A new Senate bill would protect consumers from rising energy costs for data services. Instinct is right. The analysis is incorrect.
Data centers now make up about 7% of US electricity demand, which is about enough to power every home in California and Texas combined, up from 1% 15 years ago – which is the equivalent of powering every home in California and Texas combined. That spine is rising. The four largest hyperscale technology companies are projected to spend a combined $650 billion in capital expenditures this year alone. When the numbers are so high, it’s natural to ask if the current system makes sense.
But this bill treats data centers as a problem. The truth is more interesting than that. Data center problems are a symptom of networks that have been built in a low-tech and modern way for decades, but the right data centers, designed in the right way, can help solve these problems. And the right data centers, designed the right way, can help fix it.
The Real Problem is the Grid itself
The fact is, the thing that really drives electricity up is that our electricity grid is weak in nature. It was built for the 20th century world of slow, predictable growth, a time when utilities could predict load in advance and build generation accordingly. That world is over.
But data centers are only one reason.
Electric vehicles are changing how and when millions of Americans draw power. Heat pumps are changing residential electrical systems. Industrial electrification is accelerating across manufacturing and chemicals. Each of these changes represents real economic development – new industries, new jobs, new skills. But each also puts new challenges on a network that was not designed to accommodate them. None of that is a “problem”. Even data centers.
Data centers are a visible new source of demand, which makes them an appropriate political target. But singling out one sector for the collective challenge of network development is a bit like blaming traffic congestion on new cars when the roads were already too narrow.
Transmission bottlenecks, connection backlogs, outdated planning methods that make it very difficult to bring new capabilities online: these problems were being built long before the current wave of AI-driven data center construction, and they will continue to build with or without it. Every year we slow down to improve the internet, we raise the cost of our economic growth.
Three Things Policy Makers Should Do
So what should lawmakers, in this law and beyond, do?
1. Treat variable demand as a network resource.
First, they should run an energy policy on flexibility as a grid resource. A series of reports from the Nicholas Institute of Duke University found that reducing only 0.25% to 1% of the annual use of electricity during the most stressful hours of the year could allow the US grid to absorb up to 100 gigawatts of new load – almost the entire size of the American nuclear fleet – without requiring large investments in new generation or transmission. A recent study estimates that if large data centers shifted their computing power to peak hours, the country could avoid up to $150 billion in factory electricity, fuel and transmission costs over the next decade.
Much of the capacity we think we need to build already exists. Let’s just use it properly. The law should guide regulators and network operators to recognize fluctuating demand as well as normal supply in resource adequacy planning.
2. Promote internet-enabled data centers.
Instead of restricting access to the grid, the law should require data centers to be designed for grid functionality – the ability to dynamically adjust energy consumption to match the grid’s needs. The technology to do this is real and can be used today. Data centers can be built with integrated battery storage that provides services to the network during normal operation and energy storage during outages, reducing the load within minutes of the utility signal while saving customer time.
Recent work confirmed by national laboratories has demonstrated this flexibility in full scale: these devices can reduce 100% of the load of the network in one minute of the signal of use, providing a solid capacity to restore the network with battery storage, and reconnection in a stable manner when conditions are stable. Data centers designed this way are not a burden on the grid. I am an asset to it.
The same principle applies throughout the area of demand. Efficient electrical systems now connect millions of residential devices – thermostats, water heaters, home batteries – to change usage during peak hours ~ ~, providing gigawatts of manageable capacity ~ ~. The Department of Energy estimates that increasing this network can meet 10 to 20% of peak demand in 2030, saving $10 billion annually in avoidable infrastructure costs.
3. Update coordination and planning processes.
Today’s projects were designed for slow times. They assume that all new demand requires an equal amount of new supply – and that energy-intensive systems can’t supply it either. Both of these feelings are getting worse. Flexible loads, smart storage, and advanced demand coordination should be considered as potential tools in network design, with appropriately developed control structures.
The Danger of Getting the Framing Wrong
These are not imaginary thoughts. These are proven capabilities that are currently in use. The question is whether the legislators will build on them or continue to create the challenge as a zero-sum competition between data centers and customers.
Legislation that desegregates data centers may gain political traction, but it leaves structural limitations that will continue to drive up costs for everyone. The grid needs an upgrade that intelligently accommodates all the loads of the 21st century: EVs, heat pumps, industrial electricity, and yes, data centers. Building walls around one demand group while the infrastructure is weak will not protect consumers. It will delay the updates that would otherwise do so.
The instinct to protect taxpayers from rising costs is correct. The way to honor it is to build a network that can meet the historic opportunity that is coming.
Opinions expressed in Fortune.com’s comment sections are solely the opinions of their authors and do not necessarily reflect their opinions and beliefs. Good luck.
#Data #centers #dont #destroy #internet #broken #grid #Good #luck