"In order to protect uptime of our glorious data centers, neighborhoods will begin experiencing rolling brownouts to reduce demand."
- Texas soon probably.
this post was submitted on 28 Feb 2025
200 points (99.0% liked)
Technology
63707 readers
3361 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
So why is it the duty of our country to gather all electricity possible for the richest people to waste on burning out GPUs so they can lose money on free chatbots?
For the same reason housing should be a speculative investment, and healthcare services available only to the highest bidder.
Data centers need to bring their own power.
In a well regulated way that includes oversight, yes.
To a significant extent, they do, contracting for construction of generation and transmission (very often renewable), at least at the largest scale.
But, it's (mostly) all on the grid.
With demand like that, it's not like there isn't significant negotiation with the local power company, especially because they're frequently built a significant distance from existing large power infrastructure.
Heck, all the big 3 cloud providers signed deals for nuclear generation in the last few months. https://spectrum.ieee.org/nuclear-powered-data-center
Here's just one more article about these sorts of investments: https://www.canarymedia.com/articles/clean-energy/google-has-a-20b-plan-to-build-data-centers-and-clean-power-together
Heck, all the big 3 cloud providers signed deals for nuclear generation in the last few months. https://spectrum.ieee.org/nuclear-powered-data-center
Subsidized by US taxpayers ... If data center flops, we pay hold the defaulted loan
If demand is there, microshit get cheap nuke energy and operator makes profit...
Where is the benefit to the taxpayer?
A few job and chatgpt flooding internet?!
Clown fucking world
The one state that refuses to connect to the interstate power grid and has Uber-like surge pricing on electricity? Yeah, I'm sure this won't result in regular people footing the bill for more billionaire profits.
Texas is a joke, but not a good one.
Uber-like surge pricing on electricity
We don't really: that story you heard from a few years ago was the only company that billed like that. The customers made a bet that the pricing averages through the day (lower at night, higher cost during the day) would average out in their favor over fixed-cost billing, and frankly, it did right up until it didn't.
They took a risk and got bit by, frankly, not understanding how the system works and basically ate the spikes.
Everyone else paid $0.09/kwh or so during that whole period, and the electric providers ate the cost because when you're averaging out spikes across millions of kwh, it won't lead to bankruptcy.
They took a risk and got bit by, frankly, not understanding how the system works and basically ate the spikes.
It's the exact same idea as insurance. You don't buy insurance because you think you'll take the insurance company for a ride, you buy insurance to even out your costs. If someone hits you, you don't need to fork out tens of thousands of dollars for medical bills and repairs, but you will fork that out over time instead with more manageable payments.
If you don't want to see scary bills, then pay a little higher average prices so you end up with a consistent bill.
load more comments
(18 replies)
One of the windiest, sunniest, emptiest places on earth and they want to waste water building reactors instead of renewables.
Hell, the geology means you can store energy in the ground using pressurized air.
What? I've grown up around people in the nuclear industry, and nothing I've ever learned about the function "wastes" water.
Some rambling on how I understand water to be used by reactors
You've got some amount of water in the "dirty loop" exposed to the fissile material, and in the spent fuel storage tanks. Contaminated water is stuck for that use, but that isn't "spending" the water. The water stays contained in those systems. They don't magically delete water volume and need to be refilled.
Outside of that you have your clean loop, which is bog standard "use heat to make steam, steam move turbine, moving turbine make electiricity, steam cools back to water". Again, there's no part of that which somehow makes the water not exist, or not be usable for other purposes.
Not saying you're wrong. Renewables are absolutely preferable, and Texas is prime real estate to maximize their effectiveness. I'm just hung up on the "waste water building reactors" part.
Guessing it was some sort of research about the building process maybe, that I've just missed?
How do you condense the steam back to water?
load more comments
(7 replies)
Hmm harness the holy light of the sun?
First 0 nuclear reactors will be built anywhere in US before 2035.
Texas is actually a renewables leader because, believe it or not, it has the least corrupt grid/utility sector, and renewables are the best market solution.
Even with 24/7 datacenter needs, near site solar + 4 hour batteries is quicker to build than fossil fuel plants and long transmission, and it also allows an eventual small grid connection to both provide overnight resilience from low transmission utilization fossil fuel as peakers anywhere in the state as well as export clean energy on sunnier days.
Market solutions, despite hostile governments, can reduce fossil fuel electricity even with massive demand surge. One of the more important market effects is that reliance of mass fossil fuel electricity expansion and expensive long high capacity transmission, would ensure a high captive cost at high fuel costs because of mass use, in addtion to extorting all regular electricity consumers. Solar locks in costs forever, including potentially reducing normal consumer electricity costs.
"The least corrupt/utility sector" I must be thinking of the wrong Texas, which one are you referring too?
I think they mean "the same forces that led to the grid collapsing every few years -- prioritizing profit above all else, and the government giving zero fucks-- are the same forces which trigger new development to be in renewables with zero regulation or oversight"
Conservatives always write about their broken-clock-right-twice successes in a similar way.
near site solar + 4 hour batteries is quicker to build
But is it quicker at scale? Can solar and battery production keep up with expanding demand? Can it continue to do so over 10+ years? Can it outpace demand and start replacing fossil fuels?
Usually the proper solution is a mix of technologies. It shouldn't be solar vs nuclear vs wind, but a mixture.
Nuclear does a great job at providing a large amount of energy consistently. It's really bad at fluctuations in demand, and it's also really bad at quick rollout. I think it makes a lot of sense to build nuclear in Texas over the long term because it can start filling in demand as efficiency of older panels and batteries drop off, which extends the useful life of those installations and reduces reliance on battery backups.
I also think hydrogen is an interesting option as well, since it's sort of an alternative to batteries, which can be hard to get at scale. Use excess generation for electrolysis and use those for mobile energy use (e.g. trucks, forklifts, etc) or electricity generation. It's also not ideal, but it could make sense as part of a broader grid setup.
Solar is awesome and we need more of it. I just want to encourage consideration of other options so we can attack energy production from multiple angles.
Can solar and battery production keep up with expanding demand?
China is expanding so fast that they are accused of overproducing, and so supply capacity is not only there, it can increase further.
Usually the proper solution is a mix of technologies. It shouldn’t be solar vs nuclear vs wind, but a mixture.
The main benefit of wind is in battery reduction. A capacity equal to lowest night demand. Wind often produces longer hours than solar per day. The predictability of solar allows clear power forecasts, and then enough solar for needs with a small grid connection allowing both monetizing surpluses, and having resilience in shortfalls. Nuclear has no economic or climate roles, for being both too expensive and of too long a delay.
I also think hydrogen is an interesting option as well, since it’s sort of an alternative to batteries,
Hydrogen is the solution for having unlimited renewables and being able to monetize all of their surpluses. It is a bonus to be able to provide emergency/peak power, including renting a vehicle to have bonus value of powering a building. For today, backup fossil fuel generators can still provide resilience value to solar.
For today, backup fossil fuel generators can still provide resilience value to solar.
And that's the issue. Nuclear is an effective alternative to fossil fuels and can make sense in many areas. What you need is:
- lots of space for waste disposal
- prevent disruption from activist opponents (delays drive up costs)
- enough projects that you get economies of scale for construction (e.g. specialized crews can move from site to site)
- high enough base load demand to fully utilize nuclear
France has a ton of nuclear and it is on the cheaper end for electricity rates in Europe, and they're not particularly well-suited for it.
It's not a panacea, but it should absolutely be considered as a replacement for fossil fuels if energy demand is high enough.
Using existing infrastructure for backup/resilience as renewables are ramped up is the ideal. Was German last government's approach. Cheaper (free) than even maintaining/refurbishing aging nuclear, allowing for private sector to expand renewables (also free). Standby payments to stay open and ready is cheap, and permits rapdid renewables to decrease their peaker use.
"Baseload" nuclear has the inverse problem of renewables. It needs to sell all of its very expensive power near 24/7. Costs being dominated by its initial building, means that half capacity is double the breakeven power revenue. Nuclear needs to suppress cheaper energy to be viable, and in the ultra optimistic (Vogtle took 20 years) 10 year buildout period, renewables must be suppressed so that when the ON switch is set, full power sales occur.
France has a ton of nuclear and it is on the cheaper end for electricity rates in Europe
France has understood that building new nuclear should wait until 2060s, when possible construction technology is advanced enough. The heyday of nuclear came when electricity demand was growing fast, and fears of available reserves and geopolitics affecting alternatives. Coal is also excessively polluting and dirty, in a locally displeasing way. The environment of alternatives is much different today.
“Baseload” nuclear has the inverse problem of renewables. It needs to sell all of its very expensive power near 24/7.
Excess nuclear production at night recharges batteries for daytime use, reducing the need for battery and solar rollout. Excess solar production during the day recharges batteries for nighttime use, reducing the need for baseload supply. Daytime use is higher than night time use, so this is pretty close to the ideal setup, no?
Use each non-polluting source for what it's best at. You don't need any one source to be the primary supplier of electricity, you want a diverse enough set that you get an optimal mix to keep costs and pollution low and reliability high. Mix in some wind and others for opportunistic, cheap generation.
Yes, both can charge batteries. Solar charges then at 10x less cost, and combined solar+batteries provides the same total "baseload function" at 2x-4x less cost, and can be up and running in 1 year instead of 10, and expanded the year after that. It's even a myth that nuclear uses less land. You can use the land under solar, and you don't need exclusion zones around reactors and uranium mines
It's lower initial cost, sure, but what about longer term? Surely battery costs add up long term as they need to be expanded and replaced, making nuclear more attractive after 10-20 years.
I'm not an expert here though, I'm merely saying a lot of people would be fine with a higher initial investment if the long term benefits justify it.
It’s lower initial cost, sure, but what about longer term? Surely battery costs add up long term as they need to be expanded and replaced, making nuclear more attractive after 10-20 years.
No. Nuclear also has fairly high operations/staff costs, and fuel is highly variable and more expensive the more other nuclear plants there are. You mentioned the possibility of charging batteries (Hydrogen also possible) from nuclear, to handle peak day use/transmission, but batteries pair better with solar, and as a total package can serve same "baseload" purpose as nuclear but cheaper. There are no long term benefits to nuclear... economic ones ignoring weapons motivations.
Hydrogen also possible
Yeah, I just think of hydrogen as a battery, and it can totally be a closed loop system.
batteries... cheaper
Is that actually true though? As in, if we add up initial installation cost + running cost + replacement cost long term (say, 50 years), are batteries generally cheaper?
If so, then I'd agree. But my understanding is that nuclear gets really competitive the longer it runs.
if we add up initial installation cost + running cost + replacement cost long term (say, 50 years), are batteries generally cheaper?
LFP batteries are the cheapest and also last the longest. Race car EVs want the more energy dense NMC chemistry that was the original lithium formula. With 4 hour storage/discharge instead of smaller 1 or 2 hours, LFP batteries can last 10000 cycles which is 30 years on a daily charge/discharge cycle. A couple of years ago, this battery chemistry was $300/kwh and still cheaper than nuclear. They are now below $100/kwh, with some Chinese EVs having a free car at $300/kwh price for just the battery pack component. EVs permit a private investment to provide grid service that helps pay for EV, but at no rate payer passed down capital cost.
Batteries don't really have operating costs. Nuclear has a lot of maintenance costs especially when its time to push plants past 60 years. Diablo Canyon is spending $5B for 5 year extension. That could buy 5 times the solar power (at least more total power output over 5 years) for 30+ years instead.
Diablo Canyon is spending $5B for 5 year extension. That could buy 5 times the solar power (at least more total power output over 5 years) for 30+ years instead.
Is that 5x including battery storage? And is that 5x including degradation over 30 years?
I'm down for whatever is the cheapest way to get us off of fossil fuels over the long term. My understanding is that generally means a mix of baseload supply (nuclear, geothermal, hydro), "bursty" reveals renewables (solar, wind), and storage.
Is that 5x including battery storage? And is that 5x including degradation over 30 years?
It's 5x more without batteries. The degradation level of modern panels makes them last usefully much longer than 30 years, but it's reasonable to still just use 30 years excluding the free power past that point.
generally means a mix of baseload supply (nuclear, geothermal, hydro), “bursty” reveals renewables (solar, wind), and storage.
solar is cheapest, wind is complementary reducing battery needs. Hydro is less expensive than geothermal, and the latter is not as suited to giant power projects. Both provide the opportunity to be used as batteries pumping water uphill or heat down into the reservoir for "peaker power use" later in the day or seasonally. Solar and wind can power everything, but companies with expertise in other sectors can offer to help too. It's only nuclear that is pure corruption uselessness.
view more: next ›