Utilities spread fixed infrastructure costs across more kilowatt-hours.
Data centers didn’t raise electric bills nationally from 2015–24. Surprise! Actually, they modestly lowered them. That's because big fixed grid costs get spread over more kilowatt-hours, and new demand can unlock economies of scale. @arxiv
Many users rejected claims that data centers lowered national electric bills as misleading, arguing national averages hide local grid stresses, rising infrastructure costs, and future supply constraints.
No Digg Deeper questions have been answered for this story yet.
https://arxiv.org/abs/2606.19777
@JimPethokoukis @arxiv Utilities maintain 10 - 15% excess grid capacity to manage peak loads and outages, if data centers consume 5% of that excess grid capacity, utilities need to add 5% to its infrastructure to maintain its excess grid requirements.
Flawed math.
@JimPethokoukis @arxiv And then what happened?
@JimPethokoukis @arxiv You are comparing this to old school data centers we are talking about hyperscale data centers. We're not talking about micro, edge, modular, or standard data centers here.
@JimPethokoukis The “future supply constraints” mentioned at the end of the abstract have now arrived in most areas. The task of protecting other ratepayers is getting harder as a result.
@davidmanheim Data centers have raised and will absolutely continue to raise energy bills, and this analysis is misleading and also ends in 2024, lol (no shocker there, it’s the AEI).
And besides that, energy use is still an issue irrespective of whether energy bills go up or down.
@JimPethokoukis @arxiv Problem being, this data infers that data centers work within EXISTING infrastructure. New centers, especially more power intensive ones, will require significant infrastructure build out to accommodate them. This means that the average consumer will pay HUGE $$ increases!
I think it’s odd that ppl who can grasp supply and demand intuitively when it comes to, say, housing prices, struggle with this when it comes to energy. Data centers have been adding far less energy to the grid in BTM generation than they have been consuming. This may at some point change, but it’s a pretty basic fact that has and will lead to upward pressure on energy prices.
@PaulBoudreau83 @JimPethokoukis @arxiv There's a lag in data availability, and it takes time to write a paper.
@JimPethokoukis @arxiv Now do years 2025 and 2026.
We’ll wait
@JimPethokoukis @arxiv
@JimPethokoukis @arxiv Build ‼️
@JimPethokoukis @arxiv working to include this data on our data center game
@KristenMeghan @JimPethokoukis @arxiv So show us your analysis of hyperscalers impact.
The basic math works on hyperscalers same as whatever you believe the older data centers were.
On top of this, data centers have thus far been powered nearly exclusively by consumption of baseload energy, for which all new additions have been natural gas. This is emissive! If you care about CO2 emissions, you should care about this.
Of course, the tradeoffs may absolutely be worth it, in fact, I very much think they are. But it’s not an entirely fake issue like water usage.
@JimPethokoukis @arxiv “We caution that future supply constraints could reverse the effect”
@KristenMeghan @JimPethokoukis @arxiv Hyperscalers have been here since 2015 and are a part of this. The economics for utilities is the same. They fund their fixed infrastructure costs, half of their costs, through usage rates. Lower usage; higher rates.
@PaulBoudreau83 @JimPethokoukis @arxiv Are you going to tell us what happened?
@Maxs_Ghost @JimPethokoukis @arxiv We can mandate that they fund those infrastructure improvements directly. We already do that for new power generators.
@JimPethokoukis @arxiv @BusinessInsider
