The grid you grew up with wasn’t designed for today’s demand.
The historic approach to the grid was very simple – customers connected to the grid and the utilities provided the power to meet their demands. Utilities indirectly guide consumer behavior through rate structures such as time of use charges (electricity is more expensive at certain times of the day) but otherwise have limited visibility or influence over the loads connected to them. Likewise, customers of energy have always been able to assume that the energy is going to be there for them whenever they need it. Until it isn’t or no longer affordable.
We’ve been adding a lot of expectations to the grid in recent years. Residential air conditioning is growing rapidly in the Seattle area, we are electrifying our heating with heat pumps and buying electric vehicles. This is all new load growth before we even get to the looming impact of AI data centers.
Underneath that demand is an aging infrastructure of generation, transmission and distribution, much of it decades old. That infrastructure was also not designed for the severe weather and wildfire events we are now experiencing on an increasing frequency. The upgrades necessary to rebuild and fortify that infrastructure are being laid in tandem with the load growth to drive electricity prices up.
© Peter Alspach/McKinstry
The Cost Problem
Nationally, residential electric rates have increased by over 30% since 2020. Locally, Puget Sound Energy has increased electric rates by 23% over just the 2025-2026 period and is seeking increases from 2027 to 2029 that would raise it another 30%. These increases exceed general inflation, driving up not just residential energy bills, but the costs of all goods and services.
We’ve been historically lucky in Washington, blessed with low electricity rates that have helped our economy. Our rates are still about 28% lower than the national average. But year-on-year increases so much higher than general inflation and wage growth create real stress on households. These rate increases also create challenges for climate goals around electrification.
The Load Growth Problem
For the majority of the history of the grid, load growth has tracked economic growth. This got disrupted about 15 years ago as efficiency gains (think LED lighting) offset economic load growth. This resulted in an abnormally flat period of load stagnation. But we’ve run through those easy technological wins and are now back to the norm of parallel load and economic growth.
We are also seeing the rise of new technologies that are significantly adding to our load growth: electric vehicles and heat pumps are causing us to switch energy sources from direct fossil fuel usage to electricity, thus moving significant energy uses over to electricity. Charging an EV at home is great, but it’s a new load on the grid. At scale, this is a demand spike the grid was not designed to handle.
And then there is, of course, the AI data center boom on top. While there is no shortage of information on the load growth projections of data centers, it is important to see data center load growth in context with other areas of growth. For example, as we look towards 2040, electric vehicles are projected to have the same demand as data centers, nationally and in Washington State.
The real challenge with data center growth is the scale of a single project – a single data center can easily be 100 MW, equivalent to the charging of nearly 200,000 electric vehicles, or roughly the entire current Washington EV fleet.
The Resilience Problem
So what happens when we have this confluence of load growth on an aging infrastructure? We’ve seen the cost impact. Much of that cost is going to improve the resilience challenges of the current grid. The 2021 heat wave in the Pacific Northwest caused Portland’s streetcar power cables to melt, shutting down service. The heat also resulted in hundreds of heat-related deaths and pushed the grid to the brink, with the Bonneville Power Administration considering imposing rolling blackouts.
The blackouts didn’t happen in 2021, but we were close and the conditions that caused it are ripe to repeat. Since then, our regional home air conditioning has grown from 53% to 64%. As the load increases faster than the infrastructure, the risk of blackouts rises. This is not just happening in Washington State. Utilities across the country are confronting the same math – more load, older infrastructure, and weather events the system was never designed to handle.
Meeting Load Growth with Demand Flexibility and Virtual Power Plants
Demand flexibility is a proven solution that can be delivered rapidly at scale. On September 6, 2022, with an extreme heat wave pushing the grid to its limits, California officials sent an emergency text message to millions of residents via the emergency alert system stating:
“Conserve energy now to protect public health and safety. Extreme heat is straining the state energy grid. Power interruptions may occur unless you take action. Turn off or reduce nonessential power if health allows, now until 9pm.”
According to the US EIA, within 5 minutes demand had reduced by over 2,100 MW below forecast – equivalent to the maximum output of Hoover Dam – and stayed there until 9 pm that night. A simple text message offset the output of one of the ten largest hydroelectric plants in the country. That is the power of two-way grid engagement. And just think – if a simple text message with manual adjustments can have that impact, what could real connected control systems do?
California demonstrated the power of a single text message. In demand flexibility programs the utility sends an automated signal when it anticipates peak demand challenges. Demand flexibility refers to an agreement between a utility and customer to reduce demand on a request by the utility, typically in exchange for payment. Customers can adjust their demand during that period.
Demand reduction measures can include automated adjustments to heating and cooling setpoints that temporarily reduce HVAC energy use during peak periods, without requiring manual intervention, or managed EV charging. The difference is that with a formal demand flexibility program, the customer has a real financial reward for keeping their loads offline. The utility benefits by avoiding the need to purchase expensive energy at times of peak demand or avoiding brownout or blackout conditions from a stressed grid infrastructure.
© McKinstry
Virtual Power Plants (VPP) work in much the same way, but at an aggregated scale. A VPP provider will bring many buildings together to create a coordinated large block of load that can flex at the same scale as a power plant, while also targeting areas of the grid that are particularly congested. While the load shifts all occur on a smaller scale, the controlled and coordinated response creates a more reliable load management asset for a utility. The VPP provider coordinates automated load adjustments between the utility and participating buildings, allowing customers to opt out or override when necessary.
The commercial building stock is largely untapped by demand flexibility and VPPs. We have a tremendous opportunity to use our existing assets to automatically reduce or shift load on demand. These reductions cost almost nothing to build – it is built into our existing building controls systems. How we design the programs matters – when done well, VPP programs can lower the costs for all customers.
Organizing and connecting these load reductions enable us to use the grid we already have more wisely, reducing the need for new power plants, limiting costly infrastructure upgrades, and improving overall grid reliability. Traditional grid infrastructure typically requires years of planning, permitting, and construction before capacity ever comes online. With softwaredriven solutions, utilities can bring capacity online in months rather than years, while also lowering costs for customers and creating new revenue for participating buildings. Much of the infrastructure we need is already built. We just haven’t connected it to the grid yet. The cheapest power plant and transmission line is the one you don’t have to build.
About the Authors
Peter Alspach, PE Director of Innovation, McKinstry
Jesse Rebello, Senior Director of Dawson Ventures, McKinstry