"There is a lot of work to do" is a terrible excuse to not start working.
'Close to impossible' for Europe to escape clutches of US hyperscalers
European organizations wanting to break free of American cloud operators may find their hopes dashed, according to industry analysts, for a number of reasons including a sheer lack of datacenter capacity. [I]f everybody in Europe was to move to the public cloud, it would still take about 20 years (based on historical build …
COMMENTS
-
-
-
Thursday 22nd May 2025 18:27 GMT Adair
'if climate activists had stopped complaining about nuclear being too slow to solve the problem'
Well that's self-serving bollocks for a start.
The fact that 'nuclear' has, and always has had, massive cost-effectiveness questions hanging over it, always seems to get swept under the carpet.
It was a case of: we've got this technology for making very noisy explosions, but it's epically expensive—we really need some way to make it cuddly.
IDEA: power too cheap to meter.
RESULT: epically expensive power to the masses, with added creativity to keep as much of the expense as possible as far away from the meters as possible.
-
Thursday 22nd May 2025 19:04 GMT Like a badger
"Well that's self-serving bollocks for a start."
Actually it's not. The UK has frittered enough on renewables in the past ten years (circa £350 billion) to have built ten twin EPRs that would have given a total output of 32 GW. We've got just over 5 GW of existing nuclear, so 32+5 is would meet total demand with plenty of reserve for 95% of days of the year. Instead we've got a fleet of wind and poor quality solar that are erratic and cost as much per GWh as nuclear, and we have rely on French exports and CCGT to keep the lights on.
-
Thursday 22nd May 2025 19:50 GMT Adair
We always get this, "if only they wouldn't fritter away the money on other things nuclear would save the day".
But still the 'Total Cost' of nuclear is quietly ignored. It is a fabulously expensive way to generate electricity, and not just in pure money terms. The level of technical skill is huge, all the way down the line, from start to finish. That, in itself is 'unsustainable' and adds fragility.
Any idea that nuclear is 'the answer' is just fantasy. It may well have a place, but by any reckoning it's not a sensible long term bet, certainly not for sustaining the bulk of the world's electricity needs.
We seem to have this odd idea that technological progress is an inevitable given. That things can only get better, despite all the historical evidence to the contrary, and the evidence playing out in front of our eyes that there's a decent chance we are really going to stuff up our civilization. Maybe we won't, but I wouldn't be counting on nuclear to save the day on the energy side of things.
-
Friday 23rd May 2025 18:29 GMT Excused Boots
You are correct in that nuclear power is expensive, mainly when the construction and especially the decommissioning costs are factored in, it is potentially dangerous, yes.
But on the other hand it does work, it provably works, once up and running it absolutely will provide a stable source of electricity irrespective of the time of day, whether the wind is blowing on not, without the massive environmental impact of tidal barrages, etc.
It actually does work and we know how to build and run them.
"The level of technical skill is huge, all the way down the line, from start to finish. That, in itself is 'unsustainable' and adds fragility."
Yes it does require a degree of technical skill, as does solar panel manufacture, wind turbine manufacture; oh I don't know maybe we invest in education and training in such skills, promote them?
"We seem to have this odd idea that technological progress is an inevitable given. That things can only get better, despite all the historical evidence to the contrary, and the evidence playing out in front of our eyes that there's a decent chance we are really going to stuff up our civilization."
Actually the historical record would suggest the opposite, yes nothing is inevitable, but based on what has happen in the past (yes with some setbacks, but overall), we do have to conclude that technological progress is very, very likely to happen.
If not nuclear then what do you suggest? All forms of renewable energy (with the exception of geothermal, which is only viable in certain areas), is unreliable, the sun isn't shining, the wind isn't blowing, yes tidal is reliable but not constant. All could be made to work with some kind of storage system, but the big problem is that such a system simply doesn't exist, there is no good, reliable and replicable* method of large scale electrical or energy storage which can work around the obvious issue with renewables. And it there were to be one invented tomorrow what's the bet that the technical skill required would make nuclear's requirements look like something that can be set up in an average garden shed?
Nuclear power might well be a bad choice, but it's better than all the others.
Oh and the 'too cheap to meter' claim, was always a vague marketing statement, was never going to be the case. It was, I believe, based on the assumed 'running costs' of a reactor; at the time they were comparing a number of trains full of coal trucks arriving every day, compared with a single lorry with a ton or so of Uranium every few months - yes ridiculously simplistic, but you can see where the idea might have come from?
* when I say replicable, I mean something like pumped water storage, ie use excess electricity to pump water up into a reservoir release it to generate power when required; only works in certain terrains, certain geographies, and what's the efficiency of such a system.
-
Saturday 24th May 2025 06:24 GMT Adair
'...we do have to conclude that technological progress is very, very likely to happen.'
But not as a continuous unbroken upward progress, as you acknowledge. That is a problem.
Three hundred years from now no one anywhere may have a clue, or anything like the technical capacity, to manage a rotting nuclear pile.
Thousands of rotting wind turbines are nothing much to worry about, likewise solar panels, but a nuclear legacy isn't such a happy gift to descendants.
But the real lie about nuclear is that it is cheap. It isn't even 'reassuringly expensive' in a good value for money kind of way. It is basically a massive rip-off of tax payers' money, based on hubris and greed.
A sad testimony to the dark side of 'technological progress'.
-
-
Sunday 25th May 2025 16:27 GMT Anonymous Coward
It's why I like the whole Thorium idea.
It's not just that it's a lot safer (read: cheaper to build), it's also quicker to build, which substantially lessens the capitalisation barrier. The biggest problem with nuclear plants is the long gap between drawing the capital and starting to get a relatively uncertain return, but the moment someone manages to apply SMR principles to Thorium based plants that problem is history. The delta between running costs of traditional nuclear reactors and a Th based plant should do the rest. (even though its fuel cycle is more complex, the fuel itself is far easier to obtain).
-
Monday 26th May 2025 10:37 GMT Zolko
It's why I like the whole Thorium idea
yeah, but you can't build nuclear bombs with its residuals. That's why Uranium was chosen during WW2 and the Manhattan project : because atomic bombs. Which only shows that those most opposed to nuclear energy are actually clueless about nuclear energy :
@Adair
we've got this technology for making very noisy explosions, but it's epically expensive—we really need some way to make it cuddly
no, it's the exact opposite. Nuclear was chosen for bombs, civil energy is a byproduct. As opposed to Thorium-based reactors that are only civilian, and were thus rejected. As a side-note : China has just finished a Thorium-based nuclear plant in the Gobi desert. I think it's still experimental though.
-
Tuesday 27th May 2025 10:55 GMT Anonymous Coward
China has just finished a Thorium-based nuclear plant in the Gobi desert. I think it's still experimental though.
That test has actually finished (nice idea to place a plant in the one place where you do NOT have access to water, essential for any other type of nuclear reactor). AFAIK they're heading for commercial production now, I think I saw 2026 mentioned somewhere. TBH, once you have proven it works at production level size (and they've been at it for a long time) there isn't much more to do other than to prepare for independent validation of the concept and set up mass production. Not trivial, of course, but also not a major challenge in China.
If they can add turbines running on supercritical CO2 to this (which happen to operate at around the same 700ºC as a Thorium reactor) they could get a stupidly efficient plant together which could be containerised which would drop the build price to almost irresistable levels.
-
-
-
Friday 23rd May 2025 14:20 GMT Charlie Clark
I'd like to see you get those plants built in that time and on budget! Renewables without feed-in tariffs are now more competitive than nuclear in many places, most of the time and it has a great advantage of providing power now. What we have neglected to work on are grid stabilisation and long-term backup/swing capacity for when we have too little or too much renewables. If we'd spent a fraction of the public money that has been spent on nuclear on these, we might have more practicable solutions than what's currently available.
-
-
Friday 23rd May 2025 04:41 GMT Valheru
mmmm No. The accounting is transparent in many cases and the costs are not outsized for use of Nuclear power. Swiss nuclear power generation makes a nice small example to have a look at.
The cost of nuclear electricity in Switzerland has been relatively low, ranging between 4 and 7 centimes per kilowatt hour.
If air pollution is factored in, it is far safer than combustion based power. High bound estimates of Chernoble from biased anti-nuclear sources come in around 4-5000 eventual deaths.
Air pollution contributes to approximately 7-8 million deaths annually worldwide from conservative sources.
-
Friday 23rd May 2025 13:58 GMT Anonymous Coward
> mmmm No. The accounting is transparent in many cases and the costs are not outsized for use of Nuclear power.
What you miss is that the costs of nuclear power are for a big part carried by society, not operators as it's the case for other industries.
In 2011 a study looking at the true costs of nuclear found that if the operator had to carry the full risk then the cost of insurance for a single nuclear plant would be around 72 Billion EUR per year (2011 money).
https://www-manager--magazin-de.translate.goog/finanzen/versicherungen/a-761954.html?_x_tr_sl=de&_x_tr_tl=en&_x_tr_hl=de
The reason it doesn't cost anywhere near 72 Billion EUR a year to insure a nuclear plant is because the majority of risks are carried by the public, which means nuclear power benefits from a huge subsidy which is not available to wind, solar or even oil and gas.
The other issue is of course that on average it takes 30 years for a nuclear power plant from laying the first brick to generating power, which is a very long time during which the technological landscape, especially for green energy, is changing. Already today the viability of nuclear power is in doubt (and this despite the massive subsidies!), and in 30 years when a new plant started today finally comes online it will be much more expensive to compete with then current forms of power generation.
-
-
Sunday 25th May 2025 08:58 GMT anothercynic
That's only the case (currently) because we chose to use uranium + plutonium as fuels, whereas (apparently) thorium would be a much better bet because its fission products turns into isotopes that are stable and non-radioactive a lot sooner. There was a graph somewhere a while ago showing what uranium turned into and that in particular was not pretty because the actinides created by uranium were highly radioactive for a long long *long* time.
The idea to use thorium to burn through *all* actinides to end with something like a non-radioactive isotope of lead was cooked up before the large uranium deposits were found in Australia, Namibia, and Canada in the Western bloc, and Russia, Kazakhstan and Uzbekistan in the Eastern bloc, and everyone ran for the easy option. Niger became a huge source for France, China had its own resources. So yes... running for the cheap option has cost the planet dearly (although in terms of volume, the amount of radioactive materials, including the irradiated metal structures in decommissioned reactors, is a damn sight less than the huge amounts of spoil and ash generated by coal since (and the ash, ironically, is also radioactive to a degree because as much as carbon has burnt out of the material, uranium and other actinides that co-exist in coal have not and have been concentrated).
China is looking at thorium now to see what difference that makes. Maybe it will be the case that the rest of the world looks at China and decides that maybe they are doing it right. Who knows. If given the choice between coal and nuclear, I'll pick the latter any day of the week, mostly from the daily pollution standpoint. And if it comes to base generation, I think that nuclear is a viable option, although CCGT-powered stations can ramp up and down a lot faster than nuclear can, but gas will not be endless either, and as supply dwindles, countries will fight each other over it (case in point, the scramble to secure gas supplies post-Ukraine-invation).
-
-
-
-
-
Friday 23rd May 2025 22:23 GMT Roland6
I didn’t realise the Tories were climate activists, who since the early 1990s kept kicking the ball down the road, for some future government to deal with. David Cameron did try to do something, but everyone complained about the price per Mwh, which now seems quite reasonable.
Whilst “the Greens” did campaign again new nuclear, it was the Tories and briefly Labour, who determined whether any new. Nuclear plans actually got prepared and funded.
Similarly in the US, where you would be hard pushed to call Trump a climate activist with his fossil fuels first policy.
-
-
Friday 23rd May 2025 20:44 GMT Rol
How's about a compromise?
European customers use European cloud services, who encrypt our data and store it with the cheapest cloud providers on the planet.
Enacting Uncle Sam's Patriot Act would be useless,
Okay, yes the service on a service will add cost overheads, but a large enough European service would surely have some leverage on price/GB.
Another advantage would be the common interface presented to users by the European service, regardless of which cloud provider is holding the data, which could be seamlessly moved to another storage provider as and when required.
Slowly the EU provider can onboard its own local storage as demand increases.
Like the ingredients on a food label, I want to know where my data is being stored, and if it's on an American server, and it is readily readable by the NSA, then it isn't covered by GDPR, so I'm not buying into it.
-
Saturday 24th May 2025 05:42 GMT Groo The Wanderer - A Canuck
You know, the people who are claiming it's "impossible" sound like the same crowd who said it was "impossible" for China to develop a home-grown technology and industry base... head in the sand fools who think the only people with functioning brain cells live in the US. *Bwahahahahahahahaha!*
-
Monday 26th May 2025 08:39 GMT cookiecutter
Yeah I hate this argument..getting away from the fact that if you can't figure out how to use another datacentre provider than the one you faked the certifications for ; then you shouldn't be in IT.
But now you get it about apple bringing back manufacturing to the USA. They spent $55 billion per YEAR to train the Chinese for years! Why can't US & EU firms do the same now, to ensure that in a decade or so, all that stuff is back.
Anyone who uses this line is just a lazy bastard
-
-
Thursday 22nd May 2025 11:57 GMT may_i
Not a realistic option
"selecting a sovereign cloud option from a US vendor" is not a realistic option. Your data is still subject to being silently slurped by the US authorities - US Cloud Act. Your infrastructure on that "sovereign cloud" is still run by a US company and can be turned off at will.
People need to stop making excuses and get moving!
-
Saturday 24th May 2025 01:58 GMT Anonymous Coward
Re: Not a realistic option
There needs to be a great deal more reminding European businesses, and putting of pressure onto them, that using US cloud services with any customer data: is not legal. New companies must be prevented, and reminded very strongly, of this. Existing businesses (and government(s/their contractors)) need to have increasing pressure applied.
Processing billing, storing users' names and e-mails, access and usage logs - all data that is subject to protection by European law, is it not?
-
-
Saturday 24th May 2025 14:02 GMT lamp
Re: Not a realistic option
I agree - get going with an open source cloud e.g. OpenStack, in a data centre in your country. We use OpenStack in our Vault Cloud systems in Australia. The American government can't access any of our information as none of the end to end technology is owned by any U.S. vendor.
-
-
Thursday 22nd May 2025 12:09 GMT elsergiovolador
Brown envelopes
Yes, the European Commission, boldly declaring that building a cloud infrastructure free from US control is "unrealistic". Not because it's technologically impossible - no, no - but because apparently spinning up servers and writing orchestration code is now treated like nuclear fusion. Never mind that any competent sysadmin could stand up a modest AWS clone with open source tools and a few racks - it’s just so hard, you see. Too daunting. Might as well give up and keep paying Bezos tax.
And the “industry experts” lining up to echo this narrative? Shocking. It’s almost as if they work for or benefit from the same US hyperscalers they claim are inescapable. Billions in vendor lock-in contracts, a whole continent’s data dependency, but yes, let's pretend this is about “capacity” and “skills shortages,” not vested interests.
Meanwhile, European providers like OVH, Hetzner, and Scaleway exist in reality, not hypotheticals - but they’re casually brushed aside as too small or too inconvenient. Translation: perhaps they don’t fund enough golf retreats.
This isn’t a strategic failure. It’s a surrender dressed as pragmatism. The Commission isn’t just unserious - it’s complicit.
-
Friday 23rd May 2025 10:50 GMT veti
Re: Brown envelopes
If you seriously think you know anyone who could "stand up a modest AWS clone with open source tools and a few racks", then for the gods' sake get on with it. There's never going to be a better opening in that market than right now. In 20 years you could be, maybe not quite a Bezos or a Musk, but at least an Ellison.
Except you know - don't you? - that it's not that simple. That "modest" in your statement is carrying a helluva lot more weight than it's normally rated for. Your "AWS clone" may, once you've blown out your current imagined budget by a factor of 20 or so, be capable of hosting and serving data at reasonable speeds and standards, but every actual AWS customer relies on other functionality that you haven't even thought of, and it's not something you can put together overnight.
I've worked in a bright young startup where the founders had this fantastic idea for making networks work better. They had everything you could reasonably ask for - a hugely talented team of engineers, enough venture capital to keep them afloat for years, managers they recruited from some of the biggest name companies in the industry. It was a fantastic place to work, I enjoyed it hugely. But somehow, after five years of development, the actual goal was still just as far off as it had been at the start. That's because these "big picture" types think they can just wave their hands and sketch out a solution, and the problem is "essentially" solved, the rest is "just" legwork. But they have absolutely no conception of how fucking hard it is to actually get things working.
-
Friday 23rd May 2025 12:28 GMT elsergiovolador
Re: Brown envelopes
Yes, AWS has thousands of services. No, no one needs to clone all of them to be viable. Most customers use a fraction - storage, compute, networking, some managed DB. That’s not magic. That’s Linux, KVM, Ceph, Kubernetes - all open-source, all mature. OVH, Hetzner, Scaleway, and others already exist. Are they AWS? No. Are they real options? Yes. Are they improving? Constantly.
The real issue isn’t “it’s too hard.” It’s that too many people in comfy roles at big consultancies and vendor-adjacent startups are financially and professionally entangled with US hyperscalers - so they dress up inertia and rent-seeking as technical inevitability.
Also, sorry about your failed startup, but that’s not proof of universal impossibility. It just means you spent five years proving your team wasn’t as good as you thought.
-
Monday 26th May 2025 10:48 GMT Zolko
Re: Brown envelopes
I've worked in a bright young startup ...
... that failed. So why should anyone listen to your "advice" ? Of course I understand that you "claim" it was not your fault, but bragging about a failure and patronizing other people doesn't sound very convincing. In case you didn't know, AWS didn't start as AWS. Jeff Besos didn't ever imagine creating AWS, it is a byproduct (of Amazon's computer infrastructure designed for the 2 months before Xmas, and sitting idle the rest of the year)
I might take this more personally than necessary because yes, we're trying to create something in this area also.
-
-
-
Sunday 25th May 2025 13:48 GMT Adair
Re: Well
'Lots if Eurocrats are deeply corrupt'
Either that is self-serving bullshit, or a description of bureaucrats in general because, you know, human beings.
So, quit your whining and get out there and do something that is both useful and good. At least you will keep your integrity, and may help others to do the same.
-
-
Sunday 25th May 2025 09:45 GMT Pete Sdev
Re: Brown envelopes
European providers like OVH, Hetzner, and Scaleway exist in reality, not hypotheticals - but they’re casually brushed aside as too small or too inconvenient.
While I don't in principle disagree I'd point out some problems with the smaller homegrown providers.
1. Cost. The big US providers have economy of scale.
2. Services offered aren't equivalent, e.g. Database as a service.
3. Functionality isn't equivalent. E.g. Hetzner (which we use at $WORK) doesn't currently have autoscaling for compute instances, which AWS has had for donkey's years.
4. API incompatibility
5. Lack of current capacity.
6. For European companies that trade internationally, the lack of US located resources.
If you're moving to the "cloud", and you have the in-house chops to say run your own DB servers, then a local provider may well be viable. For a company already using AWS/Google/Azure etc. migration to a local provider can likely be unfeasible at this time.
Some of the issues above are of a chicken and egg nature.
-
Thursday 22nd May 2025 13:28 GMT Guy de Loimbard
Awesome
Commentary here fellow Reg Readers.
I was just thinking along the same lines as all of the OPs above me.
We've got to stop this BS Mantra that everyone is processing huge amounts of Data that require all this "scalability".
You're dam right, just because it seems too hard, doesn't mean you shouldn't get on the journey and re-evaluate your cloud requirements and position.
If anyone realistically looked at both their internal and external data requirements, they'd likely be surprised that they don't need all this scalability and, without doubt, you are paying for features, or tiered support models that are not needed.
I've worked across a number of industries and I've yet to see any behemoth models that require pure cloud at any cost, or size or.....
Most entities on the cloud, or on that journey have been sold the lie that it will ease all of your IT worries and costs.
I'm sure there are examples of excellence to be seen, but I'm not seeing many new Meta's/Amazon's or Googles popping up
-
-
-
Friday 23rd May 2025 19:25 GMT elsergiovolador
Re: Awesome
You cannot build a product without capital. Proof of concept probably, MVP maybe. Very few can monetise their MVP and reinvest into full product development.
Reality is that you need capital and because of how tax system is structured (ensuring working class cannot amass capital) - grant, bank, VC, angel investor are pretty much your only options.
-
Friday 23rd May 2025 19:56 GMT fg_swe
IT Airbus Financing
Indeed financing is a key element of any IT startup. The American system of government and California venture capital did this brilliantly.
Airbus similarly came only about due to dozens of billions of government financing.
Logically, the IT Airbus would provide this finance, management, production, support, legal etc capability. Microsoft would deploy their full bullsh1t force against it, naturally.
To make this happen, the leftist idiots must be first flushed out of major European governments. All they know is dimwitted inflation and consumer spending.
-
-
-
Saturday 24th May 2025 10:15 GMT Anonymous Coward
Re: Awesome
That will surely change if people simply can't rely on (or maybe can't legally use) American services any more.
Most of these big American companies are not irreplaceable - they are big due to finance and business reasons, not technical ones.
And like Premier Leauge football teams, they can afford to buy the best players.
-
-
Friday 23rd May 2025 14:28 GMT ecofeco
Re: Awesome
Look, creating databases that detail which hand everyone in the world wipes their arse with and picks their nose with is of vital importance! Let alone how to exploit the teens with the latest curated rebellious fashion trends and secret jargon!
And as if that is not enough, pretty sure extra-EXTRA layers of new software protocols will need the remainder of all that capacity.
And of course...--------------------------------------------->>>>>>>>>>>>
-
-
-
Friday 23rd May 2025 14:34 GMT ecofeco
Re: "would rather get a root canal than start over"
Greybeards remember time before Total Enshitification. New is not always better. They've seen the endless vaporware parade wars.
But also, you are correct that far too many greybeards are also pathetically in love with Win7. A stunningly, sadly disappointing, number.
And let's not forget the bean counters having an outsized role in all of it as well.
-
Friday 23rd May 2025 14:49 GMT Shadowlight
Re: "would rather get a root canal than start over"
This graybeard learns something new every day, and isn't afraid of working hard.
The reason why we don't like systemd is because it doesn't add something we need on servers, It's nice though for laptops.
The only thing I'm afraid of is that systemd will become a Microsoft stranglehold within Linux, the infamous Embrace, Enhance and Extinguish.
-
Friday 23rd May 2025 19:35 GMT Ken Hagan
Re: "would rather get a root canal than start over"
The greybeards are actually mentioned in the article, though not by that name. They are the people who knew how to run stuff in-house, who were sacked to provide the short-term saving that made the cloud migration look attractive to the bean counters.
-
Saturday 24th May 2025 23:01 GMT Roo
Re: "would rather get a root canal than start over"
Weirdly, in the Linux/UNIX space it's the Greybeards that adapt fastest to this idea of bootstrapping stuff with images and tailoring them for the job at hand... Who diagnose the underlying faults in the VM infrastructure that is crippling the code etc... Meanwhile on the other side of the aisle we've got greenbeards(?) who don't understand why they can't run the Java code on Linux Virtual Machines using a Windows jvm.exe. I kid ye not.
Meanwhile the manglement are busying trying to extract their bollocks from the vice of their Cloud vendor having discovered that the Cloud is costing 3x that of on-prem bare metal (2x on-prem cloud), I can bet you can guess what their solution to the problem is: Run those highly-optimized compute intensive batch jobs on 1/3rd the capacity... Throughout all this the greybeard backed teams supporting all this guff have managed to migrate the apps from bare-metal to off-prem cloud, on-prem cloud, and back to bare-metal and mixtures there of.
The greenbeards supporting the cloud stuff are struggling - they keep running into odd-ball VM related bugs and screwing up the configuration while they're at it. It would be nice if they could understand why migrating a VM hosting a HPC app that is eating 100% of the host from one hypervisor to another is a really shit idea for example - their solution is to ask apps to leave 10% of the host unused - which is comical given that the bare metal runs happily at 100% utilization - and the whole argument for migrating stuff to VMs was to improve utilization of hardware...
-
-
-
Friday 23rd May 2025 13:29 GMT Grindslow_knoll
It depends, if you budget in the need for failover, then no, because you would have planned for multiple redundancies spread over regions and companies (or local).
But that is non-trivial because companies do not tend to favor interoperable standards, so once you're Azure, it's not easy to keep Google Cloud compatible flows going.
And now you need staff that can work in both ecosystems, and future failure is not what brings bonuses, but short term saving does.
-
-
Friday 23rd May 2025 13:33 GMT Grindslow_knoll
To the best of my knowledge, the big tech datacenters have EU equivalents for reasons of latency, so it's not an infrastructure problem but one of ownership, which would be a legal issue when push comes to shove.
It's relatively easy to motivate big tech, just double or triple costs of water and energy until your objectives are achieved (sovereignty, tax, ...).
The EEA market is too big for them to say no, and the true footprint of the compute is ignored too much anyways.
The current drought-like conditions in NW Europe illustrate that there's no more space of consuming water for very little societal benefit.
-
Friday 23rd May 2025 20:09 GMT fg_swe
Really ?
USG retains the authority to coerce ANY US company to do whatever USG pleases.
A MSFT data center in Europe must comply, or US based management can and will be thrown in jail. See the former CEO of Qwest, he thought complying was optional - jailed 6 years.
So, sovereignty cannot happen with US controlled corporations and technologies.
-
-
Friday 23rd May 2025 21:31 GMT SherbertDipper
Grey beard revolution, Open source hyperscaling.
Hmm, as one of those Grey Beards who actually responded (On behalf of a FTSE 50 company) to the UK government investigation in to the Hyperscalers and the potential for unfair competition I was a little disappointed (to say the least) in the reported outcomes. Apparently there is no anti-competitive practice going it which we all know is pretty much BS as when the Hyperscalers have their claws in to you they continually push cheaper, and faster services that ensure lock in (You can get out, but it comes at a cost, and often involves leveraging open source which unfortunately a lot of UK execs are allergic to).
To summarize, at the moment using Hyperscalers is pretty much like the Hotel California syndrome, once you enter you will never leave, or at least you will find it very bloody difficult and proposing an alternative to a company board who are heavily audited and who have non-exec directors who are almost certainly heavily influenced by hyperscalers and their ilk is likely to fall on deaf ears. In addition to this as we all know senior executives in most businesses are pretty easily influenced (In some cases 100% reliant and lack any form of self direction - they didn't get their massive salaries, bonuses and share packages, and free tickets to Wimbledon by rocking the boat) by consultancies who are often partner organizations of the Hyperscalers. So all the time the thought whisperers are continuing to tell them that the Cloud at scale is best for them or AI is the future, you shouldn't expect any freedom of thought, or anything to change. Also worth noting that the costs of building a full on alternative to the current Hyperscalers (A Euro Cloud) is likely to be prohibitive.
Where I can see a potential change movement is through the start up community, and I could see a groundswell movement coming from open-source to actually build something different. No doubt that you need the compute power, the actual electricity, and cooling when you are building data centers at scale, but perhaps we should be thinking about this problem in a different way. As an example if you look at the current proposals for Modular Nuclear reactors which are distributed and localized, these have far less of a draw on the local environment, deliver resilience and drop power on to a National grid at the point of need.. Why is it not possible to take the same approach to data centers and build small local franchised facilities, which require lower investment, and then link these together in a high performance network providing both resilience and scalability. You do need compute, and storage, but the upfront investment is at a lower level and more attractive to smaller investors, with the ability to link all of this together to provide services at scale as the network of "franchised" modular data centers grows.
This would be a much more attractive proposition and would likely gain govt investment in terms of bringing local jobs to regions. As earlier posters have pointed out much of the software and tools to actually build services in these data centers are available in the shape of open source providers, and the networks to connect this are already available, so this shouldn't be beyond the technology community. Clearly there would need to be agreement in terms of scalability and capacity, as well as some pretty strict rules on waste (which the tech industry is pretty crap at), but this could be a way of building an alternative to the hyperscalers over time. The other advantage is that any profits would be taxable locally, and not shifted to Luxembourg and onwards to the land of the Orange Chimpanzee.
For me it's certainly time to think differently, gut feel is that in the world we live in data sovereignty is going to become a real thing (not just a politicians pipe-dream), and protecting Intellectual property rights is going to become harder and harder to do in a world where data is eviscerated from it's owners by the use of AI and used and abused by hyperscalers and their cohorts.
Clearly this is just my opinion, but I started employment in 1987 working on centralized Mainframe services that were the domain of 2 - 3 suppliers, and having seen that democratized by multiple companies through the delivery of distributed systems, and the advent of the world wide web, it is extremely worrying that we are heading back in to a world where the majority of the compute power, and data is basically sitting with 3 or 4 major hyperscalers. With the advent of AI this is massively open to abuse, and there isn't a government in the world who can put this badly behaved child to bed without some radial thinking. As things stand with polarized politics throughout Europe, I find it very hard to believe that any centralized government organisation or collective is actually capable of delivering the change required. So perhaps it's down to the technology community themselves to do something about this now, or at least to lobby shoulder to shoulder with the creative industry to highlight the potential abuse of power that the hyperscalers will undoubtedly and probably insidiously execute.
Good luck all, would the last System programmer switch the lights off on the way out of the door.
-
Tuesday 27th May 2025 07:41 GMT JohnMay
Don't forget renting space in a data center (often called **colocation**) can replace cloud for businesses that:
* Have a stable, predictable workload.
* Require absolute control over their hardware and software stack.
* Have strict compliance needs that mandate owning the physical hardware.
* Possess the internal IT expertise and staff to manage all aspects of their infrastructure.
* Have significant capital to invest upfront.
* Are very sensitive to long-term operational costs that might become high with cloud egress fees or consistent high usage.
**Hybrid Approaches:**
Many organizations today adopt a **hybrid cloud** strategy, using colocation for core, stable workloads and leveraging public cloud for flexible, burstable, or development/testing environments. This allows them to get the best of both worlds.
(Summarised from Gemini)
-
Tuesday 27th May 2025 07:45 GMT GeorgieBear
ya'll went down the nuclear rabbit hole... nothing to say bout European clouds?
acc to AI,
AWS offers eight data warehouse locations in Europe: Ireland, Frankfurt, London, Paris, Stockholm, Milan, Zurich, and Spain.
Azure has regions in Ireland (North Europe), the Netherlands (West Europe), Germany (Germany Central, Germany North, Germany Northeast, Germany West Central), and the UK (UK South, UK West)
Google's European data warehouses are located in: the Netherlands (Eemshaven, Middenmeer, Winschoten), Belgium (St. Ghislain), Finland (Hamina), Germany (Hanau), Ireland (Dublin), Norway (Skien, in development) and in the United Kingdom (Waltham Cross, in development)
why are these not considered "European"? then worst case, local governments could take them over if considered national security...
-
Tuesday 27th May 2025 07:46 GMT Everoui
Thinktank/BI/research group reporting
It would be useful if news reports (in general, not just El Reg) provided a bit of background on organisations like Informa, IDC, Synergy Research Group and Omnia. I'm not saying a full background check on their funding sources, strategic partners or CVs of key employees is necessary, but links to their websites would be handy.