Your title is wishful thinking. Sorry, but Data Centers are raising power bills.
Your evidence runs 2015–2024. You concede in the piece that this “occurred while the grid still had slack.” If you were acting in good faith or a good analyst you’d drop this period as it’s no longer relevant.
PJM (which has the most data centers) independent market monitor is measuring the period after the slack ran out. They said Data centers drove 63% of the 2025/26 auction increase, that’s $9.3 billion recovered from ratepayers in one year.
The Fed also disagrees. From the June minutes: “Many participants noted that ongoing strong demand for AI infrastructure would likely sustain upward pressure on prices for technology products and electricity.”
Even your Oregon example refutes your title. Regulators approved a 29.7% increase on PGE’s largest users to cover the costs their demand creates, and residential rates fell 1.3%. You cannot lower a household bill by removing a cost that was never on it. That 1.3% is a measurement of what households were carrying.
All complaining, conspiracy theories come from leftist surrogates, bought and paid for by China. The slower the US goes , the better for China. Fortunately DJT put an end to the craziness of Biden's "anything but fossil fuels " stupidity. /ap
Half agreed: the hardware is the wrong defendant. But I built these things for a career, & the record cuts against the other half. From 2014 to 2016 US data center power held nearly flat around 60 terawatt hours while cloud workloads exploded, because consolidation mandates & efficiency targets absorbed the growth. That was policy doing what this piece says policy cannot. The layer that did that work got cut this past year, order by order. Blame policy, yes. Pick the right policies. The run, with dates & receipts: https://carlmfreeman.substack.com/p/the-fences-were-there-for-a-reason
That map alone proves these studies are bullshit and likely excluding ancillary fees from "residential rates". Power rates in the southeast have skyrocketed the past 6 years. I know; I live here.
The answer isn't simple of course, but the question is.
For a conserving residential customer in VA, Dominion reported that the bill last year from 5/13/25 to 6/11/25 was $151.51 for using 1,041 kWh (14.55 cents per kWh), and for this year same period from 5/13/26 to 6/10/26 was $176.70 for using 970 kWh (18.22 cents per kWh). That's a 25% increase on the face of it.
Solid piece with real data! Data centers can help lower average rates by spreading fixed costs — when they’re allowed to. As you note, the real risk isn’t the load itself but the potential for costs to get socialized once headroom runs out. The solution you highlight — dedicated rate classes and large-load tariffs — already works in several states. Legislation and smarter regulation are clearly the answer, which is why New York’s ban on large data centers is such a self-defeating move. I wrote about it here:
Leftists are gonna leftist, and don’t feel need to have their claims tied to reality.
I did find it strange that other wise very good piece you commented approvingly on the Oregon increase because it was “costs their demand creates.” If for required additional infrastructure, then of course I’m all for that. But if solely based on the incremental load, why are data centers additional load uniquely “costs their demand creates”? Seems a false premise to me.
Your title is wishful thinking. Sorry, but Data Centers are raising power bills.
Your evidence runs 2015–2024. You concede in the piece that this “occurred while the grid still had slack.” If you were acting in good faith or a good analyst you’d drop this period as it’s no longer relevant.
PJM (which has the most data centers) independent market monitor is measuring the period after the slack ran out. They said Data centers drove 63% of the 2025/26 auction increase, that’s $9.3 billion recovered from ratepayers in one year.
The Fed also disagrees. From the June minutes: “Many participants noted that ongoing strong demand for AI infrastructure would likely sustain upward pressure on prices for technology products and electricity.”
Even your Oregon example refutes your title. Regulators approved a 29.7% increase on PGE’s largest users to cover the costs their demand creates, and residential rates fell 1.3%. You cannot lower a household bill by removing a cost that was never on it. That 1.3% is a measurement of what households were carrying.
https://pulpitfortruth.substack.com/p/its-getting-hot-in-here
All complaining, conspiracy theories come from leftist surrogates, bought and paid for by China. The slower the US goes , the better for China. Fortunately DJT put an end to the craziness of Biden's "anything but fossil fuels " stupidity. /ap
Nope, I’m not a left winger. That we gotta beat China line would hit a lot better if the Govt acted like they really believed it.
But they don’t.
https://pulpitfortruth.substack.com/p/watch-what-they-do
Half agreed: the hardware is the wrong defendant. But I built these things for a career, & the record cuts against the other half. From 2014 to 2016 US data center power held nearly flat around 60 terawatt hours while cloud workloads exploded, because consolidation mandates & efficiency targets absorbed the growth. That was policy doing what this piece says policy cannot. The layer that did that work got cut this past year, order by order. Blame policy, yes. Pick the right policies. The run, with dates & receipts: https://carlmfreeman.substack.com/p/the-fences-were-there-for-a-reason
That map alone proves these studies are bullshit and likely excluding ancillary fees from "residential rates". Power rates in the southeast have skyrocketed the past 6 years. I know; I live here.
The answer isn't simple of course, but the question is.
For a conserving residential customer in VA, Dominion reported that the bill last year from 5/13/25 to 6/11/25 was $151.51 for using 1,041 kWh (14.55 cents per kWh), and for this year same period from 5/13/26 to 6/10/26 was $176.70 for using 970 kWh (18.22 cents per kWh). That's a 25% increase on the face of it.
Someone care to explain?
SpaceX is going to launch a million orbital data centers into space starting next year. These will equate to thousands of land-based data centers.
Solid piece with real data! Data centers can help lower average rates by spreading fixed costs — when they’re allowed to. As you note, the real risk isn’t the load itself but the potential for costs to get socialized once headroom runs out. The solution you highlight — dedicated rate classes and large-load tariffs — already works in several states. Legislation and smarter regulation are clearly the answer, which is why New York’s ban on large data centers is such a self-defeating move. I wrote about it here:
https://www.linkedin.com/pulse/new-york-just-banned-something-wasnt-comingand-killed-sarno-ognnc/
Thanks for the piece.
Leftists are gonna leftist, and don’t feel need to have their claims tied to reality.
I did find it strange that other wise very good piece you commented approvingly on the Oregon increase because it was “costs their demand creates.” If for required additional infrastructure, then of course I’m all for that. But if solely based on the incremental load, why are data centers additional load uniquely “costs their demand creates”? Seems a false premise to me.