I accept that data centers generate more load for a system. Which will make the overall system need more maintenance, which is something that others paying into the system will also have to support. But, I'm not clear on why this is a hidden cost.
Consider, if people get the new housing developments that they want, that would also add load to the system. This larger energy system will be more expensive to run, which will lead to higher costs. Adding houses would probably be even more expensive in the transmission maintenance costs associated.
Any model you do that tries to prevent this is essentially rent stabilization for early members. And that has a pretty good track record of not being a good idea.
Homes also generate property taxes and sales taxes (from the occupants inside of them). Cities nearly always make money selling to new homes -- low density suburbs are highly profitable for municipalities, even on utilities alone, both initially and over a 40 year time period.
Data Centers do not work like this. They don't generate any new sales taxes, they don't really generate much in the way of new jobs, and they often don't even pay property taxes at all (our biggest data center here, for example, got a sweetheart deal on a massive property tax exemption -- they literally don't have to pay any property tax at all)
Data Centers also don't pay standard price for their power -- they get 'industrial' power rates (locally here, our industrial power rate is much lower than what a home would pay for equivalent kwh usage, even after factoring in transmission differences).
If you just charged equivalent access (if industrial users had to pay to-the-penny exactly the same prices as a residential user, identical transmission fees, identical per-kwh prices, identical time-of-day usage surcharges, etc), it would go a long way to making the data center setup more fair for everyone.
> If you just charged equivalent access (if industrial users had to pay to-the-penny exactly the same prices as a residential user, identical transmission fees, identical per-kwh prices, identical time-of-day usage surcharges, etc), it would go a long way to making the data center setup more fair for everyone.
Why? An industrial feed to a data centre is the company dealing with a single customer contact, to a single location, probably bulk-buying up front. That's very different to serving 500 houses with very variable demand.
Most data centers almost certainly pay property taxes, as well. It is still a deeded plot of land, after all? I'm curious what data center you know of that doesn't have to pay any. I know incentives are common, but they are usually tied to some other growth metric.
doesn't really match what I'm seeing in Chelan County, WA... the county is now in a fight with the port because the port wants to create a Tax Increment Area to capture the taxes from the two data center buildouts... ~$100m over 25 years. they also estimate $150m sales tax revenue and $70m in property tax revenue for those data centers. one relevant stat is that Microsoft will become the biggest taxpayer in the county.
They're getting that rate because of the reduced cost to support their connection to the grid per kWh. It's essentially the cost of the "packaging". If this is resulting in a loss of revenue for the utility, the blame for that falls on the utility for not properly measuring costs.
Is that also factoring in demand charges? I find this gets forgotten in a lot of discussion of industrial rates, when it is often a large factor in the effective $/kwh.
These larger power-generation systems tend to be more efficient than smaller power-generation systems, not less, which should result in a cost decrease, not an increase.
Tennessee (for example) has fairly cheap electricity because the TVA uses a lot of hydroelectric, and since we have a ridiculous amount of rain and violent thunderstorms each year, every decade or two they build another hydroelectric dam and create a new lake, which generates more hydroelectric power (and a moderate increase in tourism/recreation). We don't have buried power lines (excepting in a very few places) but we've got a ton of redundant power substations and multiple transmission paths (because storms). The TVA and Corps of Engineers are kinda hardcore here otherwise the valley would flood about a quarter of the year and be sitting around in the dark for another quarter of the year.
Maintenance of the power transmission lines is paid for by the electrical customer as a part of paying for the electricity itself. This actually scales just fine. If your local electrical utility is not doing it this way, someone needs to explain to them how proper accounting works.
Calling a "hidden cost" is just a convenient way to say "We're making this up because we feel like it's right and we don't intend to show any proof."
Now let's consider a different form of govt/shared charges.
Taxes, why are individual and corporate taxes and structures different?
Both are doing work and generating income?
Why do corporations get to deductions and do so much tax magic that individuals don't?
why don't we charge them both with the same laws, and structures?
Cause corporates generate jobs? Isn't that unfair to the people born individuals...
I don't think anyone is saying there should be stabilization of electricity prices.
But costs for grid improvements for industrial, data center or AI usage should be on the said companies.
They are using that resource to generate a profit.
While people in residential homes are just living their lives, you are comparing cost of essential commodities to production inputs...
We make taxes on poorer people less as well for the same reason.
But yes if someone is saying electricity prices should be stabilized for early consumers that's definitely unfair. But I didn't see or read that here.
The easy way to justify why corporations get to deduct their spending, is it encourages corporations to spend. Something that a for profit company would not necessarily do otherwise. With the obvious note that spend sent to people is taxed as income for that person.
Now, I agree that this gets super odd when people also make an odd "corporations are people" argument.
As for my assertion that this is effectively arguing for price stabilization. The entire thing hinges on the complaint that costs have been going up. Which, of course the price of a good that has increasing demand is going up. I'm not sure how to read that other than an appeal to price stabilization for early consumers.
>Why do corporations get to deductions and do so much tax magic that individuals don't?
Individuals can do deductions if they're a sole proprietorship. The requirement in general is that the expense is used for a business purpose. Buying a laptop for personal consumption isn't tax deductible, but buying it for your consulting business is.
Make the user who requested the extra power pay for it (upfront then monthly cost)? That's what we do in the UK, then if you stop paying the monthly cost for the 'commit' you surrender it and someone else is allowed to use it without the upfront cost (but with the monthly cost). Why is that hard please?
> Consider, if people get the new housing developments that they want, that would also add load to the system.
It's a matter of scale and efficiency. Housing also adds (full-price) customers new taxpayers.
A major problem with housing development is that low density (e.g. single-family zoned suburbs) are net-negative contributors. (But this effect is primarily seen with infrastructure like roads)
> Any model you do that tries to prevent this is essentially rent stabilization for early members.
The problem with datacenters is twofold:
1) It's a demand-shock. Because of the rapid rollout of datacenters, energy production and transmission capacity simply can't scale up. This causes prices to spike locally.
2) 'Industrial' electricity like this tends to pay very cheap rates, leading to residential electricity customers subsidizing them. Especially as such industry tends to be given high tax & other incentives.
> Any model you do that tries to prevent this is essentially rent stabilization for early members. And that has a pretty good track record of not being a good idea.
A question here is whether or not we're in a datacenter construction bubble. (To anyone who says we're not: Nvidia's stock price has soared more than Cisco's during the dotcom bubble.)
If you're in charge of a major electricity company, are you going to sign off on major expansion investments right now, knowing that it will take 5-10 years?
A lot of these datacenters are not owned and operated by Big Tech itself, but instead disposable companies like CoreWeave. If there's a bubble, it'll pop, and these companies will just go bankrupt. The power purchase agreement'll be worthless then.
My understanding, the last time rising electric costs came up, is that a large part of the cost increase was finally paying for the infrastructure maintenance that companies put off for a time. Is that no longer the case?
That is, it is specifically the transmission and distribution network that is driving costs increases. Per the crappy AI answer "Electricity distribution costs now often exceed the cost of electricity generation for U.S. utilities, with transmission and distribution (T&D) costs potentially comprising over half of a customer's energy bill, up from about one-third previously."
I should say, I have no problem thinking data centers may need to pay more. If only because they are probably able to. My assertion is just that it is misleading to pin them as THE reason costs have gone up. It is far more complicated than that.
I'm in favor of them only if they are extended to renters. I don't know how to do this, but everyone should get a lower cost for their first home, even if it is an apartment. If you want a second home, apartment, campsite, or whatever additional place to dwell I'm fine with you paying a little extra for that luxury, but for your first home we should make it affordable.
Same way I feel about any rules. Worth trying and seeing how they can get the desired outcomes.
I should have said that rent stabilization may work out some day. Just it is not my expectation that it will, as it has a record of not. As such, I would expect some inflationary pressure on utility rates.
Americans are discovering the massive underinvestment in the electrical grid.
When you live in San Francisco, you can experience several power outages during the summer, and it's perfectly normal to have a generator at home.
If you live like me in France, you can go several years without a power outage.
You don't maintain your infrastructure, so you pay ridiculously low costs. And then, you're caught up in reality, here, dominated by data centers. And it's the fault of the GAFAMs ;)
This ignores that the French have massively underinvested in the replacement for the nuclear fleet, and the maintenance of the existing one.
Should we all pretend that half the French nuclear fleet was not offline at the height of the energy crisis? A large portion of the european energy crisis came from the French nuclear power not delivering.
A large renewable buildout should be ongoing in France but wasting money on dead-end subsidies for new built nuclear power seems to be what the French people and politicans wants.
We're seeing a huge number of closures coming in the next decades and the politicians can't even agree on the bonkers large subsidy program for the first 6 EPR2 reactors. Maybe targetting some new power on the grid by 2040? If we're lucky with the construction timeline, at a cool cost of ~20 cents per kWh excluding subsidies.
Americans have to deal with the cost of living near data centres, and everyone on the planet has to deal with global warming. At least the AI companies are providing a free tier that almost anyone with internet can use.
I was looking into energy markets and how they work and it is truly a cluster of moving parts all along the Eastern Interconnect. The question is when is the shoe going to really drop? You can only keep prices going up on an inelastic good before something really bad happens, and this doesn't even touch climate tail risks like heat sagging tx lines across the grid.
Energy markets are deeply complicated markets. They are heavily regulated and have a lot of challenges outside of strictly cost of power. Worth remembering they have safety requirements, uptime, national security implications, truly the backbone of our economic welfare.
Only upside I see out of this huge demand for electricity -- hopefully nuclear will clear the deck on using coal, gas and diesel. If we can build and operate nuclear again we can level our long term cost of power. Combined with renewables its a good combo.
Other than that - power prices will always be derivative to the price of natural gas.
Good news is that line replacement is definitely occurring. Especially for higher power lines, the modern conductors that are used for rewiring lines have lower loses and less sag.
I think it’s coming soon. The data center demand is just one aspect. In New York, a combination of half baked policy to go carbon free by 2030, and 100% EV a few years later, combined with insane political decisions (Gov Cuomo decommissioning the nuclear plant that supplied 20% of NYC), and the black swan event of Trump starting a trade war with Canada (Quebec is the best source of electricity) is making electric delivery rates alone increase 30%.
On the gas side, we’re both forcing ratepayers to spend billions on conversion from low to high pressure gas mains AND prohibiting new gas service.
The environmental lobby is so dumb and ineffective - the public will be demanding that coal come back.
The policy is intended to fail. There will be another wave of failed policy under the guise of “getting it right”. Because if you set the dates for 2050/2060 whenever the state is able to convert over, then businesses won’t do anything until 2050/2060. The most effective method forward is going to be continuously failed but enforceable policy to legally push industries and people into cleaner energy.
Agree that Cuomo decommissioning the plant was DUMB.
> On the gas side, we’re both forcing ratepayers to spend billions on conversion from low to high pressure gas mains AND prohibiting new gas service.
On the gas side, Trump's one BBB greatly increased the caps on natural gas exports.
The reality is that the companies extracting NG are not going to be giving it to Americans, they're going to sell it abroad and rake in way more money.
Which would maybe be fine... if we also weren't currently (and severely) artificially limiting the supply of renewables. Um, oops. There's nothing left.
This is a tangent, but one untapped source of energy savings that seems to be invisible to climate activists is Microsoft Windows' constant drain on resources relative to Linux and MacOS. It's shocking how energy inefficient Windows is even when it's doing absolutely nothing noticable for a user.
This seems unlikely to me. Most of these machines are laptops nowadays. While my Linux system idles at 4W, the plug is only capable of going up to 60W. So even if Windows brought it to the point of nearly overheating the power brick (I’m no Windows fan, but I’m pretty skeptical there!), it would only be 1/10’th of a typical 600W window air conditioner.
The untapped energy savings is, IMO, getting people to run their climate control less. We should toughen up a bit, tell our brains that 50F-80F is the comfortable range. (Depending on humidity and your workload).
I can pretty much guarantee you that the typical Windows desktop machine uses much less energy than the typical Linux desktop machine of similar capabilities. Linux power saving basically never works out of the box unless the user has carefully selected the platform to avoid the Linux kernel's numerous defects. By contrast every computer that comes with Windows has working energy-saving features from the factory because that's how Dell and HP get those Energy Star ratings.
I wholeheartedly agree. Software bloat is a hidden energy drain on corporate America as well as the general public no different than how landlords only leech rents from firms and individuals while providing nothing of value. There is no need to politicize by yelling at climate activists though.
Because it is the exact opposite? Power management and drivers are a shitshow on Linux. When I want my Lenovo (RIP IBM) Thinkpad to last more than 2 hours, I have to boot Windows.
But is is no question: the Apple Silicon chips sip power. If you're looking to minimize watts consumed, a MacBook Air is still on top (or bottom), even if OS X has many pain points too.
There's quite a bit of flim-flam in the video, but the key falsifiable claim is that residential electricity demand is not growing. I am looking at the EIA Monthly Energy Review and it appears that residential electricity grows at ~1% CAGR this century. Which raises the question of whether the bills are now due for past underinvestment in generating capacity.
Great video on how the public is getting screwed on energy deals.
Basically large tech companies have the deep pockets to push up prices at electricity auctions. But why bid in public when you can do those deals in private. That's the first problem. All that needs to be out in the open.
What really irks me is that the market is so manipulated that we can't do anything about it. Think about NEM 3.0 vs 2.0. Putting data centers in their own rate class does make sense as the first step.
>Basically large tech companies have the deep pockets to push up prices at electricity auctions. But why bid in public when you can do those deals in private.
Public utilities can't do the same? Moreover if the implication is that large tech companies are somehow getting great prices at the expense of residential users, what does that mean for the electric generators on the other end of this transaction? Why are they leaving money on the table by selling to large tech companies for cheap?
These companies are regulated and can only charge for the costs they incur plus a flat profit on top of that of 10% or so.
The datacenters give allow them to justify building a lot more capacity to serve them. That increases costs, which means that 10% added for profit is now a bigger number and they can give bigger returns to their shareholders. But those profits are extracted from the existing customers who now see higher bills to cover the costs of expanding capacity to serve the datacenters.
In the US people are complaining about electricity like it’s a scarce resource while China has drastically overbuilt energy capacity (solar and nuclear) and gives 0 shits about AI or crypto. One of these approaches seems better thought through.
Right - because their nuclear power plant construction mechanisms are just starting to scale up and they need something to backfill baseload when solar doesn't cut it. Coal growth is still shrinking and will peak in 2030 at which point it will get squeezed out by all other energy sources.
> […] Applying a new data set for country-level energy prices since 1960, this study evaluates the effects of energy prices on economic growth in 18 OECD countries by controlling for other important macroeconomic conditions that shape economic activity. Mean-group estimates that control for cross-country correlations are used to emphasize average responses across nations. Averaged across all nations, results suggest that a 10% increase in energy prices dampened economic growth by about 0.15%. Moreover, some evidence exists that this response may be larger for more energy-intensive economies.
Readit News
I accept that data centers generate more load for a system. Which will make the overall system need more maintenance, which is something that others paying into the system will also have to support. But, I'm not clear on why this is a hidden cost.
Consider, if people get the new housing developments that they want, that would also add load to the system. This larger energy system will be more expensive to run, which will lead to higher costs. Adding houses would probably be even more expensive in the transmission maintenance costs associated.
Any model you do that tries to prevent this is essentially rent stabilization for early members. And that has a pretty good track record of not being a good idea.
Data Centers do not work like this. They don't generate any new sales taxes, they don't really generate much in the way of new jobs, and they often don't even pay property taxes at all (our biggest data center here, for example, got a sweetheart deal on a massive property tax exemption -- they literally don't have to pay any property tax at all)
Data Centers also don't pay standard price for their power -- they get 'industrial' power rates (locally here, our industrial power rate is much lower than what a home would pay for equivalent kwh usage, even after factoring in transmission differences).
If you just charged equivalent access (if industrial users had to pay to-the-penny exactly the same prices as a residential user, identical transmission fees, identical per-kwh prices, identical time-of-day usage surcharges, etc), it would go a long way to making the data center setup more fair for everyone.
Why? An industrial feed to a data centre is the company dealing with a single customer contact, to a single location, probably bulk-buying up front. That's very different to serving 500 houses with very variable demand.
Tennessee (for example) has fairly cheap electricity because the TVA uses a lot of hydroelectric, and since we have a ridiculous amount of rain and violent thunderstorms each year, every decade or two they build another hydroelectric dam and create a new lake, which generates more hydroelectric power (and a moderate increase in tourism/recreation). We don't have buried power lines (excepting in a very few places) but we've got a ton of redundant power substations and multiple transmission paths (because storms). The TVA and Corps of Engineers are kinda hardcore here otherwise the valley would flood about a quarter of the year and be sitting around in the dark for another quarter of the year.
Maintenance of the power transmission lines is paid for by the electrical customer as a part of paying for the electricity itself. This actually scales just fine. If your local electrical utility is not doing it this way, someone needs to explain to them how proper accounting works.
Calling a "hidden cost" is just a convenient way to say "We're making this up because we feel like it's right and we don't intend to show any proof."
Taxes, why are individual and corporate taxes and structures different?
Both are doing work and generating income?
Why do corporations get to deductions and do so much tax magic that individuals don't?
why don't we charge them both with the same laws, and structures?
Cause corporates generate jobs? Isn't that unfair to the people born individuals...
I don't think anyone is saying there should be stabilization of electricity prices. But costs for grid improvements for industrial, data center or AI usage should be on the said companies.
They are using that resource to generate a profit.
While people in residential homes are just living their lives, you are comparing cost of essential commodities to production inputs...
We make taxes on poorer people less as well for the same reason.
But yes if someone is saying electricity prices should be stabilized for early consumers that's definitely unfair. But I didn't see or read that here.
The easy way to justify why corporations get to deduct their spending, is it encourages corporations to spend. Something that a for profit company would not necessarily do otherwise. With the obvious note that spend sent to people is taxed as income for that person.
Now, I agree that this gets super odd when people also make an odd "corporations are people" argument.
As for my assertion that this is effectively arguing for price stabilization. The entire thing hinges on the complaint that costs have been going up. Which, of course the price of a good that has increasing demand is going up. I'm not sure how to read that other than an appeal to price stabilization for early consumers.
Individuals can do deductions if they're a sole proprietorship. The requirement in general is that the expense is used for a business purpose. Buying a laptop for personal consumption isn't tax deductible, but buying it for your consulting business is.
It seems to work.
It's a matter of scale and efficiency. Housing also adds (full-price) customers new taxpayers. A major problem with housing development is that low density (e.g. single-family zoned suburbs) are net-negative contributors. (But this effect is primarily seen with infrastructure like roads)
> Any model you do that tries to prevent this is essentially rent stabilization for early members.
The problem with datacenters is twofold:
1) It's a demand-shock. Because of the rapid rollout of datacenters, energy production and transmission capacity simply can't scale up. This causes prices to spike locally.
2) 'Industrial' electricity like this tends to pay very cheap rates, leading to residential electricity customers subsidizing them. Especially as such industry tends to be given high tax & other incentives.
> Any model you do that tries to prevent this is essentially rent stabilization for early members. And that has a pretty good track record of not being a good idea.
A question here is whether or not we're in a datacenter construction bubble. (To anyone who says we're not: Nvidia's stock price has soared more than Cisco's during the dotcom bubble.)
If you're in charge of a major electricity company, are you going to sign off on major expansion investments right now, knowing that it will take 5-10 years?
A lot of these datacenters are not owned and operated by Big Tech itself, but instead disposable companies like CoreWeave. If there's a bubble, it'll pop, and these companies will just go bankrupt. The power purchase agreement'll be worthless then.
That is, it is specifically the transmission and distribution network that is driving costs increases. Per the crappy AI answer "Electricity distribution costs now often exceed the cost of electricity generation for U.S. utilities, with transmission and distribution (T&D) costs potentially comprising over half of a customer's energy bill, up from about one-third previously."
I should say, I have no problem thinking data centers may need to pay more. If only because they are probably able to. My assertion is just that it is misleading to pin them as THE reason costs have gone up. It is far more complicated than that.
I should have said that rent stabilization may work out some day. Just it is not my expectation that it will, as it has a record of not. As such, I would expect some inflationary pressure on utility rates.
Dead Comment
Should we all pretend that half the French nuclear fleet was not offline at the height of the energy crisis? A large portion of the european energy crisis came from the French nuclear power not delivering.
A large renewable buildout should be ongoing in France but wasting money on dead-end subsidies for new built nuclear power seems to be what the French people and politicans wants.
We're seeing a huge number of closures coming in the next decades and the politicians can't even agree on the bonkers large subsidy program for the first 6 EPR2 reactors. Maybe targetting some new power on the grid by 2040? If we're lucky with the construction timeline, at a cool cost of ~20 cents per kWh excluding subsidies.
Only upside I see out of this huge demand for electricity -- hopefully nuclear will clear the deck on using coal, gas and diesel. If we can build and operate nuclear again we can level our long term cost of power. Combined with renewables its a good combo.
Other than that - power prices will always be derivative to the price of natural gas.
PJM is a shitshow though.
On the gas side, we’re both forcing ratepayers to spend billions on conversion from low to high pressure gas mains AND prohibiting new gas service.
The environmental lobby is so dumb and ineffective - the public will be demanding that coal come back.
Agree that Cuomo decommissioning the plant was DUMB.
On the gas side, Trump's one BBB greatly increased the caps on natural gas exports.
The reality is that the companies extracting NG are not going to be giving it to Americans, they're going to sell it abroad and rake in way more money.
Which would maybe be fine... if we also weren't currently (and severely) artificially limiting the supply of renewables. Um, oops. There's nothing left.
All of our lobbies are fucking stupid.
The untapped energy savings is, IMO, getting people to run their climate control less. We should toughen up a bit, tell our brains that 50F-80F is the comfortable range. (Depending on humidity and your workload).
But is is no question: the Apple Silicon chips sip power. If you're looking to minimize watts consumed, a MacBook Air is still on top (or bottom), even if OS X has many pain points too.
Basically large tech companies have the deep pockets to push up prices at electricity auctions. But why bid in public when you can do those deals in private. That's the first problem. All that needs to be out in the open.
What really irks me is that the market is so manipulated that we can't do anything about it. Think about NEM 3.0 vs 2.0. Putting data centers in their own rate class does make sense as the first step.
Public utilities can't do the same? Moreover if the implication is that large tech companies are somehow getting great prices at the expense of residential users, what does that mean for the electric generators on the other end of this transaction? Why are they leaving money on the table by selling to large tech companies for cheap?
These companies are regulated and can only charge for the costs they incur plus a flat profit on top of that of 10% or so.
The datacenters give allow them to justify building a lot more capacity to serve them. That increases costs, which means that 10% added for profit is now a bigger number and they can give bigger returns to their shareholders. But those profits are extracted from the existing customers who now see higher bills to cover the costs of expanding capacity to serve the datacenters.
It's a question of aligning incentives.
Deleted Comment
And coal: https://www.carbonbrief.org/chinas-construction-of-new-coal-...
> […] Applying a new data set for country-level energy prices since 1960, this study evaluates the effects of energy prices on economic growth in 18 OECD countries by controlling for other important macroeconomic conditions that shape economic activity. Mean-group estimates that control for cross-country correlations are used to emphasize average responses across nations. Averaged across all nations, results suggest that a 10% increase in energy prices dampened economic growth by about 0.15%. Moreover, some evidence exists that this response may be larger for more energy-intensive economies.
* https://www.sciencedirect.com/science/article/abs/pii/S01409...
Though an interesting graph for Sweden on GDP growth and energy use:
* https://ourworldindata.org/grapher/energy-use-gdp-decoupling