance.

Looking Forward: The Technology Imperative

Reed sees rate capabilities as the most critical innovation utilities must embrace in the next f ive years. “Rates are the key to bending the cost curve,” he emphasizes. “Whether it’s a rate for a new AI data center or optimizing existing infrastructure, the vision remains the same – we need technologies that unlock this capability.”

The integrated approach GridX represents, combining rate design, billing, customer engagement, and analytics in a single platform, addresses what Reed identifies as a fundamental technology barrier that has prevented utilities from realizing the full potential of their AMI investments.

“Companies like GridX are blazing the trail,” he says. “We’re excited to move from theoretical to real, getting utilities’ data into our application and showing them these capabilities in weeks and months rather than years and quarters.”

For utility leaders, the message is clear: the complexity isn’t going away, and the traditional approach of building infrastructure to meet peak demand is becoming economically unsustainable. The utilities that master dynamic pricing and grid optimization will thrive; those that don’t risk being left behind in an increasingly competitive and complex energy landscape.

The cost curve won’t bend itself, but with the right tools and strategies, utilities can harness complexity as a competitive advantage rather than an operational burden.