That runaway datacenter power grab is the best news for net zero this century Datacenter power is a shocking business. The latest report from the International Energy Agency makes some hair-raising predictions, such as Irish datacenter electricity usage making up a third of that country's total juice budget by 2026. Globally, datacenter infrastructure is expected to more than double over the same period …
Not only would this help data centres, but it could put a substantial distributed base load to the grid, a couple at each motorway service area... waste heat can be used by the services, and the electricity can supply the grid, or local storage, or local chargers... that fat grid connection can work both ways.
This post has been deleted by its author
Even if the standard was agreed today and the first production units were being delivered tomorrow, it would take at least 10-15 years to get one installed in the UK. Planning permission and anti-nuclear protestors objecting at all stages and going to court to challenges every time they find an uncrossed "t" or undotted "i", appealing to the next court up when they fail, camping out in trees, digging tunnels and doing everything else in their power to slow down the transition to any form of green energy, which is, sadly and entirely unironically in their eyes, completely contradictory to their stated aims.
Best news for Net Zero would be no data centres.
Most of the digital offerings are surplus to the requirement, designed to kill boredom, stroke egos and make us over indulge in vanity. I mean if 99% of websites were gone tomorrow, sure the public would have some withdrawal symptoms for a couple of months and then the life would go back to normal.
There are certainly some politicians with this fault but some are still fortunately intelligent and self aware enough to only ban things that are actively harmful and enticing to the point that personal decision making is largely ineffective at preventing large amounts of harm. This is more like charging an extreme authoritarian (irregardless of political wing) with identifying and banning anything they regard as lacking value to society.
> What mechanism other than a ban would mean "99% of websites were gone tomorrow"
Oh, I don't know, Global Thermonuclear War?
Far more likely than any effective ban on datacentres!
At this point, a "political accident", "cuban missile crisis alt. ending" style, is looking a lot more likely than any civil nuclear accident
Politicians love banning stuff, apparently oblivious to its ineffectiveness as a tool of policy... but it's not the only lever available.
Tax would be one option. I'm sure some clever economists could devise an international regime of tax on profits from internet services, such that the only viable ones would be those that generate genuine value, or that someone is prepared to run on a non-profit basis.
Extreme privacy controls - legal or technological - would be another. No doubt most of those sites that the OP was referring to are the sort where the user is the product. If privacy controls were so tight they couldn't monetise their users, a lot of them probably would be gone tomorrow.
"the only viable ones would be those that generate genuine value, or that someone is prepared to run on a non-profit basis.
I came here to ask what you would define as "genuine value", then the rest of the sentence sunk in and it appears you mean those that at least make enough money to pay for running the service. I'm pretty sure generating cash flow is not the only definition of "genuine value".
Most of the digital offerings are surplus to the requirement, designed to kill boredom, stroke egos and make us over indulge in vanity. I mean if 99% of websites were gone tomorrow
Not sure about 99% of websites. Most websites occupy a couple of MB on a cPanel server somewhere and use functionally no electricity. They're not the ones doing the harm.
If we want to make a difference:
* Ban crypto from datacenters. If you want to launch a coin, you have to run it from an on-prem DC. If that's not viable then that's your problem - make it more efficient until it is. The entire notional point of cryptocurrency is to unshackle you from "the man". If you can't run it on an end-user device and nodes are all owned by firms with racks full of hardware then... meet the new banks. A lot like the old banks, except unregulated and probably Ponzi schemes.
* Ban tracking-based advertising and RTB. Loading 70+ trackers on a page load isn't free. It substantially raises the energy and bandwidth footprint of a site. Why should loading one page call 70-plus unrelated services (and servers)? NPO have shown that contextual-based advertising is just as cost-effective as tracker-based. Cutting out the middle man saw their ad revenue rise substantially, as reported in these pages. Plus, it's privacy friendly and there's less risk of leaking large databases of PII.
* Tax React. Make people buy carbon credits to use a chonky framework. No, I haven't thought out how to actually do that. I'm being silly. But again, reinventing the wheel isn't free. Standard ebooks serve a million views per month, as well as hosting their entire git and build infrastructure on a single-core 2GB VPS. A paean to the classic web And you know what? It's great. News websites serving mostly static content don't need a hefty client-side framework. Quite frankly, neither do most brand websites. For the most part, sites served via a complex K8s infrastructure are very much doing it wrong.
* Tax the buggery out of datacentres (and warehouses, factories and logistics parks) that cover less than 90% of their open roof space with solar panels. No, obviously solar won't come close to powering a dense datacentre (though it will for a warehouse that's mostly shelf space), but it's free land. It's frankly bizarre to see fields full of solar down the road from warehouses and DCs with bare roofs, or just a handful of panels in one corner (older buildings may not have the load-bearing capacity, but that's not an issue for new-builds). Equinix LD4, LD5 & LD10 could each get the equivalent of 0.9-1.1MW of solar on their roofs (a basic area calculation, rounding down generously - a fully naive calculation pops out more like 1.2-1.5MW each, I'm also ignoring their curved roofs - so don't downvote. I'm caveating the hell out of this!). Whilst 3MW won't come close to running Slough Trading Estate, it's also not nothing, and I've only looked at three of the many bit barns in the area. Cutting your power bill by 5% is non-trivial for a datacentre, and leaves overhead for the incoming grid supplies - particularly with people mithering about GPU-laden boxen increasing the required kVA per rack. Slough Estate should be a glittering array of solar from the air. If nothing else, 5-10MW would offset the total consumption of the town, if not the data centres.
And electricity of course is only half the problem. Sucking up groundwater to drive evaporative coolers (and subsequent disposal of brine) is a major sustainability issue for DCs.
Civilian, and safe, nuclear power production means Thorium. If we are to imagine a world where a datacenter is something that sits on or near a nuclear power plant, Thorium is the best bet.
Thorium reactors produce the least amount of radioactive waste, and doesn't last as long a PWRs. Additionally, you can take waste from PWRs and use it in a Thorium reactor, which has the potential to vastly reduce the amount of radioactive stuff we already have lying around.
Until we get fusion reactors (and that seems to be on its way now), Thorium is our best, safest bet.
And you don't need dozens of operators 24/7. You need one engineer with a beeper, who gets alerted when the reactor auto-shuts-down for some reason. Because the Thorium reactor cannot explode, nor melt down. When it gets too hot, the salt plug simply melts, the radioactive slurry is deposited in reservoirs and the whole reaction stops. After that, it's just a question of putting in a new salt plug, scooping the slurry back into the chamber and starting up again.
If datacenters could push for that design, that would be real progress.
> In August 2021, China announced the completion of its first experimental thorium-based nuclear reactor. Built in the middle of the Gobi Desert in the country’s north, the reactor over the next few years will undergo testing. If the experiment proves successful, Beijing plans to construct another reactor potentially capable of generating electricity for more than 100 000 homes.
> ...
> “To meet growing energy demand and achieve global climate objectives, the world is looking for alternative sustainable and reliable energy technologies. Thorium may become one of those,” concluded Clément Hill, Section Head at the IAEA. “We will continue our research to deliver credible and science-based results for those interested in working with thorium.”
https://www.iaea.org/newscenter/news/thoriums-long-term-potential-in-nuclear-energy-new-iaea-analysis
Thorium reactors may be an option in a decade, but is clearly not available now.
"Thorium reactors may be an option in a decade, but is clearly not available now."
True, but then again, if datacenters come together to spec out a small local reactor, as suggested in the article, it's anyway going to be close to a decade until they're ready to go. What we're talking about here is SMRs, of which there are a lot of talk and precisely zero commercial power-producing units. If, as seems likely, it's going to take close to a decade to get any SMR fired up and ready to go, might as well design them thorium-based from the beginning.
I'm not a fan of SMR, but ...
> Both public and private institutions are actively participating in efforts to bring SMR technology to fruition within this decade. Russia’s Akademik Lomonosov, the world’s first floating nuclear power plant that began commercial operation in May 2020, is producing energy from two 35 MW(e) SMRs. Other SMRs are under construction or in the licensing stage in Argentina, Canada, China, Russia, South Korea and the United States of America.
https://www.iaea.org/newscenter/news/what-are-small-modular-reactors-smrs
There is one SMR in operation and it appears that others are on the way.
You are assuming that the experiments prove that thorium reactors work and don't find any problems. What happens if it takes 15 or 20 years to build commercial throium reactors?
> Thorium reactors may be an option in a decade, but is clearly not available now.
For molten salt Thorium I tend to agree, they have a big issue with contamination, and the materials involved are really nasty.
The tech that 'excites' me the most though is the Accelerator-driven subcritical reactor. I wonder what happened to Aker's ADS reactors?
https://en.m.wikipedia.org/wiki/Accelerator-driven_subcritical_reactor#Rubbia_design
The Belgian MYRRHA subcritical reactor (mentioned on that webpage) is now more or less becoming reality. The first phase (a proton accelerator) is supposed to be ready in 2026. The rest however is not ready for 2036. It's a research reactor so not immediately useful on the commercial market. One of the advantages is that the proton accelerator can also be used for production of radio-isotopes for pharmaceutical use.
Although it doesn't use thorium as a source, it uses MOX (uranium-plutonium mixture) instead, which is made from nuclear waste. And it uses lead-bismuth for its cooling. The webpage seems to think that thorium is the only solution for this type of reactor.
Data centers accounting for two to four percent of total electricity usage - on the other hand data centers may have lead to decreased energy usage in others ways: e.g., less paper wastage, work from home, remote meetings instead of getting on an airplane.
By all means though, go ahead and nuke crypto, please.
There is one obvious location with easy cooling and carbon neutral geothermal generation available - Iceland.
The low outside temperatures make cooling far easier than (for example) a datacenter located in Arizona.
There is a lot of untapped geothermal energy available in Iceland due to its location on the mid Atlantic ridge.
As existing datacenters take years of planning to arrange power etc, the requirement to build a small geothermal plant for each datacenter should not add much delay to the process of building a datacenter.
A lot of datacenter activities (eg AI training) can easily tolerate the additional few milliseconds latency for the data transfer from Iceland to mainland Europe or the US.
Already being done, with circa 50% annual growth in recent years. Obviously certain constraints around geologically active zones, and reliant upon not-so-cheap underwater cables. A further consideration is whether such underwater cables are of particular interest to various well known international troublemakers.
Yes, with the obvious caveat that you don't want your datacenter to be suddenly exposed to a lava flow. We can push electrons around almost as easily as we can push photons around, so there's no need to locate datacenters right next to geothermal plants (or any other power source, really, what they need to have is transmission lines).
The low outside temperatures make cooling far easier than (for example) a datacenter located in Arizona.
Iceland has experienced temperatures high enough to require active chillers: https://www.plantmaps.com/en/is/climate/extremes/c/iceland-record-high-low-temperatures
So you'll still need to pay for that capacity, and maintenance of the units, which is much more difficult to do where the condenser coils frequently get iced-up, snowed over, etc. Unless you're Google, and can just shut off entire data centers.
There's a lot to be said for building data centers in deserts. Lots of unoccupied open land, efficient cooling with evaporation, ideal for on-site solar panels, few or no blizzards or ice storms (which have a habit of taking down electrical grids and make it difficult for personnel to come and go), etc.
More on-topic, nuclear power generation hasn't proven to be price-competitive with wind and solar, and data centers are a terribly competitive business where even slightly higher electrical rates raising their costs is very likely to cause customers to go elsewhere.
A lot of datacenter activities (eg AI training) can easily tolerate the additional few milliseconds latency for the data transfer from Iceland to mainland Europe or the US.
It's a logistical issue, though... How long does it take to ship parts? How difficult is it to get people on-site for occasional major moves/reconfiguration/etc. The added expense of trans-Atlantic rush shipping and costs for remote-hands could easily eliminate any savings from the lower electrical prices. Of course huge companies can make the logistics work, but most cannot.
My thought was: we can already build nuclear powered submarines so lets build a few, leaving out unwanted stuff like weapons and crew quarters. Moor these near one or more of the islands that we keep getting told will soon be underwater due to global warming. Build the datacentres on the beach, hook up to the submaines for power and soon they'll have all the water cooling they need as well.
Most heat pumps are reversible i.e. both "coils" can either condense or expand the gas inside them. The ones on the outside are probably still called condenser coils by some, as a leftover term from Air Con units that they resemble/are.
Freezing Fog, they certainly do not like.
"There's a lot to be said for building data centers in deserts. Lots of unoccupied open land, efficient cooling with evaporation,"
Pretty much the prime definition of a desert is no or minimal water so you may need to get truck or pipe water in and than try to come up with a closed loop, or near closed loop evaporative cooling system that doesn't corrode away too quickly.
Why is it that DCs need evaporative cooling anyway? I guess because it uses less leccy than "Heat Pumps" aka refrigerating chillers?
Not usually one to support the green nonsense, but it ought to be fairly easy to dispose of heat when there is a temperature differential in your favour i.e. the chips are hotter than the outside air, without needing to drain an aquifer
If ever there was a dangerously overambitious electric interconnector project, XLinks is it.
A few weeks ago the UK grid dropped to 49.2Hz due to the sudden trip of 1GW from a 2GW interconnector from France. That was a fairly close shave. If it dips below 47Hz then lots of other things disconnect such as solar and wind farms, and other HVDC interconnectors.
Our power system is designed to tolerate a trip of around 2GW at most. Having a 10GW unit of any type is asking for trouble, never mind an HVDC interconnector with no inertia, prone to trip at a sneeze.
And with 4000km of undersea cable to go kaputski, Putin and chums will be laughing their socks off if this ever gets built. Luckily it probably won't, because the price of copper has rocketed since it was initially dreamt up.
Well, the only "clean" reserve with near-instant startups are Dinorwig, Cruachan and a couple of other minor hydro facilities. Those combined, IF reserved in full (never - the owners want to use their facilities not hold them in reserve indefinitely and not making money!) maybe buy about half of what Xlinks is meant to be capable of at full tilt.
4GW is quite a lot of rolling reserve to have to keep ticking over in parallel to cover for the "next fault". Perhaps if combined with an absolute shedload of fast response storage; batteries & maybe the micro hydro setups (one of the latter is being built in Dorset as we speak) can be made to fly.
Xlinks were intelligent enough to split their connection request in two; with two landing points UK side; this does bring the load into a levels that those hydro facilities were meant to cover. But loss of the lot (say, due to an event on Morocco end) is still a risk.
On a similar front, the decision of Port Talbot to ditch the blast furnaces and (maybe) install an Arc Furnace instead causes a bunch of issues for capacity and unpredictable loading. The ones in Sheffield are already a pain - and to be honest - so lightly used I don't for one moment believe Tata's CBA on installing an electric furnace adds up to adding another one.
It's almost as though we would benefit from a central authority thinking about all aspects of the energy sector, and it's interface with industry. With the engineering know-how and personnel to build and operate stuff. You know, a board of the central electricity generating variety?
But what do I know. We have senior personnel in BEIS that don't even know what Calorific Value means making these decisions. Yes, that's the standard of Civil Service and Government appointment we have.
>>At $16 billion, it's still less than 1/3 of the cost of the new nuclear reactor at Hinckley Point C.
True dat.
>>At least French tax payers are footing most of that bill though.
Poor sweet child. The French tax payers are acting as garuantors to the loans EDF required to build1 the machine... and are direct beneficiaries of the UK power consumer's generosity in years to come!
See Private Eye's 'Keeping the light on' columns passim and, no doubt, ad infinitum on all the shenanigans that Hinkley C is associated with, not least the base cost per unit being, IIRC, about 4x the existing base cost per unit...
1for special values of build in time t for which t may be apparently infinitely variable
So as I said the French tax payers are footing the bill of circa 3 x the original estimated cost vs the original and locked in agreed price of electricity for 40 years. (not adjusting for inflation - coz the power price is also adjusted for inflation.). And those interest costs are stacking up too as it's not going live before at least the 2030s.
So as I said the French tax payers are footing the bill of circa 3 x the original estimated cost vs the original and locked in agreed price of electricity for 40 years. (not adjusting for inflation - coz the power price is also adjusted for inflation.).
I rather suspect EDF will find ways to extract the money from UK tax payers instead, and our politicians will let them. Or what I suspect is happening is EDF saw the news from Orsted and Vattenfall about rebidding their CfDs and decided to try the same trick and screw us for more money. Hopefully the NAO is crawling all over the books to find out just how EDF managed to screw this up when their original bid was already very... generous in comparison to other reactor builds.
What is really needed though is a wholesale reform of our energy markets. I was reading about 'round tripping' on our interconnectors where we for some reason import and export pretty much the same amount of power. This may be to balance the system, or just generate profits when demand, and thus prices are high. That sounds a lot like the refiling tricks telcos use to game international settlement rates.
>>I rather suspect EDF will find ways to extract the money from UK tax payers instead, and our politicians will let them.
This. The UK Tax payer is already footing the bill, and will continue to do so with the double whammy (should they use on-grid electricity) of increased base unit costs affecting the price (upwards) of all the electricity they use.
The whole point of a seemingly at the time stratospheric price for the electricity from Hinckley Point C was that all the risk was on EDF.
And yes they got bitten and not just by the UK one, but by the new reactor in France too. Hence why their share price is like a black ski run. Well that and tens of billions in future unfunded reactor decommissioning costs. As above though, it's ultimately the French tax payer footing all these bills.
Other way round. EDF pulled the trick and now the wind farm operators are doing the same "because they can".
No competition means cartels and monopolies for things you need will take the pish.
Low and behold, once again, the case for state ownership is incredibly strong. Why do consumers tolerate all these middlemen creaming off?
/me Can't decide if you are trolling or just don't understand how these things work in practise.
>>I said the French tax payers are footing the bill
No, becasue EDF are borrowing commercially. They aren't loans from the French Government (who only own 84% of EDF anyway...).
>>And those interest costs are stacking up too as it's not going live before at least the 2030s.
All of which will be repaid by the UK consumer... or tax payer, depending on what extras EDF can add to the contract or persuade the Govt. to add as subsidy to the scheme (as I said, have a look at "Keeping The Lights On" in back issues of Private Eye).
Agree. The potential downtime for failures on that system is immense. Much smaller systems in more benign offshore environments are still very hard to look after. The Atlantic coast is particularly unforgiving compared to North Sea.
Even if Xlinks does land in the UK, you still have to get the power from that landing point(s) to where it's actually needed. TEC register would suggest they are talking about landing in the South West and South Wales. That'll need more onshore reinforcement.
Cue planning permission hell.
A/C for what should be obvious reasons.
As Whitehall squashes any attempt to build anything other than huge mega-reactors, which will never be finished.
I suspect this is a mixture of leftover-hippie "Nuclear power? No thanks!", and an attempt to stall all other energy sources until "we have enough wind power never to need it".
Not only have Rolls-Royce effectively been told to do one, but a proposal to make use of Sellafield waste by a new company was dismissed. They are now going to build in France.
The ONR (UK office for nuclear regulation) is staffed by a lot of CND types, as well as busybodies and jobsworths who love making work for themselves and other people.
This is a symbiotic relationship as the CND anti-nuke types can use the busybodies to push up the cost of anything 'nuclear' to astronomical proportions, while the busybodies use the CND types to justify their own miserable existence.
Meanwhile there are few people left at the ONR who actually have a clue about nuclear engineering. I refer you to Page 39 of the latest Private Eye (#1615) for a column by 'Old Sparky' on the current exodus of staff from the ONR.
This is why Sizewell and Hinkley are both billions over budget and a decade late - because they are both on the hook for thousands of engineering changes ordered at the last minute by the ONR, which itself has no accountability against being obstructive
So I agree with Mr Semicolon's sentiment. Also I think you needed to explain the meaning of "Do one" in the modern British vernacular to our left-pondian friends.
It's not just hippies that don't want Nuke in their back yard.
Your average Tory doesn't even want a Pylon. See Priti Vacant joining the Just Stop Pylons protests in the East Anglia despite it being an utterly obvious pre-req for Sizewell C, the interconnectors and various offshore generators connecting in that area.
I hate to remind people, but "Net Zero by whenever" is in fact, a government policy despite the best efforts of the government and it's media bosses to sabotage and lambast the execution of it's own policies.
Small reactors have all the same planning problems as large ones and so, if you're going to solve the planning problem pissing around installing a couple of hundred MW makes no sense whatsoever. Go big or go home, so to speak.
Time and again just about the only thing the British public WILL tolerate is a little gas turbine plant; which, I'll grant, tend to be very quiet and out-of-sight. With the minor problems that they are entirely at the mercy of the global supply chain and totally incompatible with net zero.
...people fear nuclear. They fear it to the point of irrational panic.
It's become one of those things about which it's basically impossible to have a reasoned discourse because so very few people actually properly understand it. I was at least taught physics at highschool by a qualified nuclear physicist who had worked on nuclear power plants in the 80s, so I like to think I have some grasp on the complexities of the issue. The reality is tho, that I think nuclear power is good because someone I have reason to believe understands the subject told me it was. I don't really understand it myself.
A lot of people got told it's really, really bad, and at this point that's going to be a really hard notion to dislodge because they don't understand it any better than I do.
Nuclear power is regulated to the moon and back - which is probably for the best, given the potential for it to do harm if we cock it up - but given how many people are just plain scared of it I rather suspect that getting regulations relaxed enough to allow any kind of small modular power plant to actually be built is going to require a great deal more than a bunch of large companies agreeing that they'd like one and setting out a standard for what it should look like.
I have no idea how we get enough of those people on board that the issue isn't so politically toxic that no one with the power to do anything wants to touch it.
And the flip side of that coin is not fearing fossil fuels enough. I mean, even aside from the CO₂ & climate change angle, just the dirty stuff from coal is killing millions on a regular basis when it's working as planned, but oh, a few people died from radioactivity that one time, so we can't have that! That's what really sticks in my craw.
Humans are not good at estimating risk - and we're really bad at the idea of composite vs absolute risk.
There's something very immediate about "If this blows up it will kill everyone in a half-mile radius"* it's something people can understand. Something has exploded or it hasn't. People have died or they haven't. True or false. It's absolute.
Trying to get people to wrap their head around the idea that "Statistically, running this thing for 40 years will lead to everyone in a 20 mile radius having a 4% higher risk of cancer and a 6% increased incidence of respiratory diseases, which on average will amount to an additional 100 deaths per year over that time frame"... that's just too abstract. People don't really engage with that in the same way.
*...and yes, I know Nuclear power plants don't do that as a rule, but for the sake of this example it's a thing that people are afraid that they will.
*...and yes, I know Nuclear power plants don't do that as a rule, but for the sake of this example it's a thing that people are afraid that they will.
But non-nuclear industrial plants can and do. Yet we don't have the same level of hysteria about those.
Even a Carbon Capture and Storage plant has the potential to asphyxiate everyone in a half-mile radius if it goes horribly wrong in calm weather.
The specific hysteria about Nuclear is so stark that i think it has to have been pushed by a combination of CND types and Big Oil in the background. That and The Simpsons hasn't helped.
And the flip side of that coin is not fearing fossil fuels enough.
Don't they? Have you seen the popular opposition that appears when the construction of a new power plant is announced in an area? Maybe people are just bad at appreciating far-away threats?
a few people died from radioactivity that one time, so we can't have that!
Nuclear power plants are a different type of risk, entirely. Unless you're directly downhill or downstream of a massive fly ash tailings pond, no form of disaster at a coal power plant is going to render your property completely worthless and uninhabitable for several generations to come.
On a country-wide scale, nuclear power is a better proposition than coal. But on an individual scale, nobody wants to be the one to take the risk to their own safety and assets.
...people fear nuclear. They fear it to the point of irrational panic.
Yup. As I kid I got to walk around an NPP and was disappointed that I didn't develop super powers. Or the control room didn't show temperatures of thousands of degrees in the reactor core. A little knowledge went a long way. Then as a bigger kid, submarines. Their reactors occupy a relatively small volume inside the vessels, and are cared for by a small percentage of the already small crew, and through the time naval nuclear reactors have been in service, they seem remarkably safe.
So I'm a big fan of SMRs, just there's a lot of political challenges around who can own and operate those. And as you say, overcoming decades of anti-nuclear FUD from the Greens.. Who are the real problem. Especially given this comment in the article-
What none of them are is proportional to the threat of climate change, a threat which you may have noted is universal and terrifying. It's not something that companies can fix through competition.
Which is the same problem. People have developed an irrational fear of the climate, which has developed into huge business opportunities for the Green Blob. Who then campaign against nuclear and promote windmills instead, and windmills are pretty useless for powering datacentres due to cost and intermittency. But it's also a toxic issue for politicians, who are hell bent on pursuing 'Net Zero', even when the data and advice they're relying on is garbage. So for example this-
https://dailysceptic.org/2024/01/24/revelation-that-u-k-climate-target-is-based-on-one-windy-years-data-threatens-to-unravel-net-zero-credibility/
The Royal Society analysed decades of local wind speeds and found the electricity system needed the equivalent of at least a third of green energy to be stored as backup. Such a cost would be astronomical. Now it appears that the Government’s Climate Change Committee (CCC) fudged the issue by using just one year of high wind data in persuading Members of Parliament in 2019 to donkey-nod through Theresa May’s insane legislative rush to Net Zero by 2050.
Nuclear of course isn't reliant on the weather.
Perhaps you could read the actual Royal Society report rather than Toby Young's view.
I have only read the Executive Summary which includes this paragraph.
> In 2050 Great Britain’s demand for electricity could be met by wind and solar energy supported by large-scale storage.
https://royalsociety.org/-/media/policy/projects/large-scale-electricity-storage/Large-scale-electricity-storage-report.pdf
Some context about the "The Daily Sceptic".
> The Daily Sceptic is a blog created by British commentator Toby Young. It has published misinformation about COVID-19 vaccines[9] and climate change denial.[16]
https://en.wikipedia.org/wiki/The_Daily_Sceptic
https://www.desmog.com/toby-young/
Perhaps you could read the actual Royal Society report rather than Toby Young's view.
Perhaps I already have? Perhaps the problem is still that the 'CCC' cherry-picked a single year with above average winds to project their sponsor's profits far into the future? Perhaps if you actually read the RS report, you'd note that it points out low wind conditions are common, thus the storage requirements, and costs are far, far higher than the CCC's projections?
https://www.desmog.com/toby-young/
Perhaps you don't realise Hoggan is a PR slimeball who's made a career out of PR for both the fossil fuel industry, and the 'Green Blob'?
https://en.wikipedia.org/wiki/James_Hoggan_(public_relations_expert)
Note no mention of his PR work for fossile fuel companies. If you want to learn more about how this garbage works, perhaps read up on David Suzuki and his role in PR as well? From fruit-fly botherer to climate 'expert'.
Again you make the point though. People who believe in anti-nuclear FUD and Thermageddon are often fuelled by the same popular delusions, which is not science..
" In 2050 Great Britain’s demand for electricity could be met by wind and solar energy supported by large-scale storage."
Not at any reasonable cost through. Solar is a very poor resource for the UK, with pathetic capacity factors, and a profile that maximises output at times of lowest energy demand, if the RS think it's a useful resource, it's because they don't understand the UK energy system.
Wind, especially offshore, is a resource we could make more of. but large scale storage is problematic. The essence of storage economics is that to be cheap you need frequent (paid) utilisation. Smoothing out minor diurnal and regional variations in wind power can be done largely by the appropriate mix of sites across the UK. Which means any storage is used infrequently. To cover a big winter anticyclone, we'd need storage capable of delivering 90% of wind capacity for around five days each year. Storage would have some use outside of the extreme low power event (<10% of installed capacity available), but if you'd like to estimate the cost of a battery to power around 60-70% of total UK demand (inc heating and transport, and allowing Hinkley and Sizewell) for five days, you'll have a very big number - of the order of 10 TWh of storage. Divide the cost of that by the actual power output, say for a total annual run time of 15 days, and you'll have a cost per MWh that I can tell you now will be unaffordable.
Worse still, that five contiguous days is a normal annual event. If you look across a twenty year horizon, then <10% of capacity wind events can last for ten days or so.
if you'd like to estimate the cost of a battery to power around 60-70% of total UK demand (inc heating and transport, and allowing Hinkley and Sizewell) for five days, you'll have a very big number - of the order of 10 TWh of storage. Divide the cost of that by the actual power output, say for a total annual run time of 15 days, and you'll have a cost per MWh that I can tell you now will be unaffordable.
So? Just fire up gas power stations for 5 or 10 days per year as required.
Sure, the electricity from these would be very expensive, compared to running them all the time. But as we only talking about (say) 5% of total annual demand coming from gas, having that 5% be really expensive is spread over the other 95% of usage.
Forget "net zero". If we were able to reduce CO2 emissions to 5% of 1970 levels, we'd all be fine. Just make good use of that 5%. As a side benefit, we'd also make our existing fossil reserves last 20 times longer.
(However, note that today only about 1/6th of our energy consumption is in the form of electricity, so there's going to have to be a big shift from fossil to electric for heating and transport too. Even with heat pumps, this will most likely treble both our generation and transmission requirements)
"Just fire up gas power stations for 5 or 10 days per year as required."
Even the fastest gas power stations can't just go from cold to full power in a few minutes, they are kept on standby to be ready to go at a few minutes notice. This also consumes (very little) power, but more importantly needs maintenance and taking care of. If our local weather forecasting abilities improve so much that we can confidently predict a 5-day windless spell a month in advance, this would be a good option..... except that doing so you would still need your standby gas stations to cover a very large percentage of the whole of the country's power needs (IIRC a study that I once saw, close to 80%). Of course, having large-scale interconnectors over a wider area means that when there isn't any wind in UK, you can buy it from Greece or Norway or Poland, but even that generally means much-larger-than-usually-necessary wind/solar farms and interconnector infrastructure pretty much everywhere.
Peak 'what aboutery'. De carbonizing our economy will be difficult,, but not impossible. Yes you can't just fire up gas/steam powerplant in minutes. But then the wind doesn't go from blowing to not blowing in minutes. But you can spool up a gas turbine. You can also arrange your grid so that you stop doing somethings when the wind doesn't blow.
If the Register commentariat had been around in the 1800s they would have railed against the impossibility of transitioning lampposts from gas to electricity, told us that infernal combustion engines could never replace horses, and pointed out that no ship could sail the Atlantic under its own power because there were no coal mines in the middle of the ocean.
If the Register commentariat had been around in the 1800s they would have railed against the impossibility of transitioning lampposts from gas to electricity, told us that infernal combustion engines could never replace horses, and pointed out that no ship could sail the Atlantic under its own power because there were no coal mines in the middle of the ocean.
No, they'd probably be telling us we should stick with windmills and sailing ships. We don't need those new fangled steam ships, and electrickery is the work of the devil!
It's sad, but the Greens prove the saying that those that ignore history are condemned to repeat it. We abandoned windmills once before because better technology became available, now they've been dusted off and given a new spin.
Equally we used to put lead in petrol and CFCs in refrigerators, and we stopped doing that because we realized that the environmental costs were disproportionately high. Given the pace of technological change over the last 200 years the rea strange thing is any assumption that the fossil fueled world we inhabit today is in any way the only answer.
You might be right about the 'green' response in 1800, but I stand by the theory that the readers of Ye Olde Reg (1/2 a farthing, published weekly) would have argued against change for green reasons. We much prefer the 'it will never work' , 'its no use because [add weird and highly personal circumstance]' and 'I would never make that mistake' arguments. Not forgetting the old 'I never read these trashy broadsheets and 'I never gossip in coffeehouses', while still being all too happy to read, comment and gossip on everything and anything.
Those 5 to 10 days would be predictable from a week or so in advance and pretty precise demand 72 hours out with current met forecasts. Anticyclones and cloud don't just appear out of the blue. Given our dash for gas we could have a large legacy fleet to call on without having to build more plant. If only we could fire them with green hydrogen created when we have excess wind and solar to spare.
Quote
"...people fear nuclear. They fear it to the point of irrational panic."
To the point the word 'nuclear' had to come out of NMRI scanners because the public were scared of the things.... now they go 'meh' and complain about having to lay in the tube for 20 mins.
We had a science teacher at skewl explain about the safe shielding needed for plutonium, basically a crisp packet that can be sealed to stop the plutonium dust escaping, if you want to go over the top then the foil/plastic bag your phone came in would do, big fat alpha particles cant make it through 4 inches of air let alone a plastic bag.
It was from him and his geiger counter we learned about how many normal materials put out radioactivity..... its not just granite table tops you should be worried about....
It was from him and his geiger counter we learned about how many normal materials put out radioactivity..... its not just granite table tops you should be worried about....
If people wanted to avoid harmful radiation, they should live inside a reactor hall. People would then be protected from stuff like this-
https://en.wikipedia.org/wiki/Oh-My-God_particle
The Oh-My-God particle had 10^20 (100 quintillion) times the photon energy of visible light, equivalent to a 142-gram (5 oz) baseball travelling at about 28 m/s (100 km/h; 63 mph). Its energy was 20 million times greater than the highest photon energy measured in electromagnetic radiation emitted by an extragalactic object, the blazar Markarian 501.
I also think one of the challenges with understanding radiation and risks are the number of types, and units involved with radiation. Plus not necessarily fully understanding the effects, ie people living in Cornwall or Colorado are routinely exposed to 'dangerous' levels of natural radiation that would be banned in a workplace.
People fear human greed.
Silicon Valley remembers all the chip fab toxins illegally dumped. There are still extraction machines and dead zones here and there. It wouldn't have cost much to recycle those chemicals but it there was money saved dumping it.
Repeat convicted felon and annihilator of cities, PG&E, has new permission to raise rates to maintain investor profits during equipment safety upgrades. That money really is super-honest going to fix infrastructure problems this time, unlike the last few decades where maintenance records were falsified and the money was pocketed.
Witness Buncefield (UK, 2005), where a data centre situated less than 500 m from a bulk fuel storage depot was completely gutted when an explosion occurred at the depot.
>It's the kind of low-risk, low-cost, environmentally positive move that governments can get behind
No, they can't. Not in the current political climate. The environmentalists are mostly anti-nuke to the point where it overrides any other environmental concern, and the populists are mostly pro-fossils and never gave a crap about environmental concerns to begin with. These factions will bleed votes off the opposition moderates if they support you on this, and the opposition moderates know it, so they won't support you either - this is true regardless of whether you're left or right. And you can't do anything with just half the moderates and nobody else.
I'm not saying this dynamic is everywhere, but it's in a lot of countries right now. Enough that I doubt an international consensus can be easily reached. And, of course, if anything at all goes even slightly bad, then anyone whose name is on this is never getting elected to anything ever again, and not even saving the world from carbon will save them.
It might work, but, politically, it's very far from low-risk.
Think you're missing another part of a bigger picture. The environmentalists that oppose nuclear just are not that big of a group, and while the fossil lobbying is loud and controls a mass of stupid that go nuts at anything, it's still a matter of good old fashioned capitalism. Nuclear is too expensive and doesn't shunt the appropriate amount of excessive profits over a short enough time for it to be worth it. All of which are stupid reasons, but the last one is a big issue. Either find a company that will be satisfied with a reliable return on investment in the long term and keeping a payroll of well paid, highly educated and skilled workers (good luck) or it will have to be a government funded/backed approach (and then you get to deal with the group of which the fossil fuel people are a subset : government can't do anything well, so they shouldn't and if a private company won't do it, it shouldn't be done...obviously).
Sure, but my impression was that the point of the article is that standardized nuclear could be the answer to the "too expensive" problem, that everything else is just politics, and the author felt optimistic about that. To which I was pointing out that politics is not at all an easy or minor problem, and I feel pessimistic about that.
It's true that the permanently rabid are not very large groups, but it's also unfortunately true that it's easy to get poorly-informed moderates temporarily rabid on this topic. Just spread FUD about the risk and/or costs, or outright lie if you have to, and you can get moderate voters pissed at their parties, maybe nab a bunch of them, even. On the other hand, there's exactly zero a moderate party can do to get any of the extremists on board with this plan. So there's little incentive to get cross-party cooperation.
"To which I was pointing out that politics is not at all an easy or minor problem, and I feel pessimistic about that."
And, so far, no one has even mentioned "TERRORISTS!!!!" yet :-)
Lot's of nuclear material scattered in small units all over the country/world. Much harder to keep secure than a few large, well fenced off sites. There will be "DIRTY BOMBS" going of all over the place, mark my word!!!!
Hence put the reactors offshore, where none of the moderates have any concern whatsoever. The UK and Northern Europe are uniquely placed with massive amounts of relatively shallow open water readily available and established industries for building platforms on them.
The increased cost of the platform and connection is massively offset by the enormous reduction in red tape for putting your nuke at an onshore location.
Getting a bunch of datacenter operators to agree on a specification would be nice, but hardly the end of our travails. The next step would be to get one or more countries to accept it without messing with the spec. too much, and for them to continue to do so into the future without making unnecessary regulation changes. Nuclear power plants get regulation changes (or unrelated design changes) all the time to the point where costs don't decrease for the next-in-line power plant from experience, it keeps on costing about the same per megawatt delivered as the first. Which is one of the biggest reasons why nuclear energy is about the most expensive type presently in use and why there's only a trickle of new nuclear power plants coming online, mostly in autocracies or near-autocracies.
Getting a bunch of datacenter operators to agree on a specification would be nice, but hardly the end of our travails. The next step would be to get one or more countries to accept it without messing with the spec. too much, and for them to continue to do so into the future without making unnecessary regulation changes.
That's really the point of SMRs. Rather than having to go through very expensive approvals processes every time, you just order a COTS nuclear reactor from the factory. Datacentre operators don't really need to be involved in the process because there is already a specification. Like they just want affordable, reliable MW. The SMR industry is a bit different, ie RR's SMRs are relatively high power compared to others, but the theory goes once those designs are standardised and approved, you can just order SMR NPPs in 500MW, 250MW or <whatever> units. Then run in multiples, if you need more power.
Which is one of the biggest reasons why nuclear energy is about the most expensive type presently in use and why there's only a trickle of new nuclear power plants coming online, mostly in autocracies or near-autocracies.
Nuclear is much cheaper per MW or MWh than 'renewables' given it's relative simplicity, and it doesn't need as much transmission infrastructure, or the real biggie, backup power. One of the biggest con tricks the 'renewable' industry, and their pet lobbyists at DECC or the CCC pulled off is the 'levelised costs' scam. By excluding trasmission, back-up and/or storage costs, 'renewables' is made to look cheaper than it actually is. The other con is of course the way the 'renewables' scumbags lie is by telling us costs are falling, which they're not and are increasing.. Which should be obvious to anyone who's looked at their electricity bill and noticed it's been going up rapidly rather than falling, as promised. That then increases energy poverty and inflation. Tobacco duty gets massive increases, that feeds into CPI or RPI figures, which are then used to inflate 'renewables' indexed supply contracts.. Which then has other interesting effects, like in Germany, Europes biggest condom producer closed down due to high energy costs.
This is all part of the deindustrialisation the 'renewables' scumbags have delivered the UK and Europe.
As for autocracies, they have the political advantage of JFDI. Rosatom has a lengthy order book from nations that don't believe the Green garbage. Westinghouse's designs are also doing well, and we used to own them before G.Brown Esq decided to flog off most of the UK's nuclear industry to help prop up EDF, which just happened to employ his brother. And strangely, China licensed EDF's reactor design, and has those operational far faster and far cheaper than EDF can manage. But then we're still helping to bail out EDF.
"Nuclear is much cheaper per MW or MWh than 'renewables' given it's relative simplicity"
No, that's simply incorrect. In rough terms, comparing the strike prices under the UK CfD arrangements, renewables are roughly half the price of Hinkley C per MWh. To put some numbers on that, Hinkley C's strike price as at September last year was £128/MWh. Last year's CfD for onshore wind was £53/MWh. The strike price for this year's CfD round will be in the £63/74 range (the higher figure for offshore), but equally Hinkley's strike price will go up by around 4%, to £133/MWh.
It is most certainly not simple, even so I have no problem with nuclear power, it's a good and reliable power source. Unfortunately it's inordinately expensive and painfully slow to construct. By committing to nuclear baseload at this outlandish price, politicians are trying to bankrupt the country.
To put some numbers on that, Hinkley C's strike price as at September last year was £128/MWh. Last year's CfD for onshore wind was £53/MWh. The strike price for this year's CfD round will be in the £63/74 range (the higher figure for offshore), but equally Hinkley's strike price will go up by around 4%, to £133/MWh
I think you're confused-
https://cfd.lowcarboncontracts.uk/cfd-register/register/NUC-HPC-198/
which is perhaps not suprising given CfD's are often quoted in 'old money'. Hinkley's strike price may actually reduce given their contract is discounted if Sizewell gets the go-ahead. The next auction round has been characterised by special pleading given news like this-
https://www.reuters.com/business/energy/orsted-cease-development-some-us-offshore-wind-projects-2023-10-31/
Nov 1 (Reuters) - Renewable energy firm Orsted (ORSTED.CO), opens new tab on Wednesday halted the development of two U.S. offshore wind projects and said related impairments had surged above $5 billion, as the industry grapples with supply chain delays and higher costs.
And much the same with their UK subsidy farms. And then-
https://orsted.co.uk/media/newsroom/news/2023/12/orsted-takes-final-investment-decision-on-hornsea-3-offshore-wind-farm
In July 2022, Ørsted was awarded a contract for difference (CfD) for Hornsea 3 at an inflation-indexed strike price of GBP 37.35 per MWh in 2012 prices. The CfD framework permits a reduction of the awarded CfD capacity. Ørsted will use this flexibility to submit a share of Hornsea 3’s capacity into the UK’s upcoming allocation round 6.
Which illustrates why our useless 'leaders' shouldn't be allowed to write contracts. The 'renewables' scumbags made a lot of press about the initial strike price, then Orsted discovered it can't actually deliver at that price, so wants to rebid in the upcoming rounds where offfshore is more likely to be >£100/MWh. Difference is the US contracts include penalties, the UK just allows price gouging. Again it's also not a like-for-like comparison because the costs of dealing with wind farmer's unreliability and intermittency are not included in those strike prices. Add the costs of batteries and delivery, and wind is far more expensive.
Again this is why our energy bills have been rising, not falling. The more 'renewables' we add to the system, the more expensive and unreliable our electricity supply becomes.
"to the point where costs don't decrease for the next-in-line power plant from experience, it keeps on costing about the same per megawatt delivered as the first"
Well, the EPR's being built in Europe seem to be getting more expensive over time, so the cost per MWh goes up rather than staying constant. EDF recently said Hinkley C will be completed around 2030, at a cost of about £34bn. Unfortunately, in the Wanky World of Energy Policy (tm), Hinkley costs are always quoted in 2015 prices, so in 2023 prices it's ***already £45bn***, and likely to have an out turn in 2030 of around £52bn (assuming no new and notable downsides between now and then).
Whilst any sane person expected it all to go wrong, I wonder if they'd have gone ahead if government hadn't been such f***wits? The original construction estimate for Hinkley C was £18bn, and that give or take a year's indexing is what has now become £34bn.
Where can I buy one of these really cheap nuclear power stations? I looked on Amazon and they just have books and mislabeled batteries. Ali-Express would only sell them in packs of 500. I thought I found one on Etsy but all they sent was an embroidered cushion cover that doesn't even glow in the dark.
Thanks to the shills for the commentless downvotes, I'll take it as a badge of honour.
I welcome reasoned arguments to the following points:
1) Waste. Funnily enough, no article or comment advocating nuclear mentions the elephant (foot) in the room. Presumably companies will expect the taxpayer to take care of it, as usual.
2) Cost. Nuclear power at scale is more expensive than other methods. I fail to see how small scale plants will magically be cheaper per GWh.
In the very different context of a submarine:
I) The tactical advantage is worth it
II) The military doesn't give a shit about cost.
3) Security. Existing large scale plants have unsurprisingly high security requirements, including their own police force (CNC). How will that be taken care of? Perhaps companies are hoping the state will deal with that too, thus pushing the true cost even higher.
Waste: Bury it at sea. That's what happens to the core of a sunken nuclear sub, isn't it? Nobody seems to care about that.
Cost: This is driven by an attempt to appease political opposition. It doesn't need to be anywhere near as expensive as it is.
Smaller means lower cost because risk is cost, time is cost. SMRs are less risky and can be built faster. Once they are built, public opinion will change, and the cost and risk of even larger plants, or farms of multiple SMRs will come down. It's also much cheaper to fabricate multiple small reactor vessels than one really big one. Safer to operate too.
The CNC are useless. They are nothing but a retirement home for corrupt, abusive, lazy and otherwise unfit-for-duty policemen.
Reactors are thankfully not dumped at sea at EOL. The RN's approach to date has been to just moor the dead boats indefinitely - which is not a solution as they are running out of moorings. There is an expansion underway at an RN facility with more dedicated decommissioning capability.
You might not be aware, but the US provides funding to Russian decommissioning activities; overseeing and enabling the old reactors from cold war boats to be cut out and safely stowed away well away from anyone on land.
On cost, the reactor is not the cost issue, and it never has been. All the ancillaries around the site you need are where the cost is... And good luck with the planning (zoning) permissions. Anyone that thinks they risk will losing the next election if they say YES is never going to approve.
I maintain that nukes on offshore platforms are a reasonable compromise between the paranoid safety perception and ability to get stuff built relatively cheaply and inexpensively. Nobody will meaningfully complain about a nuke plant on a platform 150km offshore.
The reactor is never the problem in the context of any reactor. Durham university had (has?) a small reactor in the late 1990's. The problems arise in security, civil engineering, and generic NIMBYism far outweigh any engineering considerations of the reactor itself. To the point that they dominate overall cost of ownership.
Don't rule out decommissioning cost. A system designed from the outset to permit decommissioning won't be as bad as the 1960's designs now hitting EOL. Of course, those reactors, the nuke operators were meant to put away a bit of cash every year for that future need. Successive raids on public finances stole those funds and now consumers are "paying twice" for the job (probably more than that, in fact.
Insightful couple of posts, cheers.
Well, when I said sunken I meant sunk by the enemy or by accident. Are there not quite a few of those buried at sea? Can't we reasonably expect some to occur in future?
On cost, I thought that the EPR vessel and inner containment were a significant cost factor, especially when they had to be re-made following minute cracks being found. My point is that it's easier to make a small vessel metallurgically perfect than a large one.
But I agree, the ancillary stuff probably does dwarf the technical cost. That is a political and regulatory issue though.. I like the idea of using offshore platforms, although I worry it would just be bait for russian saboteurs.
"Office of Nuclear Energy: NRC Certifies First U.S. Small Modular Reactor Design"
“The U.S. Nuclear Regulatory Commission (NRC) issued its final rule in the Federal Register to certify NuScale Power’s small modular reactor.
"The company’s power module becomes the first SMR design certified by the NRC and just the seventh reactor design cleared for use in the United States.
"The rule takes effects February 21, 2023 and equips the nation with a new clean power source to help drive down emissions across the country..."
“...Each power module leverages natural processes, such as convection and gravity, to passively cool the reactor without additional water, power, or even operator action....”
https://www.energy.gov/ne/articles/nrc-certifies-first-us-small-modular-reactor-design
Read it here.
This article is data center simp approved. Would you like to tell us about how the next gen of hardware is so much more efficient and imply that it will decrease energy use? So, Yes, obviously a nuclear reactor means data centers are no longer a massive problem. They will be free to build as many of them as they want to help collect even more personal data and help write more code for useless apps that just collect more data and are of basically no real benefit to anyone. Power is the easy part. Stupid fears and stupid planning/construction the justify stupid fears are somewhat more difficult. Seems like there is some other component you totally ignore....hmmm....what is it, what is it.... Phew, getting hot in here with the 4090 going full on, one sec... Sorry about that, had to redirect human drinking water to cool this...right, cooling! That's it! I'm sure single phase cooling will be just as safe and drawback free as consuming entire rivers or unleashing floods of all the fun two phase chemicals. So yes, just about the power. Also the small issue of no one wanting to pay for nuclear reactors because it doesn't make the profit fast enough, requires skilled workers that no longer exist and will take a stupid amount of time to make new ones, etc... etc... etc...
As Jeavons noted in his book on the Steam Engine and usage of coal : improved efficiency does not decrease total demand. Improved efficiency means that there will be more users of said technology; and the net effect of that is increased demand even if the per-user quantity is reduced.
Jeavons extrapolated out the known coal reserves of his era to predict the downfall of the British Empire. While this did not know about Oil or the two world wars to come, it was on the whole, surprisingly accurate.
In datacentre land, there MUST be a point where there are diminishing returns on collecting more data (probably already reached in fact). What MORE does Google need to catalogue about your preferences to target it's ads even more (in)effectively?
Sure, I bought a new bike 3 years ago for short commutes. Spamming me with adverts for luminous spandex is not going to be successful. I drive a quick car as well but I don't need a fireproof racing suit and helmet to drive to B&Q.
Quantity seems to be in vogue rather than quality.
> What MORE does Google need to catalogue about your preferences to target it's ads even more (in)effectively?
The scope will creep (has crept) from targeting ads to targeting You.
It's not enough to build a detailed model of each and every person. The goal is model-feedback control.
"So imagine if the world's datacenter industry got together – yes, a fantasy, but these are needful items - and issued a specification for a standard, modular, small nuclear power plant." That's not a promising outlook. Take the Rolls Royce PWR3, an adaptation of GE's S9G. Even those small variations in specification have made the project five years late, mostly because of delays in construction of test facilities. Moreover, it's not just a specification which is required. Most of the infrastructure for the development, test and maintenance of these plants is at classified facilities which are unable to accept commercial tasks. So the datacentre industry would also be paying for a building and staffing programme for all of those facilities. The small design means that there is only limited opportunity for re-use of test facilities designed for larger commercial plants.
All this delivery risk is occurring when solutions to power generation for datacentres are needed within a short timeframe.
If latency is unimportant -- as the author claims but I doubt -- then there's a far simpler solution: build a global ring of datacentres to operate at times of local peak PV solar.
Moreover, it's not just a specification which is required. Most of the infrastructure for the development, test and maintenance of these plants is at classified facilities which are unable to accept commercial tasks.
I guess this is something that could be changed. Prior to SMRs becoming more mainstream, I had fun wondering why naval reactors couldn't be modified for civil power. But the usual answer to any questions about naval reactors was 'It's classified'. Now we're in the age of YT though, there's video of reactors being removed from decomissioned Russian subs, or people working around the reactors.. and they're tiny.
If latency is unimportant -- as the author claims but I doubt -- then there's a far simpler solution: build a global ring of datacentres to operate at times of local peak PV solar.
Latency can be very important, ie if you want to do synchronous replication for high availability designs, there's a limit to acceptable latency that then translates into maximum distance between datacentres. The bigger issue is one we've known about for centuries, which is the cost. When the wind doesn't blow, the miller sits idle, the baker doesn't get their flour and the bread doesn't get baked. Now, the 'renewables' industry have polished that turd in re-imagining windmills as 'modern' technology, even though wind turbines have been around for well over 100yrs. And the fundamental drawbacks are still the same. Problem with the ring of datacentres is for hours each day, they won't be able to do any useful work because it's night time.
Datacentre operators drank the Green Kool-aid and gush about '100% renewable', but if they actually were, they'd be offline a lot of the time.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
