London has 400 GW of grid requests holding up datacenter builds While the UK government wants to turbocharge datacenter construction, a newly published report says there are already 400 GW worth of outstanding requests for connection to the power grid around London, and regulator Ofgem estimates 60-70 percent of these will never happen. These figures are drawn from the EMEA Datacenter …
Latency. They want to be as close as possible to the internet exchange that their clients will connect to. Which will be near population/business centres for the same reason.
You can transport power across the country with minimal loss, but the speed of light is fixed so you can't reduce latency by anything other than building closer.
Internally latency matters, but the additional couple of ms isn't relevant for the vast majority of applications.
Power availability should be a much bigger factor - and I bet it would be if we had regional pricing for energy... The reduction in cost from running in scotland would immediately outweigh the trivial latency increase
"The reduction in cost from running in scotland would immediately outweigh the trivial latency increase"
The major avoided cost of regional location would be avoiding power grid costs. The wholesale price wouldn't be much different. And the unfortunate thing is that the grid costs have been incurred, so they'll get lumped on other energy users, there's little real saving in system costs. There is also the question of what Scottish DCs would run on when wind power is low. If they run on their backup systems, that'll likely be dirty and relatively expensive, if they need to go into the wholesale market then they'll be hammered if they haven't got contracted or hedged supply contracts.
Energy prices do of course vary regionally at the moment, but not by a lot. There are proposals that the net zero extremists like of much more regional pricing, effectively having isolated energy markets by region. This is (like smart meters) one of those wildly expensive bad ideas that hasn't been properly thought through, but stands a good chance of being implemented. Which leads me to wonder what government think will be the foundation of the wealth to support public services in an economy where energy is too expensive for any commercial purposes more intensive than a coffee machine.
There might be lower cooling costs for a datacentre in Scotland though and it would help redress the London-centric imbalance that the UK has. As for grid charges I thought Scotland had a lot of windfarms and was a net exporter of energy?
But I have to admit that my first reaction to the headline was 'Good. That's helping reduce the environmental impact'.
There would probably be slightly lower costs of cooling in Scotland simply because of lower ambient temperatures, but I can't see the 2-3 degree difference being a big deal in the wider picture.
As for Scotland being a net exporter of energy, that doesn't mean much. Suffolk is a net exporter of energy, but it doesn't mean any dramatic reduction in the cost of power. In almost all scenarios any sizeable Scottish DC isn't going to be in the relatively remote locations of the large wind farms, so needs a grid connection, and the wind farms sell into the UK wholesale market which sets the price (although a different price may have been agreed under the contract for difference model that the government uses to subsidise the winners it has picked).
And that prices is actually set by the costs of the most expensive generator that can pick up the last few % of demand.
Which turns out to be gas. *
That's what runs when the sun don't shine, the wind don't blow and the dams are empty.
BTW studies show that while both retail and wholesale gas prices are roughly the same as other parts of the OECD yet we have some of the highest electricity prices in the developed world.
That combination of paradoxical facts suggests a)Government involvement in the problem b)Market reform is needed.
Having a more electric economy in the UK would be a good thing if electricity was more in line with prices in the rest of the developed world.
*Which raises the question could coal (given the UK sits on about 200yrs worth) be cheaper at greater environmental impact?
In Australia, a little while ago, it was discovered that a solar power operator was disconnecting part of its production from the grid to ensure that more expansive generators would be put online, providing that operator with a far biggest revenue than otherwise.
This is the invisible hand of the market at its finest...
Given the closure of much of Scotland's energy intensive industries there are plenty of brownfield sites with existing grid connections that could be utilised to deliver the additional data centres. Unfortunately our crass London-centric leadership is too myopic to appreciate that such investment in these brownfield sites is beneficial to the environment - saving on construction in the South East, lower transmission losses for the power and no shortage of water. The latency issue is largely a red herring unless you are involved in exceptionally time sensitive activities, e.g. some of the arbitraged fi
Scotland !?
Don’t need to go so far! :)
I’m a little surprised there isn’t a queue of buyers for former steel work sites eg. Corby (aka “little Scotland”)
Vast tracts of land at very cheap rates (with urban regeneration funding available) and large scale grid connections (now standing idle).
Thanks to electrification, it’s 1hr 10min from/to London…
Not got an inside track but several are just parking lots for new car imports ( including Rockingham Speedway). Others have been redeveloped into traditional warehousing/logistics hubs. Yes they have started demolition of the main town centre site with the sub station across the road, dont know whether it has been sold or whether they are preparing the land for sale. But the “surprise” is the absence of anything other than low value logistics/storage.
Given the closure of much of Scotland's energy intensive industries there are plenty of brownfield sites with existing grid connections that could be utilised to deliver the additional data centres. Unfortunately our crass London-centric leadership is too myopic to appreciate that such investment in these brownfield sites is beneficial to the environment
Problem is there can be competing demands. So I once wanted to build a datacentre in Ipswich on the old RS&J site. Brownfield, good grid connnections and fibre routes LINX-AMSIX practically on the doorstep. Slight snag, outbid by housing developers and a supermarket who could spend more and didn't want to share the site for a decent sized bit-barn. It can be much the same with other urban-ish brownfield sites because those are usually easier to turn into housing.
lower transmission losses for the power
Not always entirely true because Scotland's power is very heavily wind, so when the wind doesn't blow, it relies on imports from England and interconnectors. Which may also be future fun given Estonia, Latvia & Lithuania just announced they've disconnected their grids from cheap Russian electricity and now adding their load to the much more expensive, and wobbly EU grid.
The UK has had academic studies showing the viability of wave power for base load (if you install enough in enough different places which not only are we not doing but don't plan to) Hydrogen burns clean and can be harvest from biomass using pyrolysis or seperated from water using excess power from variable sources, wind can also be fairly reliable rarely calling on back up if it is geographically disparate enogh (again not only is it not in the UK we don't plan to make it so) whilst I would argue the cost and risk are unnecessary given viable alternatives nuclear electricity generation does not emit carbon. The problem is that our politicians and oil companies and whoever else rubbished reports like this in the past when the wealth of the nation was tied up in oil and were in denial for too long (only apologizing to the widow of the scientist who did the study decades later)> So we started late and lost the advantage we had in technology. Worse still when it comes to taking the technologies over and above the mature wind and hydro systems we already have to commercially viable scale this government and the previous government dont seem to have a clue. Both main parties seem more worried bout appearances (green washing) and what nigle farage is up to than they are about actually delivering legally binding emissions targets to avert disaster and give the Uk econamey a sustainable industry to export (renewables once scaled up in general are cheaper to operate and maintain that fossil fuel plants, understandable given you are not subject to the oil price volatility).
To solve this issue we need the Tories and Labor MPs as well as government ministers to stop worrying about reform uk and both plan a sustainable grid as well as deliverenogh interim capacity to bridge the gap balancing the climate issues and economic imperatives. God help us.
If you think that wave power and hydrogen from biomass can be any worthwhile part of the UK's energy mix, then clearly you don't know much about either.
Wave power can't be economic at scale (previous employers of mine threw enough money at wave power to find out the hard way), and there simply isn't the volume of unexploited raw biomass to make H2 synthesis worthwhile. Likewise, the same employers built a large pilot scale plant to investigate hydrogen from wind power, and the economics of that were equally shocking. Systemic losses are grim due to the multiple process stages, and if the electrolysis plant and compression systems are only used when there's "spare" wind then your asset utilisation (and thus capex cost per unit of output) skyrocket.
Just because something works technically doesn't mean it is sensible or viable. Green hydrogen is an especially good example of that.
I think you are confusing viability with your previous employers capability to make a profit.
The technology and available energy exist in wave power but to be economic it needs a transmission system that can reach it which we have not built and the technology needs to be mature and cheaper. However companies like orbital marine are generating energy.
What the government are not doing is investing in research getting from kilowatt and megawatt wind turbines existing to a system to put the energy on the grid required research at universities funded by governments as well as government and private investment. to get from where it is today to a sustainable system wave energy would need to be generated near big consumers and there would need to be a number of interconnecting sites. even if there can be profit with the limited imaginations of exiting companies and governments in the long run, national grid and generators would need loans at least to start that kind of thing.
In terms of nuclear i notice no objections. The overall point was we needed large investment decades ago to have this solved now and if we want it done by 2050 we need large investments now not just into the tech i mentioned but into anything with merit to see what comes out as the best technology and we need to build a grid that can take it.
"The technology and available energy exist in wave power but to be economic it needs a transmission system that can reach it which we have not built and the technology needs to be mature and cheaper. However companies like orbital marine are generating energy."
Stop believing the interwebs greenwash. Whether it's private sector or public money, if an investment doesn't cover its cost of capital then it destroys value, and overall society is poorer. In the case of tidal power, the reason it doesn't scale is because there's very little energy you can actually capture from a wave. Consider a hydro generator - they'll work at very tiny scales, to actually be worthwhile at a system level you need a head of around 70m. Now think of a wave generator, how much head is there? Around 1.5m seem a fair guess for the average. So the power from each wave is pretty minimal. Due to the reciprocating nature, you're only getting power for say half of the time, and with the low power available you need to blanket vast areas of sea to collect anything at scale. That costs money, a LOT of money. Now think anchoring systems that'll be proof against a force 11 storm, that's necessary but expensive. Then flexible, waterproof power couplings that don't sink the generator units, a transmission system to get it ashore...etc etc. Some of the cleverest engineers and technicians worked on this project, and still it couldn't be made viable. And then there's the fact that wave motion is only a second order effect of wind, so why not collect the energy from wind rather than using the sea as a very inefficient wind power collector? That's why people build offshore wind turbines, not wave generators.
I'll agree with you there's lots of sea, lots of waves, and that makes for big numbers on energy. But recovering it is difficult and very expensive, and experience shows that the majority of the designs simply aren't durable enough. You could make them more durable, at greater cost and compromise, but yet again why bother when wind is a better power source if you want renewables?
Consider a hydro generator - they'll work at very tiny scales, to actually be worthwhile at a system level you need a head of around 70m.
Low-head. high-flow systems are perfectly viable. The Galloway Hydros scheme, for example, has power stations with heads from 20m (Carsfad, Earlstoun) to 115m (Glenlee) using different turbine designs as appropriate.
Now think of a wave generator, how much head is there? Around 1.5m seem a fair guess for the average. So the power from each wave is pretty minimal. Due to the reciprocating nature, you're only getting power for say half of the time
Nobody suggests building wave power systems which generate using head difference like a hydro electric system. Instead they extract kinetic energy from very large quantities of water moving to and from over fairly short distances. To and fro because KE doesn't care whether you're going up or down, so the "getting power for half of the time" thing is wrong too.
I think you have been confusing wave power with tidal power systems.
The technology and available energy exist in wave power
They don't. I'm afraid. There isn't a single technically - let alone commercially - viable technology for wave power. Floating things (Salter's ducks, Pelamis, buoys) are too complicated and vulnerable, shore based systems (the Islay LIMPET) can only be used ina very small number of places and ocean bed structures would be prohibitively expensive to make, except in a very limited range of areas.
<......."Just because something works technically doesn't mean it is sensible or viable. Green hydrogen is an especially good example of that."......>
On the subject of hydrogen power, I think Lord Bamford, Mercedes-Benz Trucks and Manitou may not agree with you ;)
The responses to this original comment have been most enlightening - even more so when I realised(*) that I had read 'tidal power' when responders were actually discussing 'wave power'. There is no discussion of tidal power; why is that? Same issue of insufficient Kwh per £million invested?
(8) Most pissed off that The Register flags my correct spelling of 'realised' as an error. How dare you?
<i(8) Most pissed off that The Register flags my correct spelling of 'realised' as an error. How dare you?</i>
I think that's down to your browser, not the site. As I write this, "realised" is fine but "realized" gets the Wiggly Red Line of Doom, despite being the correct spelling here in Scotland, from where the Americans took it.
There is no discussion of tidal power; why is that? Same issue of insufficient Kwh per £million invested?
And to address that ... there just aren't many places where it would be technically viable. You either need a very large range, which for the UK limits you to the Bristol Channel or the Solway, or you need a very high flow rate, which limits you to the Pentland Firth, Strangford or a few other places on the west coat - Jack Sound, Kyleakin, the Sound of Luing and so on. The former sort raises all sorts of environmental problems which is why it has only be done one (Rance Barrage) and the latter only works in quite small areas.
Actually I think tidal stream power will probably make a useful contribution in due course, but it won't be a huge one and that "four times a day thing" will always be a problem.
"Energy prices do of course vary regionally at the moment, but not by a lot. There are proposals that the net zero extremists like of much more regional pricing, effectively having isolated energy markets by region. This is (like smart meters) one of those wildly expensive bad ideas that hasn't been properly thought through"
Nope... it's one of those ideas that has been very well thought through.
At the moment we pay both to turn off wind turbines in the north and to run gas peaker plants in the south, and then later to run gas peaker plants in the north and pay to switch off renewable generation in the south.
If we allowed that renewable energy to be sold to consumers at a lower rate, without requiring that same discount be provided all over the country then all regions would end up with lower pricing than they have now...
The concept of curtailment should be an absolute last resort - give the energy away to local consumers who can adjust their usage to suit, that will reduce demand earlier/later in the day/week, and therefore require less infill of burning stuff.
"If we allowed that renewable energy to be sold to consumers at a lower rate, without requiring that same discount be provided all over the country then all regions would end up with lower pricing than they have now..."
Excepting that regional energy markets will have significant additional costs themselves, the overwhelming majority of consumers won't be able to respond to randomly timed price changes, and by the time there's sufficient variable "smart demand" on the system to make much difference we should have EGL1 and EGL2 operational and the "need" to incentives consumers to shift demand will have gone.
You really think that the overwhelming majority of consumers can't adjust their usage?
Even if that was true, it doesn't stop them benefitting from the adjustment of those who can.
Regional energy markets don't have significant additional costs, and there will always be a benefit to shifting demand.
We lack transmission capacity to get power from where it is generated to where it is needed and this is not a quick fix. Mostly due to regs and lots of NIMBYs. Same in the US but more profit is involved there as people want lots of money for lines to run across their land. The UK grid is still largely based on the placements of the big thermal plants.
We NEED to get away from the marginal pricing model but this is Europe wide and makes nice profit for the renewable generators who are not on CfD arrangements. This can be done without smart meters but the savings won't be quite a good.
We lack transmission capacity to get power from where it is generated to where it is needed and this is not a quick fix. Mostly due to regs and lots of NIMBYs.
Tell me about it. Here in SW Scotland were infested with fanatical anti-pylon nutters. The absurd and divisive "Kingdom of Galloway National Park" proposal is largely their attempt to get statutory backing for their hysteria.
"At the moment we pay both to turn off wind turbines in the north and to run gas peaker plants in the south, and then later to run gas peaker plants in the north and pay to switch off renewable generation in the south."
So, if DC builders were encouraged to add battery backup and be able to charge up when renewable energy is producing so they could draw on that when peaking plants come online and wholesale rates would go up......
Datacenters aren't big employers so it's odd that politicians are falling all over themselves to have companies building them all over the place. What do they tell their constituents when those DC's are sucking up all of the power and residential supplies suffer or new homes can't be built due to infrastructure insufficiency?
Power constraints can be upstream generating capacity, but are far more commonly on the grid or the distribution networks. In all large UK cities, getting land adjacent to power distribution with unused power capacity is rarer than hen's teeth, for several reasons, the main one of which is that whilst the network was built with some spare capacity, population growth and new housing has largely used up both that capacity and available brownfield land. Adding new distribution or grid capacity in built up areas is phenomenally expensive, and usually somewhat less popular than smallpox.
"Latency."
That would matter for some applications but not for others. "
Stock traders that high speed trade by the millicent put their centers right across the street from the trading offices to shave nS off of the latency, but even a gamer would boggle at the costs to do that.
For the wider world, the latency can be all over the place depending on the resources that are up and running and how traffic gets routed. People sign up for Starlink when they could save some mS by getting fibre instead.
Yes. And for almost everything else, 3-6ms is not only irrelevant, it's undetectable.
For almost every application, what matters is sufficient power, bandwidth and having a small group of servers in the same building.
The distance to the end user matters a bit, but anything under perhaps 500 miles is absolutely fine.
Other exceptions might be game streaming, which might want to keep round trip latency below 15ms - but if you can do the frame in under 10ms you're still fine.
Latency is really only an issue for the algorithmic electronic traders. (Hence why stock exchanges now ensure all connected clients get the same latency. youtube.com/watch?v=d8BcCLLX4N4)
For the AI peddlers and cat video pushes, a couple of extra milliseconds of latency is not a problem.
If the data is travelling at the speed of light doesn't it get there before it's sent? If I look at the sun (not advised) the light is from 8 minutes ago as I see it right now but if I went to the sun I would be seeing light 8 minutes in the past. I also failed at science in school if that's any help.
That’s largely bollocks in the modern world with fibre a plenty available - most of my former employers retail SaaS Servers were in AWS EU/West-1 in Dublin…. They covered the UK, Ireland, Mainland Europe and around the world. Doesn’t seem to bother Netflix, YouTube, Disney+ for real time Video feeding and Twitch for game buyers and other organisations providing real-time payment authorisations.
If a few need to be close to ‘The City’ … that’s what Docklands DC’s are for.
I'm going to opt with optics.
London is the capital city, and is the 'heart' of commerce in the UK, so it's the place to be, and more importantly, be seen.
There's also the idea that London has a priority on power supply, and if power needs importing from France, it'll go to London first.
Add in that major cities attract the best and brightest: Makes it easier to recruit people you'd need onsite, be it in London or Birmingham or any other major city.
Costs a fortune in wages, however, but hey, more pay = more tax!
There's also the idea that London has a priority on power supply, and if power needs importing from France, it'll go to London first.
Not really how a national electricity grid works - you either have the whole place connected together running at one frequency, or you have separate grids which aren't interconnected. London may be where some of the subsea interconnects come ashore (those from The Netherlands for example) but it doesn't mean that "London has priority on that power".
Was talking about perceptions rather than actual infrastructure.
As to the grid: I take it you missed the warning a few years ago that there could be planned power outages to 'save the grid'? It's possible to shut down sections and prioritise others for power if there's not enough production. My point being that London is perceived to be of a higher priority than more isolated areas, so the grid might be shut down to remote areas just to maintain supply to the Capital and other priority areas. Although... there was talk about data centres being declared as critical infrastructure so would get a higher priority for power... but again, it's how those who are behind the building of these data centres see things... so back to 'Optics' (aka perception)
"Add in that major cities attract the best and brightest: "
I find that specious. Many of the jobs the best and brightest want to get are in major cities, it's not that they want to live there.
I've run the numbers and have had plenty of headhunters try to entice me to Silicon Valley. It's stupid expensive and crowded. So many want out on weekends and bank holidays that one needs to take days off to beat the traffic. If I worked in a bigger city, I'd not own a home, I'd likely be carrying a bunch of debt and be burning out cars in stop/go traffic on a daily basis. It's not what you make, it's what you keep. If I had a family, the last place I'd raise kids is in London, NYC, Paris or Madrid. Big city life is ok in your 20's, but after that, it wears.
I was making a point about Optics: Perception.
It is perceived that the major cities attract the best and the brightest, not stating that they do or that people what to live or work in a city.
Personally, I like being out in the countryside as it's more relaxed and I can go out for some rather lovely walks through woodland. Couldn't do that when I worked in Birmingham. But when I was younger, that's where I lived and so that's where I worked.
Again: I'm referring to the Optics: The perceptions.
Businesses congregate in larger towns and cities because they believe that's where they'll get the best candidates. However, this then costs them more in both rent and wages as costs in cities tend to be higher, and in London, they're much higher. That's why some businesses have relocated away from the main cities, and they've not had the problem getting skilled staff. But the perception remains that the big Cities, and in particular, London, are the places to be.
Oh, there are other reasons such as where key infrastructure is located, but it really shouldn't matter - and recent years have shown that remote working can be a viable alternative to large offices and local workers... unless there's a requirement for staff to attend their office for a few days in the week, which looks to be a growing trend. But optics: What the 'big bosses' see: That's not changing quite so quickly, and that's affecting the decisions as to where things are built.
Availability of workers is a key one, especially the lower-paid jobs like cleaning, groundskeeping, security etc. Those people would be less likely to work for somewhere with a long commute as it would eat into their already meagre pay.
Also, you want your site to be accessible to your customers, so a connection close to a major A road or motorway is preferable so that your customers can come in and out of the site easily as needed. You also want contractors (electricians, HVAC technicians, mechanical engineers, etc) to be able to get in and out easily.
Plus if you're having big bits of kit delivered (think electrical transformers, large tanks etc) you want easy road access, not somewhere down a country lane that's as wide as a tractor but no more.
Major cities also have more interconnects and require less digging up of roads etc to get cables laid (distance-wise) to get your site connected in, since you probably want multiple diverse paths for both power and data cabling.
There's other ancillary stuff too like availability of nearby food/shopping/hotel accommodation for your staff and visitors (to go and buy lunch, or if they have to travel long distances to get to your DC they may stay overnight), plus the weather - you don't want your site to be snowed in every winter so that your customers or staff can't get in or out.
>” Availability of workers is a key one, especially the lower-paid jobs like cleaning, groundskeeping, security etc. Those people would be less likely to work for somewhere with a long commute as it would eat into their already meagre pay.”
Which basically says locate the facilities in places where these people can afford to live and have the “rich” Londoners et al commute…
"Plus if you're having big bits of kit delivered (think electrical transformers, large tanks etc) you want easy road access, not somewhere down a country lane that's as wide as a tractor but no more."
That's really going off into the weeds. Getting a large tank/transformer/whatever delivered downtown can also be a major headache. Many large cities have very narrow roads, lots of 'street furniture' and require loads of permits to get a big shipment allowed. Hiring police to man road closures and needing to do these sorts of things in the middle of the night at double rates are also jacking up the costs. Between the two are middling size cities where there are two lane roads and plenty of infrastructure.
If I were working a lower paying job, I think I'd be stuck between commute costs and housing costs. Living and working in London at minimum wage cleaning offices is a trap. I'm not sure if I'd be better in Leeds, but maybe in Preston. Those places have positions for house/grounds keeping, security, etc and should be easier to find affordable housing.
Because that is where the majority of the engineers are...datacentres in the sticks are a good idea, but in terms of risk it'll never be as good as a datacentre near a city simply because it takes longer and costs more for engineers to get to it to perform maintenance, repairs, respond to emergencies etc...not to mention the time for spares, kit, fuel etc to be delivered...held up by weather and so on.
Sure, we could all disperse and follow the DCs but that's not going to happen. The UK has learnt it's lesson about "following the work" and relying on it to not just go bust or disappear.
I wouldn't move somewhere to specifically work there...that is big dumb on a grand scale...you also have to factor in that the vast majority of highly skilled and experienced engineers tend not to work for just one organisation...trying to encourage them to move to the promised land to work at a single DC just isn't going to work.
Engineering is also, at it's core, quite a collaborative profession...it's quite important for engineers to be around other engineers...whether they work together in a team or not...whacking a mjaor DC in Carlisle or somewhere like that and training/employing locally would end up creating a small isolated team of engineers...that's not great.
The only way that remote datacentres works is if engineers are paid more and have their travel costs covered...which would make a remote datacentre quite a bit more expensive than a city based one in the long run.
TL;DR it's not practical.
>”The only way that remote datacentres works is if engineers are paid more and have their travel costs covered...which would make a remote datacentre quite a bit more expensive than a city based one in the long run.”
Much depends on what you mean by “in the sticks”. But a sensibly located (and sized) datacentre will be cheaper to operate and maintain if located outside of London/SE even if the engineers live in the SE and have to be paid travel and accommodation expenses. However, idiots like to nit pick and so will take exception to the seemingly increased engineer expenses whilst failing to grasp these are lost in the huge costs of maintaining a site in the SE.
For many years I worked for companies who maintained a “London office” which wasn’t much more than a few desks and meeting rooms, the real work being done (ie. Offices where 500+ worked daily), were well outside of London.
We have seen similar with schools, where it was significantly cheaper to operate larger schools and bus children in than to operate many smaller schools where people to could walk to. The trouble is small minded people see the transport costs as some thing they can cut and pass on to the parents…
"The trouble is small minded people see the transport costs as some thing they can cut and pass on to the parents…"
That's until there's a big spike in oil prices and diesel doubles in price within a week. There's no controlling that cost so one just has to live with wherever it goes. I just had something like that happen yesterday where the price of petrol at the lowest cost station went up $0.16/gallon from when I went by in the morning to when I was coming home in the afternoon. I was noticing that prices had all been marked up that day so I went by my local station that hadn't gone up and topped up the tank.
The cost of transportation was exceptionally low for some time and that led to goods being shipping from the other side of the world and consolidation of things like schools and shopping. Now those costs are going up again to the point where they aren't as insignificant for many things. There have also been interruptions to supply chains due to piracy, theft from train cars, weather and other things that illustrate that the perceived efficiency can turn around and be a major problem if things aren't smooth sailing.
Travel costs were only one aspect that I mentioned, I wasn't intending to imply just putting money in the engineers pocket...there also has to be some kind of will and motivation for engineers to want to travel out of their way and stay overnight somewhere just to work in a poxy DC. It's awful spending time in datacentres...they need quality of life improvements.
Datacentres are depressing as fuck, it's made even worse when you have to go away for night to work in one...not least because the reason you're there is probably critical, time sensitive and therefore stressful.
It's lonely as hell and really uncomfortable in a datacentre with the noise, heat and cold...asking someone to also go and stay in a shitty Travelodge and spend hours commuting there makes it far worse.
As an engineer I have to say "fuck that" to this. I don't care how much money a datacentre in the middle of nowhere might save...it's not enough to justify dragging my ass there...whether it saves money or not is of no consequence to me. I'm pretty sure most engineers would feel the same way...why should engineers have a significantly worse time of it to save a business / the taxpayer some money?
If they really want to save money on datacentres (which I suspect they don't) and justify putting them out in the middle of nowhere, then datacentres need quality of life improvements. They need fucking good accommodation onsite with catering that costs nothing to the engineer. If the accommodation was to a high standard and the catering was high quality, I might...*might*...consider going out of my way to visit a DC...otherwise, fuck it...send a junior that doesn't know any better and see how far you get.
Nothing pisses me off about datacentres more than the token efforts at creating engineer comfort...e.g. an overpriced and badly stocked vending machine.
I've slept on datacentre floors before and been there for days and the one thing I learned during those experiences is that nobody gives a shit...it's not really the customers fault though, they haven't a clue what working in a DC is like, let alone trying to sleep in one...it's abstract to them.
At the very least, if you build your DCs near to where most of the engineers are, the engineers can go home and not be trapped.
If they really want to save money on datacentres (which I suspect they don't) and justify putting them out in the middle of nowhere, then datacentres need quality of life improvements. They need fucking good accommodation onsite with catering that costs nothing to the engineer. If the accommodation was to a high standard and the catering was high quality, I might...*might*...consider going out of my way to visit a DC...otherwise, fuck it...send a junior that doesn't know any better and see how far you get.Nothing pisses me off about datacentres more than the token efforts at creating engineer comfort...e.g. an overpriced and badly stocked vending machine.
You had a vending machine?!?
But been there, done that, remember when Telehouse got some creature comforts like some sofas and a vending machine. Also sunbathing on their roof and being told off by their security. But having spent many hours in many DCs, I totally agree about the lack of creature comforts in most. Even when some have been staffed 24/7, the lack of any decent rest/break areas has been depressing. So guessing your downvote was from someone who's never had to be onsite in a bitbarn for any extended period. DCs that were intended to be 'ILO' were the worst because their designs often don't account for humans at all.
So I agree that al DCs should have a rest & desk area because something will inevitably happen that needs engineers and sysadmins in the building. I think this should also be part of DR/BC planning because although they may be intended for remote management, what happens if the business can't do that? Especially when whatever is in the DC is generally business critical.
Two vending machines technically. One was for crisps and chocolate which cost a minimum of £3 an item and only had 3 or 4 items left that were shit...and a coffee machine that cost £2 per drink and was awful...like a 90s style coffee vending machine. Didn't have real milk in it and everything was made from concentrate not proper tea or coffee...you couldn't even bring your own milk in because there was no fridge.
Yeah, I agree.
A lot of people haven't got a clue what it's like to be stuck in a DC especially when it's an unexpected stay. Some of the hardest times in my life were when I was stuck in a DC...it's one of the most hostile environments I've experienced for extended periods...and I've slept in the back of vans in sub zero temperatures, had dengue fever in a remote part of a jungle and various other shit in my life.
Aside from the physical extremes in datacentres, the mental impact is hard to explain...your senses are constantly under heavy load in a DC...noise, bright lights, blinking LEDs everywhere...even after a few hours it grinds on you...all while you're under pressure to resolve a critical problem...because lets be honest, most of the time you're visiting a datacentre to resolve a serious problem...you don't just go on a whim.
Also add to the extremes the fact that quite a few of us in tech probably have mild undiagnosed ADHD and the environment becomes very extreme very quickly. Even people with no ADHD or anything like that would struggle to maintain focus for more than an hour at a time in a DC...
Aside from the physical extremes in datacentres, the mental impact is hard to explain...your senses are constantly under heavy load in a DC...noise, bright lights, blinking LEDs everywhere...even after a few hours it grinds on you...all while you're under pressure to resolve a critical problem...because lets be honest, most of the time you're visiting a datacentre to resolve a serious problem...you don't just go on a whim.
One of my good deeds was visiting a DC to do some installs and discovering their helldesk staff. They had desks in the machine room wedged in between some of the air handling kit, and were working 12hr shifts.. Which were supposed to be 3x8hr, but some mangler, somewhere, figured it was cheaper to pay OT for 2 shifts than staff 3. But it explained why we could rarely hear them when we phoned. Luckily I had a mate who worked for the councils evironmental health dept* and arranged for a 'random' occupational health inspection. That resulted in a long list of potential penalties for having an unsafe working evironment, resulting in the staff getting a decent office.
"So I agree that al DCs should have a rest & desk area because something will inevitably happen that needs engineers and sysadmins in the building. "
There also needs to be space to prep new machines when things are being changed (I hesitate to use the word "updated".) Trying to do that in the aisles is a major PIA. Storage space is also important as kit is being delivered so it doesn't get stacked in the area meant for rest & desk. Murphy states that the biggest volume of stuff will arrive weeks before a small box of critical items that are required for the install turn up.
There also needs to be space to prep new machines when things are being changed (I hesitate to use the word "updated".) Trying to do that in the aisles is a major PIA. Storage space is also important as kit is being delivered so it doesn't get stacked in the area meant for rest & desk. Murphy states that the biggest volume of stuff will arrive weeks before a small box of critical items that are required for the install turn up.
Yep, staging, goods in and storage are also important. So a typical job I might do would involve tin from multiple vendors arriving at different times, then when it was all in place, we'd go install and commission that. Which is all part of DC planning because often they were designed for customers installing a few u's of tin in pre-installed racks. For big jobs, that doesn't always work, especially if the tin doesn't fit in standard racks, which messes up neatly dressed lines of racks. And then that can really mess up things like hot/cold aisles, or just HVAC in general.
Which having been on the DC side, can lead to interesting conversations with sales, who've sold 8 racks of collocation! And then digging into that sale, discovering it's not 8 standard racks and the standard price, and pre-sales are there for a reason. So that 'simple' sale may turn into say, 30 racks worth of HVAC, needing a cage for security taking up more footprint and perhaps additional cooling because some kit generates a lot of heat per rack. Then sometimes additional FUN! with things like storage arrays because spinning rust drives are heavy, so a rack of those might also need spreader plates to stop it falling into the suspended floor. Or in one case, weight of the storage rack exceeded the load capacity of the goods in ramp, so that needed to be reinforced, along with equipment hire and engineers to help install the customer's kit safely.
But that's the joy of network and IT planning. Some of those issues might not apply to these new bit-barns as those might be intended for a single tenant rather than common collocation. Other stuff would still apply, ie staging and secure storage areas for spares etc.
We need 400 GW, and probably more. We are building 3 GW and have 9 GW planned. So only a 97% shortfall.
Maybe we should just say "no". Ideally before we are all relocated to Gibraltar and the Falklands and this green and pleasant land is turned over to electricity generation.
@Jason Bloomberg
"Ideally before we are all relocated to Gibraltar and the Falklands and this green and pleasant land is turned over to electricity generation."
The green and pleasant land wouldnt need to be turned over to electricity generation if we drop unreliables. We could instead build actual power generation and even make enough for cheap and plentiful! Thereby reducing inflation and maybe even cause a little deflation to our benefit!
@John Robson
"Obviously the timing of that power isn't ideal for our conditions, but space isn't the issue here..."
This is where the issue of energy generation falls over when they forget the aim is energy generation. Lots of monuments to a sky god but a lack of energy
>” when they forget the aim is energy generation.”
I have had a number of quotes for PV + battery system.
It is clear the majority of installers including the major energy generators, do not take this viewpoint, so propose undersized systems.
One installer, humoured me and allowed me to play around with their model: design a “standard” system and pay out circa £12k, it will pay for itself in circa 12 years, however, I would still be buying significant amounts of electricity at expensive times. Max out the PVs and have a battery large enough to carry me through a weekend/72 hours, the price increases to £20k but the payback drops to 7 years, plus over the 30 year life it will generate an income of circa £100k a circa 16% Pa. return on investment.
"Max out the PVs and have a battery large enough to carry me through a weekend/72 hours, the price increases to £20k but the payback drops to 7 years, plus over the 30 year life it will generate an income of circa £100k a circa 16% Pa. return on investment."
Did you also work out the return if you took all of that money and stuck it in a timed savings account?
Step one is hunting the low-hanging fruit. My home is very nearly 100% lit with LED (LED oven bulbs don't work long). Stopping up gaps to better seal the home and making sure there's sufficient insulation that's properly installed. I found heat leaks in my home due to fiberglass insulation that was not installed correctly. My home has been heated with solar thermal since 2016. Last year I finished getting the evaporative cooler to run on solar with battery b/u. I do heat the bathroom with an electric heater prior to bathing and have an electric blanket which I love. Two more solar thermal panels should be up by the weekend if the weather holds. I'm putting rooftop solar PV off until I can afford to replace the roof. A battery has a very long ROI, so I'll wait on that until I have numbers on the PV. The cheapest and easiest things tend to have the most ROI.
>” Did you also work out the return if you took all of that money and stuck it in a timed savings account?”
Yes, short of becoming a money lender, I’ve not seen a low risk way of getting a 16% pa return on investment. Okay, a risk is that the prices I get paid for electricity I send back to the grid crash, but the batteries and intelligent management system, allows me to only sell when “the price is in my favour”. Another risk is selling the house, but then I can ask a price premium…
I also worked out what I would be paying out (to the energy companies) without the panels…
Agree about the low hanging fruit, I decided solar thermal wasn’t as simple and flexible as solar voltaic, plus looking ahead as I replace gas boiler etc. I will be using more electricity.
Would be interested in knowing more about your multi year experience ie. You have had it long enough to have gone through a couple of winters and hot summers, and so found the flaws.
> LED oven bulbs don't work long
It seems daft the bulb is in the oven rather than outside and light tubed or fibred in.
> I found heat leaks in my home
Yes a cheap thermal image camera on a winters evening (outside temperature even -1C) is really good at finding these otherwise invisible leaks. Although I have still to resolve why a 2x2m square of my kitchen floor is measurably colder than the rest of my ground floor, even though it’s a single suspended concrete floor.
> Stopping up gaps to better seal the home
Need to remember the draft free trickle ventilation to prevent mould etc.
> I do heat the bathroom with an electric heater
Infra red / radiant bar / radiator ?
As a child I was always impressed by the infra red radiant bar they had: walk into cold bathroom, turn on and within minutes the room felt warm.
"Yes, short of becoming a money lender, I’ve not seen a low risk way of getting a 16% pa return on investment. "
I put that in since so many take it as common wisdom to stack money in retirement accounts, CD's and similar things that don't pay back very well. I'm all for having a SHTF pile of money stored away, but if a solar system has a 10% PA return, it's crazy to be putting money in an IRA. The same goes for putting the money in that account while having a mortgage unless you've locked in super low rates.
My old oven has the bulb inside and I have an extensive light bulb collection so I'm set for life on filament lamps. Estate sales for the win. I like the low wattage filament lamp in the oven as I often leave it on in the winter to have a warm box to let dough rise, make iced tea, that sort of thing. The light is incidental.
I have a CAT S60 phone I picked up cheap that has a FLIR IR camera and checked the house early in the winter for thermal performance one night and could see where heat was leaking out. It's an old house so all but one window are single pane. That's something that's on the list. The solar thermal heating works a treat. I found a really great website where an aerospace engineer that was off work for some reason did a bunch of investigation and wrote up a very complete paper on what he came up with including the science and math. The site isn't around any more, but I did download the paper before it went. Fundamentally, it's insolation (W/Sqm) x Sqm and a box that's glass:air gap:black panel:air gap:outside house wall with a passage that goes in from the house behind the black panel, up, then down, through to the gap between the panel and the glass, up and through back into the house. The reason for the Up/Down is to make a "cold lock" so it won't run backwards. With a blower that is rated for 250cuft/min, the panel temp peaks about 50C. With the blower off, the panel gets up to 90C or more. You can find vids on YT of people using soda cans or rain gutter tubing to make something like this, but I found that it's way more cost effective to use 24gu sheet steel painted black. I get the clear glass from salvaged shower doors I find online (free or very low cost). Shower doors are safety glass so while a bit less transmission, much more impact resistant than plate glass.
The bathroom heater is an oil-filled radiator. Cost me ~$5 at an estate sale (3@$5ea).
How much different is that 2x2m patch of kitchen floor? How even is it? You may want to drill a hole and investigate as long as you don't think you'll go through an embedded water or waste pipe. If you drill in and find water, there may be a leak that's wicking heat away. Could be an air gap. Could be a secret tunnel that leads to a super villan's secret lair.
If you open up a couple of windows and vent the house every so often, the loss in heat is not that great but it does exhaust humidity. I'd do that before having a constant air leak.
I'm not putting solar thermal on the roof as it self shades from the eaves in the summer by having it on the south side of the house and PV is a better use of that surface. I can always siphon off any excess power generation for heating if I have it. There are even power management devices that will do that automatically so you only export when you have plenty of hot water, the EV is charged, the house is warm and nothing more is needed. Sorry, I can't remember the name of the one I've seen that impressed me.
> How much different is that 2x2m patch of kitchen floor?
It’s a good few degrees colder in winter, and thats through the flood vinyl - you don’t want to stand in it in bare feet. I suspect it is missing its insulation sheet under the scree as it has welll defined straight edges.
> Could be a secret tunnel that leads to a super villan's secret lair.
That’s under a large bush I planted some years back, on grounds of plausible deniability as it goes under the neighbours driveway, which hasn’t yet suffered a cave in…
I actually suspect it is part of a water course as before houses the area was locally known for having springs.
I actually suspect it is part of a water course as before houses the area was locally known for having springs.
Hmm, I wonder if this is where I had a bunker researching ZPE? Or it might be worth digging into further with your council or BGS to try and find out more about the local hydrology. Thinking if there are springs, they can often have a bad habit of surfacing, so your cold problem might turn into a damp problem, especially if insulation is missing.
"If you open up a couple of windows and vent the house every so often, the loss in heat is not that great but it does exhaust humidity. I'd do that before having a constant air leak."
Heat recovery ventilation systems are cheap(ish), have been around for decades and solve that issue nicely
"Heat recovery ventilation systems are cheap(ish), have been around for decades and solve that issue nicely"
If you are having work done that gives access, yes, they can be fantastic. If you are just renting or routing all of the ducts would be difficult, it's easy to open a couple of windows.
"Yes, short of becoming a money lender, I’ve not seen a low risk way of getting a 16% pa return on investment. Okay, a risk is that the prices I get paid for electricity I send back to the grid crash, but the batteries and intelligent management system, allows me to only sell when “the price is in my favour”. Another risk is selling the house, but then I can ask a price premium…"
I wouldn't count on feed-in tariffs in ROI calculations. The power companies would like those to go away completely so it's only a matter of time before they do. I'd be sure to keep good documentation so you can support a price premium when/if you sell if it will take some selling at that juncture. In another decade, enough people will see the value from their own experience. Even if the panels are nearing the end of their useful life, all of the wiring and fitted hardware makes putting new panels on very easy. Less than half the price of my solar projects has been the cost of the panels.
The 400GW sounds a bit spurious, given that the consumption of the entire country is generally in the 35 to 45GW range at the moment. Suspect units are wrong and its 400MW, but that sounds on the low side to be getting worked up about. There's probably at least the same again for EV charging queued up.
Whatever, AI is absolutely the last thing we should be spaffing the over-stretched resources of this country over.
It's not spurious. The power demands backed up around London are extreme
When Royalmail wanted to electrify their london fleet they needed to outfit depots with charging facilities to make it viable. At the scale involved they were quoted a 20 YEAR backlog on the order
"When Royalmail wanted to electrify their london fleet they needed to outfit depots with charging facilities to make it viable. At the scale involved they were quoted a 20 YEAR backlog on the order"
EV's are mainly about distribution since it displaces the massive energy usage used for oil refining. Trying to electrify a whole fleet all at once is madness. They need to start with centers that run 4-5 post vehicles and work their way up. Charging 4-5 EV's slowly overnight doesn't require a line of 100m pylon to run the power. Depending on the routes, it can be equivalent to turning on a few kettles.
"There is no way there is pending demand for about SIX times the entire capacity of the UK grid."
Have you done the math on replacing Britain's domestic oil/gas heating with electricity? It's not helped by most heatpump installations being vastly underspecced and giving it a bad rep, but it's coming
Electrical generation is only 1/3 of our carbon emissions and it's going to take a 6-8 times annual generarion increase to replace them
Can you do THAT with renewables?
"Either they meant MW or they are talking about GW-hours - but that seems less likely in this context."
It may also take into account both feeding in and taking power away. The initial grid wasn't designed to arbitrarily route power from multiple sources. What was assumed was that power would come from just a few sources and fan out from there. South Australia had a big blackout when the ins and outs lined up in the worst possible way after replacing a lot of large power plants with a bunch of renewables.
South Australia had a big blackout when the ins and outs lined up in the worst possible way after replacing a lot of large power plants with a bunch of renewables.
At least SA has sun. Well, a lot more frequently than the UK-
https://notalotofpeopleknowthat.wordpress.com/2025/02/12/why-solar-power-does-not-work-in-britain/#more-85707
Crazy Ed Miliband wants to cover the countryside with hundreds of industrial scale solar farms.
Yesterday the combined 18 GW of solar capacity which we already have produced the princely total 3.1 GWh, having peaked at 0.7 GW at midday. That meant they were running on average at just 0.7% of their capacity.
Yesterday being Feb 11th, and the UK being it's typical grey & overcast self.
"At least SA has sun. Well, a lot more frequently than the UK-"
That's great until you melt down your distribution network since it hadn't been designed nor upgraded to be used with inputs in so many different places.
A big part of electrifying stuff is distribution as much, if not more, than generation. I love the Connections episode that covered a major blackout in New York. One small thing causing a big portion of the Eastern US to go dark that was like a domino chain falling over.
That's great until you melt down your distribution network since it hadn't been designed nor upgraded to be used with inputs in so many different places.
Plus the min-max problem with solar and wind. So-
18 GW of solar capacity which we already have produced the princely total 3.1 GWh, having peaked at 0.7 GW
Ed Millibrain wants to build a lot more solar (subsidy) farms. When they're at their minimum, they're pretty much useless given the projected demands caused by bit-barns and Net Zero. When it's a nice, clear sunny day, their output might be more useful, but their output might not coincide with demand. So then solar (and wind) farmers get paid billions in constraint payments to not despatch power to the grid.
But that's the problem; if the barns get first dibs on what's generated there will be nothing for the rest of us.
This is why they want direct connection to the generators, want special non-grid deals - As their needs rise they expect to get it, the grid to be cut or reduced to provide for that.
Currently, connected to the grid, they will have to rely on Smart Meter disconnections to provide the power they need for those barns
"...they will have to rely on Smart Meter disconnections to provide the power they need for those barns" - exactly.
The UK came perilously close to being unable to meet the demand for electricity due to a Dunkelflaute* event in January. Given the current (and previous) government's genuflecting to corporate interests over us sans-culottes, it's perhaps unfortunate that the lack of sufficient supply didn't result in smart meters receiving the commands to turn them off as it might have made people realise that it's not just a massively costly and largely pointless exercise in consumer shaming but a means to ensure that the right type of consumers (e.g. bit barns) get the leccy whilst the peasants can go shiver around a fire.
* alas not an occult prog rock band heavily dependent on woodwind, it roughly translates as "dark doldrums"
whilst the peasants can go shiver around a fire
I don't have a fireplace in our new house, so I've just bought a generator and installed a transfer switch, for exactly that reason. I don't trust the numpties in Westminster, of any shade, to keep the lights on.
"But that's the problem; if the barns get first dibs on what's generated there will be nothing for the rest of us."
It's also means that in a given area, if there's a bitbarn hoovering up leccy, it can mean that no new housing can be built and a proper company that employs people can't locate there as power is fully subscribe. Not too far from my house is a satellite neighborhood that can't have any more homes built. There's not enough water supply and it would take a fair size investment to drill and operate another well. It would take a developer coming in and building a large number of homes for the numbers to work out and that's not likely in that location. There are enough resources in the main part of town to support more housing so that's where builders are buying lots and erecting bare minimum homes to sell. (they are really poorly built) It's better to buy a "used" home and invest the savings on improvements.
Did a decimal slip somewhere in the report from Ofgem?
I don't think so, numbers of that order of magnitude seem to appear elsewhere on Ofgem's site, e.g. this blog post, which says "The connections queue now stands at 701 Gigawatts (GW) with estimate this could rise to 800GW by the end of 2024, an amount of electricity generation that is over four times more than what is predicted we’ll need by 2050" (That's for the whole country, not just London).
But 400GW just for Datacentres in London?
The articles's "OpenGraph meta-title" (the text that you see when sharing the article on something like Slack or WhatsApp, seen in the HTML headers) says "London has 400 GW of grid requests for datacenter builds" - whereas the actual title is (or has been changed to?) "London has 400 GW of grid requests holding up datacenter builds" which could mean that there is 400GW in the queue in total, including solar farms, gas/diesel generators, EV chargers, housing developments, etc. and all this is holding up the approval of the er, much-needed datacentres.
But then again, the article says "These figures are drawn from the EMEA Datacenter Market Update for H2 2024" so it might be talking about just datacentres, but it's one of those evil Javascript PDF "flipbooks" so I can't be bothered to read it.
Edit, OK I relented and read the ghastly flipbook - London has 265MW DCs under construction and a further 1260MW planned, so more like 1.6GW, not 400. So the reg "meta-title" is misleading clickbait.
Well, the "misleading clickbait" that I was commenting on is the "meta:og-title" which appears only when you share a link to the article (or read the HTML source, line 11 of view-source:https://www.theregister.com/2025/02/10/london_has_400_gw_of/), which says that the 400GW is all Datacentres, which OFGEM does NOT say, nor does the actual article or its title.
Well, the reason for the backlog is essentially because the DNOs (Distribution Network Operators) and TSO (Transmission System Operator) are receiving requests faster than they can deal with them.
That is partly due to duplicate submissions and 'zombie' projects that were never intended to be built, but objections to "pylons" and substations from NIMBYs (and so-called NOBYs or BANANAs, who say "Build Absolutely Nothing Anywhere Near Anything") do obviously reduce the ability to clear the backlog, so it is logical to say they are partly responsible for it. And as I have said before, the NOBY anti-pylon argument is misinformed.
(That said, I think the NOBY brigade are right about Solar - The place for solar panels is on roofs and the odd sheep field, not high-grade arable farmland)
But, we really do need more overhead pylons and substations, net zero or no net zero. The grid is dangerously unstable without them, and there will be deaths if we have unexpected blackouts. Imagine your granny going upstairs to the loo for example when the lights go out. A certain percentage of the elderly/infirm would fall or otherwise injure themselves, as well as road accidents when the street/traffic/building lights suddenly go off, it becomes very dark indeed
>” Well, the reason for the backlog is essentially because the DNOs (Distribution Network Operators) and TSO (Transmission System Operator) are receiving requests faster than they can deal with them.”
My neighbour is very happy, there is a shortage of high voltage engineers, so he is pricing work for 2+ years down the road…
> But, we really do need more overhead pylons and substations
Shame, due to government dilly dallying, the only UK company making the new design pylons went bust, the last I read, we would be purchasing from China…
Similar problem with transformers etc. these used to be made in the UK - remember GEC… now imported with lead times measured in years….
Much of the “NIMBY” is self created by government et al. Rather than intelligently engaging with people, they delay and delay then complain when they want to do stuff in a hurry that people are raising questions.
> The place for solar panels is on roofs
Agree, only problem is it tends to cut out the leeches/middle men - much favoured by the Tories…
> That is partly due to duplicate submissions and 'zombie' projects
Suspect also some are putting their application in early. We have just had the initial consultation documents for a massive field based expanse of panels across several square km of arable farm land currently being used to grow “weetabix” but will be used for sheep grazing once panels installed; they have already put the grid connect application in.
No. The 400GW number is in the referenced report and probably correct. The fact that this potential 400GW demand is holding up new bit barns is also correct. People's interpretation on the title assuming this 400GW of power is for new bit barns is incorrect. The consequential leap which many people make assuming this must be for AI is consequently also incorrect.
It's a headline that leads a lot of people to wrong conclusions. Perhaps it's clickbait, or perhaps it's Hanlon's razor.
If you take as read the intent is to move people off carbon fuels on to electricity the numbers don’t look so far out.
For example the typical home uses 4x more gas KWh than electric, so (very) roughly that 50GW of electricity needs to become 250GW, if we then add in the energy we currently put into our ICE vehicles …
Agree the headline is misleading.
So, yeah, you need to pull 400GW out of your pants to supply everybody. Bit barns don't care where they are located, as long as they have the power. You can put them on the most unsavory of places too.
Let's see:
- You need to produce a hell lot of power, year-around.
- Production and consumption can still be located anywhere.
-The consumer has no ecological concerns whatsoever, so the production doen't need it either.
Nuclear plants check all boxes. The newer design ones can generate more maneageable waste than ever. It doesn't even need to be in anybody's backyard, regardless it pollutes or not.
Just put both in places nobody needs to go.
"Power issues aren't just a problem for the UK, and the report even claims that in Germany, demand is reported to be higher than the total power available for the country."
It hasnt been difficult to see this. Germany leads the way, we can easily see what a bad idea all this green madness is and yet we follow. Kick Militwit out, get someone in who wants to build actual power generation. Someone willing to cut off unreliables if they fail to provide and not expand on the wishful thinking unless they manage to make the technology work. Rush through permissions to get power plants built and get our supply up.
Who'd 'a' thunk it.
Flooding the planet with AI bit barns: means flooding the planet with power stations grid and water infrastructure: means planning mega-bitchfights, mega-investment and mega-buildout, over many years.
Take a tip from the old academic supercomputer greybeards [icon]: Quicker and cheaper to wait a few years while Moore's Law does its stuff, then just drop a latest-model toy in your back yard and plug up.
"Flooding the planet with AI bit barns: means flooding the planet with power stations grid and water infrastructure: means planning mega-bitchfights, mega-investment and mega-buildout, over many years."
Given history of these sorts of things, many ventures will go Tango Uniform leaving behind useless buildings and loads of concrete to break up and haul away leaving brown sites that were built on productive green land when they got built.
Everyone with a brain and a basic common sense could have seen this coming.
It's almost like, selling off basic utility providers and infrastructure and allowing them to extract untold profits whilst never bothering to invest in upgrading said infrastructure.... was a very bad idea.
Cue a crumbling infrastructure, pollution like we've not seen since the industrial revolution and toothless govt and regulators who refuse to do what needs to be done... nationalise all utilities, remove the 'for profit and fuck the consumers over again and again' side of the equation. Run them at a small profit that is ring fenced to invest and upgrade in the infrastructure and renewables... it's the only fucking thing that makes sense... but goes against the wishes of the govt's corporate overlords and paymasters.
.The infrastructure was already 'crumbling' as you put it well before privatisation, and investment in upgrading and improving the state of the grid transmission equipment has been considerably greater and more effective than it was under the CEGB prior to 1990, when they basically sat on their hands and did little beyond very basic maintenance and emergency repairs.
The UK grid currently supplies ~40GW (it can go to ~55GW).
Assuming all the grid connection requests are "real" (they won't be), a lot of extra generation capacity is going to be needed. Wind/solar are too unreliable to cover a base-load requirement like this and it takes 15+ years to get a 3.2GW nuclear plant approved, built and commissioned (it used to take a lot less, but is most likely even longer these days).
They are only going to get the power they need in the short to medium term if they invest in their own generation capacity.
"They are only going to get the power they need in the short to medium term if they invest in their own generation capacity."
Or buy existing capacity which was never intended/planned for them like Amazon who recently bought half the output from a windfarm in the Moray Firth.
Not entirely sure how the likes of Amazon/Meta/MS/etc owning the output of renewable energy plants really benefits the UK......
It seems pretty obvious that the solution to this is to shift business out of the capital to areas where there's the power capacity and space to accommodate the kind of business that needs gigawatt datacentres but no, that will never happen while swinging your (tiny) dick around is such an important part of your business model
So the UK energy user will have to pay extra to fund these connections and maintenance and then the UK consumer will have to pay for the electricity plus the profits of the data center.
Wow business just cannot lose can it, if it goes pear shaped then the UK tax payer has to step in and pay.
Sort of a heads we win, tails you lose and oh by the way we are keeping the coin and leave your coat here as well before you leave.
But how do OFGEM decide which project to allow?
Once they've weeded out the zombies they will have finite resources to study the rest, so how do they allocate those resources?
It's a tricky question.
For example there are 1200 livestock and poultry farms in the UK that meet the US "Mega farm" criteria. IOW every one of them will have at least a hectare of roof that could be covered with a PV array, as well as a shedload of biomass generation (as well as acting as a possible hub site for anaerobic digestion to surrounding farms and fertiliser).
IMHO between them that could put 1-2GW on the grid in years not the decades of Hinkley Point C.
But does Ofgem have the staff to conduct reasonable oversight for so many?
Once they've weeded out the zombies they will have finite resources to study the rest, so how do they allocate those resources?
Badly, usually. Case in point being the new deal for Drax, which is variously being spun as a good or bad thing. But new CfD strike price increased by £20/MWh to £158/MWh, which combined with ROC sales gives it a subsidy of around £2bn a year. Also curious how trying to limit it's hours will work in practice. It's easy to stockpile coal in heaps outside, but not so easy for wood given it doesn't like getting wet.
And btw, gas prices have dropped back from the Ukraine peak, so 'renewables' are back to being the price setters-
https://www.ofgem.gov.uk/energy-data-and-research/data-portal/wholesale-market-indicators
For example there are 1200 livestock and poultry farms in the UK that meet the US "Mega farm" criteria. IOW every one of them will have at least a hectare of roof that could be covered with a PV array, as well as a shedload of biomass generation (as well as acting as a possible hub site for anaerobic digestion to surrounding farms and fertiliser).
If that were economically viable, then it would have been done. Digestors are a thing to deal with 'waste' and convert that to gas for heating. Snag with roofs on livestock barns is they're not designed to deal with either the static or wind load of being covered in solar PV. So it's easier/cheaper to just cover a field with them instead. Assuming planning permission and regulations to try and export any surplus energy, costs of interconnection etc etc. So it'd be easier to just do an installation that could let farms be more self-sufficient and save on their own electricity costs.. but then the numbers often just don't add up.
IMHO between them that could put 1-2GW on the grid in years not the decades of Hinkley Point C.
In the context of 'Net Zero', decarbonisation and new bit-barns, 1GW is.. err.. chicken feed. RR SMRs would make more sense, but also the theory goes that Hinkley C is a bit of an outlier and covers 'FOAK' costs.. even though it's not the first NPP of that design to have been built. Theory goes that subsequent NPPs will be cheaper & faster to build, but will still include inevitable costs like lots of security staff to help unglue protestors blocking access roads. Germans did this faster/cheaper by just revving up disc cutters, removing a patch of road along with the protestors and then fining them the costs of patching the roads back up.
<...."Digestors are a thing to deal with 'waste' and convert that to gas for heating."....>
Not really, whilst there are a number of anaerobic digesters that use waste (primarily chicken litter or vegetable/food waste) there are a failry large (and expanding) number of AD plants for which a specific feedstock is grown - mostly maize or rye which is made into silage before being fed into the digestor. They produce methane gas which is then captured and used as the fuel for a gas turbine generator to produce electricity which is fed into the grid - as they can be started and stopped quickly they are brought online to provide power during high demand periods of the day.
There are a few AD plants which also utilise the excess heat to warm greenhouses built next to them - British Sugar have one such plant at Wissington (running mainly on the waste pulp from refining of sugar beet) producing power for the beet factory and the grid, and the heat for warming about 40 ares of greenhouses alongside which were initially used for producing tomatoes, but now produce a crop for use in the pharmaceutical industry. Dyson Farming have a similar set-up at Carrington fed by maize and rye silage which produces electricity (principally to feed into the grid) and heats a 26 acre greenhouse which produce strawberries (https://dysonfarming.com/strawberries/).
Not really, whilst there are a number of anaerobic digesters that use waste (primarily chicken litter or vegetable/food waste) there are a failry large (and expanding) number of AD plants for which a specific feedstock is grown - mostly maize or rye which is made into silage before being fed into the digestor. They produce methane gas which is then captured and used as the fuel for a gas turbine generator to produce electricity which is fed into the grid - as they can be started and stopped quickly they are brought online to provide power during high demand periods of the day.
This is where waste-to-energy can get silly, or dangerous. So I think there are 'good' uses, ie British Sugar, or the chicken/turkey waste plants, although NIMBYs object to some of those. But ones where crops are grown specifically to burn can just end up gaming subsidies, and wasting crop land that could otherwise be used to produce food for humans or animals. The EU had this problem when they promoted biofuels, and farmers promptly switched to growing crops for that instead of human consumption. Or in the US where a lot of crops are grown for ethanol conversion.
<......"For example there are 1200 livestock and poultry farms in the UK that meet the US "Mega farm" criteria. IOW every one of them will have at least a hectare of roof that could be covered with a PV array,...."....>
There are two potential problems there:
Firstly, most chicken sheds are of fairly lightweight construction, and the roof will not bear the combined weight of being covered in panels. To build chicken sheds with stronger roof and walls to take the weight would be costly, and would probably make such a project unviable.
Secondly, there is a fairly significant number of chicken sheds which are designed to be moved on skids - when one batch of chickens has been fattened and sent to slaughter, the shed is towed/winched from over the accumulated excrement and shavings mixture of the previous batch of birds, then pressure washed and thoroughly disinfected before being winched further across the field to a fresh, uncontaminated pice of ground before the next batch of birds is put into it. This is to minimise any disease carry-over, and obviously these sheds are built particularly light weight to make them easier to move, and obviously they would need to be stronger and heavier to carry solar panels, and moving them about every 6 or 8 weeks is going to present problems for the connection to the grid.
Apart from being a valuable agricultural fertiliser, a considerable amount of chicken dung is already used for power generation - power stations at Eye, Ely and Westfield all burn chicken litter to produce electricity, and there are several AD plants in operation or planned which use chicken litter to produce methane to drive a gas turbine powered generator.
Here are the stark NetZero choices:
1. Stop growth, reduce GDP or 2. Build "sustainable" power generation, reduce farming and increase power costs even more.
Notice how all governments are really pro datacenters and AI? That's to replace you. Think they want to reduce their wealth keeping non-productive people alive? Sarcopods will be subsidised.
This amount of power is required for Vehicle charging which has been known as an issue from day 1 f any green infrastructure ideas. Resources thatsuck wads of power around london will come second as this affects our 2035 targets agreed with Europe. (its actually 2050 but well done Boris et all promising something thatcant be done)
Except they'll not be using smart meters, the plan for rota disconnections (or, in other words, "too little power -> unplug some residential") will power things down at the substation level.
This will also only take place after paying increasingly large sums to industrial users willing to take payment in exchange for reducing their demand, e.g. a hospital powering up their generators and running off those instead for a handsome fee.
I don't believe your figures. The entire generating capacity of the UK is around 100GW and you're saying there are 400 further GW being requested in the London area? How does that even sound reasonable to you
They're both reasonable, and unreasonable, and all part of the 'AI' explosion along with Net Zero. So there's a pipeline of prospective DCs working through the planning stage. Securing power contracts is obviously an important part of those plans, and a long lead time item. So if all those DCs get planning consent, are built and occupied to max capacity, then they're all going to need power. Some probably won't ever be built and are speculative planning applications just to boost land value. Land with planning permission is a lot more valuable than without, especially if it involves 'green field' locations.
But that also has a huge cost implication, ie building a bunch of new NPPs to feed speculative bitbarns that might never get built, and who pays for all that.. hence why this is something of a hot topic given the cost and risk implications.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
