The issue is not lack of european providers. The issue is almost natural lobby from the likes of Microsoft and Amazon.
All the unis are infected with former corpo programmers who push their preferred employer stacks. Students get free credits to use on Azure. Some of them become gov employees or consultants. You end up with all sides agreeing Microsoft Azure is the most trustworthy solution.
Everything gov in my country seems to be built with asp.net to point where its often requirement in the contract.
Lets start with erasing all the proprietary tech taught at universities. Microsoft can do their education. Universities dont have to do it for them for free.
Also Microsoft consultants are everywhere, and management consultants like McKinsey usually work closely with Microsoft, and many corporations are already locked in.
In many ways its kind of really scary how so much of the digital infrastructure even of key enterprises and entire states is owned and operated by a private company.
It didnt use to be open-source. Its closely tied to MS which can do whatever about it. It will always have better support on MS and Azure. It still MS stack and if you want to get out of that you have bazilion better more free options.
I didn’t see any mention of EU’s last attempt to build a sovereign cloud: Gaia-X. I think that didn’t really go anywhere due to disagreements over technical direction. Much like Quaero.
Don’t get me wrong, I hope it goes better this time and the author acknowledges the many pitfalls of such a complex undertaking.
" Innovation Mindset: Willingness to explore new technologies – especially related to trust frameworks and associated standards –, staying ahead (no kidding!) of main-stream industry trends (This is a real requirement, not ChatGPT – just look into our GitLab repos!)."
I checked their public GitLab and it's puny. Moreover, I suspect based on the use of bold and bullet points that the job description itself went through ChatGPT.
It's more or less covert subsidies for the usual suspects. In Germany that means the like of Fraunhofer-Gesellschaft, Atos, Bosch, Deutsche Telekom, SAP, BMW and Siemens get some millions (and by "some" I mean around 200 million euro). If there's no outcome, like in the GAIA-X case, it's just bad luck for the taxpayer.
Disagreements over direction should be solved by making separate orgs/initiatives to go test each approach.
Better than funding “one true initiative” and force people to work together to keep the funding.
The VC funding model succeeds in experimental domains by letting many experiments run at once. Kind of a form of Darwinism. How to do that with funded initiatives is a really interesting challenge.
The impression I have, but I may be wrong, is that investors in Europe want a return on every investment. Failure is not an option. Which kind of choke innovation.
Without protection, a EU hyperscaler will never emerge. It's not that any hypothetical future EU hyperscaler would be an inherently less competitive. It's just plain old first mover advantage. Cloud computing is a VERY sticky good, and once folks got on AWS/Azure no matter how good your product is there's no way larger customers are going to migrate off to some "risky" startup (thus keeping scale small).
Source: I worked on a sovereign cloud PaaS in Germany. Admittedly the project was a disaster, hiring talent in the infrastructure space here is really tough. Probably mostly due to a lack of opportunities to acquire experience domestically and noncompetitive wages limiting migration of those with experience.
Gaia-X is a total dumpster fire though, perfect example of what's wrong with the EU. All the money was spent paying bureaucrats, consults, and standards committee folks to architecture astronaut and design by committee sovereign cloud "standards" bullshit bingo. None of the money went to the engineers.
We should really copy the Chinese approach here, straight out clone a subset of the best/most popular parts of AWS/Azure/GCP. I'm open to contracting opportunities in this space BTW (email in profile).
I don't think bureaucrats can build a successful cloud by enforcing laws, adding even more regulations, making policies and using subsidies.
This is not how the market likes to work. There has to be a demand and the free market forces will build a supply. Governments can help by reducing regulation and bureaucracy, reducing taxes.
Policies are often used to create markets or - like in this case is propose - to create that demand. The article makes the argument that free market failed to deliver in this sector in Europe.
One of the explanations is that once you lag behind competitors, if policies don't force to value specific parameters that can be fulfilled only by other competitors, no competitors will join the market because nobody is going to choose them. So the author argues for example to impose regulation for public tenders such as "must be subject to only EU laws". This creates a demand, which is not currently matching any offer in the market and creates market incentives for new players to compete.
So regulations can absolutely work where "free" market fails (quotes because even the big 3 are/were pumped full of money by government/defense contracts).
> There has to be a demand and the free market forces will build a supply
Not in all cases it doesn't.
If the return on investment isn't there or if companies are unable to fulfil the demand then the market has failed. Recent example of this has been EVs where regulation needed to step in to force the market forward.
>If the return on investment isn't there or if companies are unable to fulfil the demand then the market has failed.
If there's no demand, there's no problem. Also there's no money to be made. If there's demand there will always be companies to supply that demand. Of course, it takes a time.
Politics shouldn't dictate to economy.
>Recent example of this has been EVs where regulation needed to step in to force the market forward
Then why should people be forced to use EVs instead of whatever they like? That is not freedom.
> If there's demand there will always be companies to supply that demand.
I want to walk around town without cameras tracking me. Can companies supply my demand?
I want PFAS and microplastics out of the evironment. Can companies supply my demand?
I want net-zero CO2 emissions in all the products I buy. Can companies supply my demand? How do I trust they are not green-washing? Do I depend on even more companies? How do I trust those companies?
Sure seems simpler to have the government enforce privacy and pollution laws.
> Politics shouldn't dictate to economy.
Which means you think William Wilberforce's most famous work was completely wrong.
> Then why should people be forced to use EVs instead of whatever they like?
You should consider how government-subsidized oil production and distribution has forced you to use fossil fuel cars instead of whatever you like.
You should consider even more how government-subsidized planning forces you to use a car in the first place, and often closed off alternative solutions like walking, bike, and mass transit.
Because we have wider goals as a society that go beyond absolute freedom. However things like EV incentives don't make you less free, you can still buy whatever you want.
Free markets like to make money for capital holders, and tend towards a monopoly or oligopoly as that maximizes profit extraction.
Name a free market where there is an economy of scale and network effect, which does not tend towards monopoly or oligopoly.
More fundamentally, if (say) Germany wants to move all of its government operations off of US-based IT, should it wait until the market has provided a solution? Or should it do like all governments have always done, and change the marketplace?
Nor is this unique to governments. Some of my clients require me to agree to their ethical practices, like "no child labor." They are using their purchasing power to change the marketplace.
Why do the countries with higher taxes and government regulations and policies over health care have better overall health outcomes than those with lower taxes and regulations?
There is a demand for clean beaches. Where is the free market for pollution control?
The topic we're discussing would seem to undermine your claim, as cloud is neither a monopoly nor an oligopoly despite having scale and network effects.
Could you explain then how so many European governments ended up one of a handful of US clouds (Amazon, Microsoft and Google)?
Almost reason I can think of comes down to scaling or network effects.
I say "almost" because I've thought of a couple more:
- it's already a Microsoft shop and it's hard to change? Lock-in effect.
- you never get fired for choosing IBM, err, AWS? I don't have a name for this effect, though it seems related to scaling.
Note that "monopoly" does not mean there is only a single provider. "De Beers controlled 80% to 85% of rough diamond distribution and was considered a monopoly" during the 20th century. https://en.wikipedia.org/wiki/De_Beers
> In economics, a monopoly is a single seller. In law, a monopoly is a business entity that has significant market power, that is, the power to charge overly high prices, which is associated with unfair price raises.
1. Invested massively. In the case of Microsoft and Google, ahead of demand.
2. Have lots of developers already familiar with their stacks, especially Microsoft via Windows and .NET which are mainstays of nearly all European non-tech businesses.
3. Indeed, have lots of "never get fired" effect; the word you're looking for is brand equity. Especially the case with non-technical execs already familiar with their consumer products, and it's often non-technical execs who make the decisions about which brand they're going with. Everyone knows MS Office and likes it, everyone knows Amazon and buys from it, everyone knows Google and searches with it. Their products work well, so their clouds must also work well. In government especially the decisions are always non technical so this matters a lot.
Monopoly does actually mean only a single provider: that's the definition and your Wikipedia link opens by stating exactly that in its first sentence. Sometimes the term is stretched or abused by governments. That happens because everyone across the political spectrum agrees to anti-trust laws in principle, and then the left can use them as a stick to beat companies with if only they can twist the definition of monopoly far enough. This so-called legal definition isn't a definition at all, it's a purely political construct which is why it includes ambiguous terms like significant, unfair, overly high etc.
Even using this non-definition though, cloud providers don't qualify. They're actually famous for not raising prices. Raising prices a lot on existing customers is a typical sign that a company has significant market power.
"Invested massively" because the could get scaling advantages. Both companies built off of existing OS and ad tech monopolies which should have been regulated long ago.
"Lots of developers already familiar with their stacks" is a network effect.
> Monopoly does actually mean only a single provider
Is a tomato a fruit?
I was very clear to give my definition of "monopoly", to include "oligarchy" for those who wanted to be picky, to point to how others describe De Beers as a "monopoly", consistent with my use, and to show how even the Wikipedia page on "monopoly" is not disagreeing with me.
You'll note that same Wikipedia page for "monopoly" lists De Beers and Standard Oil as monopolies, when they were not sole providers of diamonds or petroleum, respectively.
> They're actually famous for not raising prices.
Right. You are using Bork's 1970s influential re-orientation of monopoly power away from general anti-competitive behavior to focus only on "consumer welfare". I agree with Lina Khan (starting with "Amazon's Antitrust Paradox") and others, who argue that Bork's standard is too narrow.
Or, to use your terms, Bork's right-wing viewpoints twisted established law to allow abusive behavior by powerful companies, so long as consumer pricing was not affected. V.P. Vance likes Khan because he wants the right to use the stick to beat companies which don't support the right. This is only possible because all of the big tech companies already abuse their market power, because Bork gave them the freedom to do so.
> This so-called legal definition isn't a definition at all, it's a purely political construct
All laws are political constructs. What is "ownership"? Do you own the water which falls on the land you own? What is a "reasonable search"? What is "cruel and unusual punishment"?
Good law is precise and clear, so everyone knows how to follow it. The meaning of the word "ownership", to take your example, is precisely defined via legislation and case law.
Anti-trust law is neither precise nor clear which is why - and we are both in agreement on this point - that right now both left and right are happy to pick it up and use it to "beat companies" for political reasons. They can't do that with property law because property law is well defined.
The big advantage of the Bork formulation is exactly avoidance of this outcome. By requiring there to be actual evidence that harm was done, it hugely reduces the ability to abuse these laws for political ends and keeps them focused on behaviors that are actually detrimental to the market.
Ownership is not precise and clear. There are always new cases to figure out what it means.
Like, if there are two pieces of public land, which meet at a corner, and two pieces of private land on either side of the corner, then are you trespassing on private land if you step from public land to public land across the corner?
I think all of the big IT companies deserve to be beat, broken up, and strict laws placed in future power. That's different than using the cudgel to only go after the ones which disagree with whoever is in power.
> for political reasons
My view is non-partisan. If Amazon were to become a worker-owned co-op tomorrow, infused by the spirit of Joe Hill, I would still want it broken up.
> They can't do that with property law because property law is well defined.
The US has some very big cudgels already.
The government can pass laws changing what property law means. The government has very broad power to seize land, and exercise eminent domain. The government can nationalize companies.
If the US decides to invade Canada and Microsoft turns off services to the US military, do you have any doubts that the US will nationalize Microsoft? I don't.
> evidence that harm was done
See, fundamentally we have different interpretation of market abuse. Bork's view is simple - only look at the effect on consumers.
That simplicity doesn't make it right. Antitrust laws are meant to encourage competition. The free market does not like competition.
The law is full of edge cases. That's why we have the courts.
Since there isn't a precise and clear definition of "cruel and unusual punishment", should we get rid of it?
I kinda disagree on both accounts. Many many people give their all for less than that. More money will get you better (foreign) talent, but the correlation between pay and effort is much lower.
I work in the EU for an American company. The consensus here is Americans work more, but not harder. We deliver better work in less time.
The edge Americans have is a can do attitude and not being afraid of failure.
If you want to attract the best people in the EU it's the fact, and not even for a discussion, that you have to pay them the good $$$. Those people are found in places where the most complex infrastructure software is being built. And that is for worse or for the better is only found in FAANG or HFT or similar very large-scale software where even saving a few cents per line of code is bringing big profits to the company. How many such EU-based products are out there?
Your perspective of "delivering more work in less time" is cute. Put yourself in a pool of 95th percentile of engineers on the market and you will soon find that your experience is not representative.
I do not. Even with those skills you're still looking for a lot of hours to burn in. It only happens that those skills are more or less a prerequisite for more chance of success.
$200k is €180k, in Germany you'd take home 57% or $112k. In California you'd take home $134k
Not sure of the relative costs of housing in San Francisco vs Berlin, but it doesn't feel like a massive difference. Personally I'd rather live in Berlin on €112k net than San Francisco on $134k net
Hey, but in return you receive government tokens which will pay back in worse times... if you satisfy a lot of conditions, and wait in many queues, and send many letters.
This misunderstanding about EU/US salaries is so baffling to me, how this opinion can come up still in every EU conversation. Do people really just look at $$$ without a second of thought of how governments/taxation works and what they get for that salary/taxes?
People end up thinking this, but they're just looking at tech roles. However, the difference becomes clear when you compare other jobs. In tech, the US pays way more than Europe, for other jobs, they don't. Take teachers, for example:
- Median Salary in USA: $~58k [1]
- Median Salary in Berlin: 35k EUR [2]
So while there is a difference, it is by far not as huge a gap as in tech where Europe pays ~60-70k (very and some countries, e.g. Spain or Portugal way less) and in the US 200k-300k are not unusual.
The difference is that US companies understand the huge value creation potential of tech jobs (especially over the past 2 decades) whereas most european companies think value creation comes from having middle aged balding managers.
It is unclear to me if that salary includes pension, employer taxes, etc.
Comparing raw values also does not put job value into view: you cannot be fired on the spot. Healthcare cost almost nothing compared to the US due to some of those taxes. You can have sick-leave thay is paid for multiple years or parental leave for a year (in Germany). All of that while your employer keeps paying. This pushes salary down as risk for employers goes up, but life quality and happiness can also increase across the workforce.
All-in-all super hard to compare salaries, even though having a 200k salary in the Netherlands I would not say no to of course
There are 44 countries in Europe. What makes you think that your experience of getting the most of your government/taxation whatever you want to call it matches the experience of any other European country government/system?
Who's in the better position? Me completing a Masters degree in the EU, spending anywhere between 10k and 20k EUR, and landing a job that pays 30k EUR/year and living a pipe dream of getting myself salaried with 100k EUR, and only perhaps after 10-15 years of continous effort (and luck). Or is it someone in the States who needs to spend as much as 5x more for the same level of degree (but not education and opportunities!) - but in the very first year of employment gets 100k per year and has a vast more opportunities to build their careers up?
My perspective: I'm an (Technical) Information Security Officer in the Netherlands and I'm at €120k if I include pension contributions (Which Dutch people almost never include in their 'gross' salary).
And this is with all the employee protections you get when being salaried in the Netherlands.
If you elect to go for a more American style contract with your employer where you have little to no protection you can easily make €200k or more. Two of my colleagues get €20k a month. Downside is that they're the first to go since they're more expensive and the company can just let you go for whatever reason they want essentially any time your contract is up for renewal.
These are called 'independent contractors' in my country but essentially they are just employed 'at will' as Americans call it.
I feel like most people online that complain about €30k salaries in Europe (which is minimum wage here by the way) are just parroting online meme talking points.
This is a great way of discrediting somebody's else opinion and also incredibly a very rude one.
I earn more than you do but that's besides the point since I'm most likely in 95th percentile of the market. The point that you're missing and that I'm trying to reinforce regardless of the very good position I'm in is that 50th percentile of the market is not anywhere near +100k EUR figure. I worked with 100s engineers across different domains in Europe so the sample size should be pretty representative. I also hired many engineers so I am familiar with the market value as well.
Finally, you're also missing the point that your sample of dutch market isn't representative of all European countries markets.
I'm sorry for the rude remark, but claiming software engineers and other IT professionals are working at the legal minimum wage came across as if you were mocking us.
My apologies.
And yes, perhaps the Dutch market is not representatie for all of the EU. But then again Silicon Valley isn't representative of all of the US but these salaries are usually posted for benchmark.
Still your argument doesn't apply since I wasn't comparing salaries of experienced professionals but out of the University starting salaries.
Considering the whole European market as a whole I think I was even optimistic with 30k figure since there are many professionals with years of experience still not managing to earn much more than that. And I'm talking about building smart stuff, algorithmic design.
100k USD figure OTOH was pessimistic and it didn't include SV starting salaries which would obviously be much larger than that.
I'm talking about the EU countries and not the third-world countries but since you're ignorant and disingenuous there's no purpose of continuing this discussion anymore. More graceful interpretation would be that you're limited in capabilities to understand distributions, percentiles and some fresher year statistics, in which case it should be easy to fix.
I don't know why they don't just hire the people who built Azure / AWS / GCP etc...
If you offer these people a good salary, tax cut and a EU passport after 4 years I suspect many will be open to move to London / Paris / Amsterdam for a few years.
Subsidise the high speed train between all these cities for 5 years, offer crosscountry startup cash and investment and you might just end up with a functioning IT sector in Europe.
The triangle London / Paris / Randstad really isn't much further apart than many Sillicon Valley HQs.
It's not false. Those are the people who have already proven that they have the necessary experience to build those products so the rate of success is simply much higher rather than seeking the talent elsewhere. Startups and US companies apply this strategy all the time - you want success, poach the good people from competing companies.
I think people outside the domain often underestimate the complexity of Internet-scale infrastructure software and hardware. You don't find such challenges very easily elsewhere so simply paying more money to random engineer doesn't stand more chance than paying the same money to already established engineer in that domain.
> so the rate of success is simply much higher rather than seeking the talent elsewhere
This is true to some extend, but I don’t want the riffraff that built AWS building the EU infrastructure. It works to some extend, but some consistency would be a very welcome addition.
I don't understand what is your concern. What consistency? If you need engineers with storage engines domain expertise at scale, or query execution challenges, where else do you find those people if not in existing businesses that have already proven they have a workforce delivering exactly that type of work?
FWIW I've recently seen this exact thing play out, I can't really go into the nitty gritty but someone who had VP/Director positions in well known US public companies was brought in on a big project that's local to Europe and he's failing spectacularly.
So your theory might not work as well in practice, just my 2c from what I've seen in the real world.
Right. But how about: if he's failing spectacularly here, he's not winning over there. On the net, this is actually a win for Europe, because failure here doesn't make our IT industry much worse than it already is, while winning in the US can be leveraged into more winning.
Not that I condone such "crabs in a bucket" mentality, but put this way, it's no longer defeatist :).
The cultures across Europe are anti growth and anti speculation
That facilitate it for other people, but not enough interest culture wide
It’s kind of interesting “why don’t we just make the same amount quarter after quarter” like, wow who knew you could do that? But speculation drives innovation. Permission to fail results in faster selective pressures.
Perhaps you mean "investment" rather than "speculation"?
All mainstream European politicians are in favor of more growth, more investment, more industry, more jobs, and so on, across the continent. The parties who are about "de-growth" are marginal, even as the climate crisis is on full steam.
The actual problem is lack of pro-risk capital. Almost all of Europe's capital is concentrated in 1) conservative industry giants, and 2) huge conservative pension funds.
Basically all capital in Europe, public as well as private, is risk averse. There are very few actual "VC"s. I'm not sure that's unambiguously a bad thing, but it is the difference.
Well, I'm of the opinion that one of the other blockers of major EU growth is the lack of a singular financial market; instead of one market like the USA, we have 27 markets. Twenty seven different sets regulations that all need to go out the window; exchanges in the double digits, most of which should shut down.
This balkanization will never lead to a growth-oriented modern financial system.
Hetzner is not small, I would say it's several times bigger than OVH in turnover, but this is just a guess based on their perceived markets and market share in those markets.
The cloud offered by Hetzner is the perfect sweet spot for me, but the vast majority of IT crowd and non-IT decision makers in EU want AWS-like.
But all EU alternatives that mimick AWS will be worse than the alternative.
Pretty much like wvery MS Word copycat was worse... until Google Docs shifted the paradygm.
OVH (french) is very well known and I like them a lot. Used them for domains a lot, because they are very cheap and their management is nice.
I also like very much ScaleWay (french also) for price and quality of service, have used them for years on my startup, can highly recommend.
Also heard a lot of good things about Infomaniak (swiss), but never used them myself.
Would love to hear about european cloud providers with comments from users.
Calling yourself a cloud doesn't make you one. Can you do auto scale groups with dynamically scaling load balancers yet on OVH? AWS has had that for 15 years now.
I dont think you realize how small scale majority of EU gov software has to be.
When you have country with population under 10mil and your gov form is used by 10% of those people a year… Thats 80k submissions a month, split by 20 workdays = 4k submissions a day over 8 work hours = 500 submissions hour or about 8 submissions a minute.
I know these are wrong numbers and there are peaks etc. But many would dare to put this on single server with sqlite. Even if you 10x that.
You do need autoscale groups or self healing architectures. Govs requirements are CRUD apps whos biggest issues are design, accessibility, data permissions etc. Not scale.
While the often advertised point of auto scaling is to handle some imagined large load, in reality it's used as a reliability principle.
For example, it enables seamless rollout strategies for frequent cicd. Any regressions in application performance are automatically handled by scaling up, etc.
> The cloud side isn't polished enough to pretend to be a cloud provider.
I mean, see Azure and Google Cloud a few years back. For quite a while the market was AWS, and also some joke services which nobody who wasn't required to used (notoriously, in 2012 Azure was substantially entirely down for _over a day_ due to a leap day).
China was able to build its own IT technologies thanks to decades of protectionism and smart maneuvers from the party. Europe, being under the US "umbrella", was in an almost impossible position to do so, along with the very poor attitude of European elites, which is one of rent-seeking rather than investment.
Regardless of the attitude, I don't see this happening without some form of protectionism, trump sperging out might be the chance of a century to reboot the European military and IT industries.
If Europe wants cloud protectionism then they need apply the usual methods: tariffs on foreign cloud services and requirements on public funds to be spent on the European cloud providers.
They could apply Chinese policy and require foreign cloud corporations to have locally owned partners that will own the technology produced locally.
> They could apply Chinese policy and require foreign cloud corporations to have locally owned partners that will own the technology produced locally.
This feels like a misunderstanding of the situation. China used that tactic in cases where the Chinese people didn't know how to how to build or do a thing, and it would have taken a long time to rediscover. Saudi Arabia uses the same tactic for the same reasons.
Europe is full of people who know how to build a big cloud. It's not easy, and there's certainly learning-by-doing required, but as the author observes many of the people who work on the US clouds are from Europe or India or still live there. The knowledge transfer has been happening continuously since the start and in both directions.
The reason Europe hasn't produced a competitor on the scale of AWS is simply because that business model is the exclusive preserve of already large tech companies. AWS, Azure, GCP and Oracle Cloud are all the products of huge firms that already had tens of thousands of software engineers and (with the exception of AWS) pre-existing very profitable businesses. They could afford to sink vast sums of treasure into buildouts well ahead of demand, and subsidize their clouds using the other businesses until they had been able to catch up with Amazon. In addition, they all had large pre-existing tech ecosystems to leverage.
European countries have failed to produce tech companies on that scale for all sorts of well analyzed reasons. Moaning about cloud specifically is of no use, it's just a symptom not a cause. The causes meanwhile are the usual grab bag of uncompetitive compensation (a symptom of being poor, it's a feedback loop), bad laws that retard innovation and yet which never get fixed (due to the prevalence of governments that are explicitly suspicious of companies being successful), and finally the culture of US firms which is very globalist, collaborative and internally open, making them pleasant places to work for the people in Europe who have the right skills.
Look at it this way - would a skilled engineer rather work on AWS, having global impact under a global brand selling to every kind of customer imaginable and justifying its existence through technical innovation, or on an explicitly parochial "Eurocloud" whose highest dream is merely cloning what AWS was ten years ago, and which exists for no better reason than some domestic politicians being unwilling to resolve their differences with one specific US administration?
It's just a fundamentally unappealing proposition even if pay was good, therefore, it won't happen.
Even if you manage to build the product (R&D) now you have to incentivize the private sector and governments to switch to the EU product. There's only few options of making this happen and tariffs could be one of them. Natural transition won't happen for many reasons.
Tariffs are not particularly different to subsidies. Effectively they are a subsidy except that consumers are paying for it more directly (being forced to pay more for inferior local products) instead of through other taxes. The source of funding and effect on prices are the most meaningful differences.
> Tariffs are not particularly different to subsidies.
Tariffs are the exact opposite of subsidies; though a tariff applied globally to all imports of a class is, I guess, sonewhat similar to a subsidy to domestic products of that class.
In what way are they so different? Both increase prices for the the consumer it’s just they end up having to pay for the subsidy indirectly through taxes and such instead of having to pay a higher price. From the sellers perspective as long as subsidies are only available for domestic producers they are even more similar to tariffs since they reduce the competitiveness of imported products.
Subsidies give money to selected market participants.
Tariffs take money from selected market participants.
They are, exactly, opposites.
They are similar in that they are government interventions in the marketplace, which under simplistic ideal-market assumptions have net costs which must be born somewhere, but at that level every possible government action is "the same thing".
> Both increase prices for the the consumer it’s just they end up having to pay for the subsidy indirectly through taxes and such instead of having to pay a higher price.
No, they don't. "The consumer", "current taxpayers", and "future taxpayers" are different-but-overlapping groups of people.
The recipe is simple but nobody supposedly wants to do it.
1. Find all (e.g. LinkedIn) the EU-based engineers working or having worked for American cloud companies.
2. Offer them a proportional salary to join and start building the EU cloud product.
3. Incentivize the engineers to top-performance by giving them bonuses, stocks, whatever ...
4. Go back to the garage engineering and cut-off the bureaucratic BS with endless PMs, milestones, JIRA boards, scrum masters, chapter masters, architects, and all other similar BS.
But EU doesn't want to do it - they rather spend money in ways so that "EU funds", which are supposedly public, end up running up to their pockets directly. Dozens of examples out there but "AI factories" being the last embodiment of that.
You describe a plan for building it. That is but one piece of the equation. How will you sell it? Why should people pick you over established AWS/Azure/GC? Will you target new projects, or get people to migrate? How does the strategies there differ? What is the timeline to becoming profitable? How will you fund it in the meantime?
I'm aware of that. Please see my other comment that already addressed your concern. The point rather was that EU strategy won't even get us so far because it doesn't have or doesn't want to have a plan.
The EU is ran by lawyers and political science majors. They have absolutely zero idea how to setup proper incentives. Any hope of EU getting its act together is wasted. Much like our tax euros.
There’s nothing in the EU quite like DARPA: oriented to long term, fundamental, high risk, high reward projects.
Furthermore, this European DARPA would be an organization with its own physical structure - its own labs. Say, with each different discipline in a different EU country.
Cloud is highly overrated for a lot of projects. While in theory you could scale up or down based on demand, my experience the need for it was rare. More common was provisioned resources sitting there for months.
And S3 encourages you to basically save everything
And there was another aspect, since compute / storage was a consumable. Devops were less shy about spinning up services as if they were candy, rather than being more economical with resources that were fixed in the short term.
It also takes care of backups, disaster recovery, maintenance, HW replacement and offers many services right out of the box that would require at least a single extra DevOps guy to do on-prem.
Managed databases and managed storage like S3 is where it's at.
Australia needs to have a rethinking of its geopolitics strategy. It still seems to think it's part of Europe while at the same time being one of the members of Five Eyes, which makes it an extension of USA and UK's intelligence network (which has become problematic from the EU's perspective). In reality, Australia's biggest trade partners are all Asian (China, Japan, South Korea, India). If anything, Australia should be seeking to join efforts with its Asian partners and neighbours.
I'd argue Aus should really seek a policy of neutrality, and strong Asian trade. Remaining neutral requires strong diversification and the EU is a great independent technology supplier to keep on board.
I think the big problem is this: That NLNet thing he mentions gives subsidies of the insanely large amount of €5-50k that’s not even worth getting out of bed over, much less quitting your job on the off chance you are successful in the 3 months that gives you.
> There are also false prophets—we all want quick solutions, and sometimes we get (too) excited about people who promise them. For example, in Germany, some claim they have “made Microsoft’s cloud European” through SAP. However, this turns out to be not quite the case. If the U.S. refuses to cooperate, they (by their own admission) would be down within a few months. Amazon also claims they are going to do something similar, but they’ve been saying that for years, and they conveniently avoid addressing what happens in case of a conflict with the U.S.
An EU sovereign cloud by Amazon though would be built in the EU and staffed by EU employees though.
The effect of the US trying to insist it won't co-operate would be similar to what happened to most Russian based American franchise locations: troublesome but it's not like the facilities, people or inventory just vanished.
> An EU sovereign cloud by Amazon though would be built in the EU and staffed by EU employees though.
At that point, why is Amazon needed? Don't we really have the expertise and capital to do this in Europe without Amazon (or Google, MSFT)? I somehow doubt that very much.
Make all tech knowledge available, spin it into another completely separate company with 0 ties to the parent and then we can actually start talking. Rest is fluff and PR.
I am really questioning that initial premise that "we need an European cloud". Here I roughly understand "we need a few big European hyperscaler with a dozen huge datacenters we can migrate to" because this is the model we understand. But you are fundamentally "fighting the last battle", the wrong battle.
The idea of an European cloud has happened several time before and I was once strongly encouraged to select one so I investigated, here are my findings:
About 10 years ago all those countries needed a cloud to put their government and sensitive industries on, so public money and contracts was given to friendly corporations.
Once the initial chunk of free money ran out a lot of those companies got rid of the business. In France a lot got sold to Orange, the big telco partially owned by the state.
I considered it because according to some metrics they had the biggest cloud in Europe.
I found out quickly they actually immediately given everything to Huawei. So it was all managed by Huawei but if I remember correctly still hosted on Orange premises.
Strangely information about it is more difficult to find now but you can still read about it on third party websites [0].
By the way all of this is GDPR whatever compliant.
I just did a quick search and it seems they had to shut it down in 2024 because Huawei became such a radioactive topic. Information is hard to find about it in English.
In the end the companies who were forced to select a "sovereign cloud" probably went through hell and moved around several time in the last 10 years.
You are told you should/must go on a sovereign cloud, you then end up on a Chinese cloud and you are then kicked out of it.
I am not pushing for the big american hyperscalers but be very careful before listening to politicians.
You have a French perspective, but EU is big and varied.
Here in Belgium we do have our own Government Cloud which, while hosted using Broadcom (VMWare) and IBM (RedHat Openshift) software is locally owned and controlled.
Now, this "cloud" is a misnomer. It is not as automated as AWS. I would rather call it a government hosting provider, but it's doing its job.
I have seen a lot and usually the smaller the better.
I remember the Flemish gov (7M people so quite small) was doing great in this domain. France where everything needs to be centralised, big and controlled is doing terrible I guess.
This is why starting the discussion with the concept of "EU" and "Cloud" is wrong. We do not need an aircraft carrier, we need a lot of drones.
I agree but for a slightly different reason. EU cloud is not about chasing AWS or GCP or Azure scale and complexity. It's about hosting our businesses and most importantly governments locally.
Thus, "cloud" is not an objective but a means to a goal.
It seems we are pushed into major conflicts and disruptions so we have to decentralise it as much as possible.
The current model is very awkward, we have been pushed into those very centralised, hyperscale solution that most businesses and individual do not really need.
Most businesses do not really need much to do their job.
We are probably gonna have major connectivity disruptions on of those days or what happened when your Hertzner or Orange/Huawei datacenter is down?
And I know they have "better security" but if we are attacked they are gonna be taken down anyway.
We do not wanna look at what a major cyber disruption would do to our societies. I believe just the consequences on the electricity grid and hospitals would make us wave the white flags in a few weeks.
You wanna own your hardware because if there are disruptions it is gonna be hard to get your hand on some because of supply, logistics or just monetary disruptions.
How hard is it technically to setup something like Ubuntu Microcloud [0] for your own use or local businesses you provide that service for?
From my experience the biggest risk and headache is all those EU regulations like the EU Cyper resilience Act and in some cases energy costs.
So when Ursula tell you Europe needs to prepare for war and authorise/push your country to go into even more debt so her friends can build expensive tanks that will never see a battle it is ridiculous.
They should focus on removing all the obstacle to resilience they setup themselves, find a way to reduce energy costs (pretty hard now...) and maybe encourage moving out of those big tech cloud (starting by doing it themselves).
Yes there is good chance we are being pushed into one.
What I mean is the tanks, aircraft carriers or F35s probably won't really matter.
You can create enough pain on current western societies by other means such as major cyber disruption so that the classical battlefield won't really matter.
I am not OP but let me give you a Microsoft Office example.
For many years, Open Source and businesses tried to make a Microsoft Office competitor by mimicking Microsoft Office. Naturally, all clones were worse than the original.
Until Google changed the paradigm and rolled out Google Docs that had a unique feature of online collaboration.
Then was the turn of Microsoft to mimic collaboration features of Google Docs in Microsoft Office and be worse at it almost by definition.
Another example is S3.
For years, businesses tried to have POSIX-capable filesystems seamlessly scale in size and in availability. It took Amazon to roll out a simpler alternative that, by having a smaller set of features, enable so much sought properties of distributed file systems in an efficient and commercially viable way.
I think Hetzner Cloud is a sweet spot of a cloud. Instead of reimplementing AWS, EU should standardize of something like it.
You are so right about computers, except from Raspberry Pi (UK) not much as been done in Europe.
Regarding the eu cloud, it is definitely about sovereignty, specially since the CLOUD Act (H.R. 4943). In the context of the global trade war initiated by the US of A, it also makes sense from a European Union perspective.
Sorry to be a killjoy, I see these sentiments everywhere across the EU sovereign cloud space. The dream of distributed "cloud providers" made up of individual companies products, all cooperating and communicating together in some utopian way. That's the stuff that Gaia-X was huffing.
Make no mistake, this will absolutely be curb stomped by vertical integration.
But I get the sentiment, we need to figure out a way of collaborating on this. Perhaps airbus is a good example.
No he couldn't. That site uses a non-geographical definition of Europe that excludes Britain. "EU, EEA, EFTA, or DCFTA member country" also excludes Serbia, Turkey, Belarus and Russia but includes several countries that aren't able to join the EU at all due to corruption or misaligned legal systems.
This really shows the incoherence of the whole Euro project. There's no such thing as European-ness: when the sort of people who wave the blue flag use the word Europe they are imagining an ideological construct subject to inconsistent and ever-shifting definitions. They don't even agree with each other what European means. One minute Britain is in Europe and Ukraine isn't, events happen, and suddenly Ukraine is European via DCFTA and Britain isn't. Switzerland is a similar case: sometimes it's considered to be European by these types, and other times not.
Why should I, a man born in Britain now living in Switzerland who has worked on two different US clouds, want to apply that experience to a Eurocloud given this history? This of thing is why it will never inspire much loyalty, and why Bert instantly gives up on the idea of a European cloud being used because it's actually good. The resort to force of law underpins the entire project because the European identity is a sort of social engineering programme, not something organically developed.
To be fair, the non-geographical definitions that excludes Britain, actually only excludes Britain because Britain excluded itself of the European Union in 2020.
And yet that website doesn't use the EU as a definition of European, so it clearly doesn't matter in this case. That's what I'm getting at: the word European doesn't seem to mean anything because the people who use the word the most are relying on definitions that yield unintuitive and self-defeating outcomes, like deliberately excluding one of the countries in Europe that actually does have a bit of a tech ecosystem.
I've done quite a bit of government contracting for governments in Europe, including various government levels in Belgium and the EU itself. Every single time, I encountered a lot of friction regarding hosting. The applications were always hosted by a partner [1] that won a large public tender, usually by virtue of supposedly being the cheapest. Those hosting partners were invariably awful and incompetent. Simple tasks would take weeks. Complex tasks wouldn't happen at all. There was enough red tape to go to the moon and back.
Every project I worked on, developers (consultants like me) advocated for finding a way of bypassing the hosting partner, and either doing their thing on top of platforms like Hetzner, or switching to AWS/Azure. Not because they particularly love AWS, but because the default hosting solution sucks so much arse.
I can totally see EU cloud infrastructure working for the public sector, but ONLY if they come with the same level of self-service management as the big cloud providers. Without that, it's a non-starter.
[1] One of those is a large French company on the verge of banruptcy.
In my experience this is mix of lack of knowledge sometimes mixed with self-dealing/corruption.
But it can be prevented by writing hosting specs well. If in your proposed solution you specify really detailed hosting envinronment then most of these old incopetent companies wont even apply.
Unfortunately, the way my governments have dealt with the issue is by writing one very large tender, from which a Hosting Framework Agreement is derived. Individual projects have no say in it, and they're certainly not consulted when the agreements are drafted. So basically they're stuck with whatever is on offer, which isn't much.
A choice example was a hosted K8S solution: it was supposed to be highly available, but the framework agreement only specified 2 datacentre locations. The hosting provider thus couldn't build a reliable stretched cluster, and instead built two identical clusters. Forcing every project to figure out how to deploy their applications in a HA fashion on top of two clusters. Much fun. Not a giant waste of resources at all. /s
Hmm one other thing - USA is completely English but EU is quite multi-lingual so to be competitive within EU market services need a lot of more efforts.
If you take the triangle between London - Paris - Randstad (Amsterdam+) then we're talking about 60+ million people all connected via <4 hours of high speed rail where essentially everyone can speak English.
(Okay maybe in Paris only people under 40 can but in IT it's usually okay imo).
You don't need the entire EU, just a chunck of it that can carry the industry like Silicon Valley does for the US.
> The European cloud industry has struggled to take off, partly because existing players lacked the will or vision to compete effectively (among many other reasons).
The EU would have come after them. It is coming after Google, AWS and Azure even though they have diplomatic support from the US proper; how is an EU cloud provider supposed to compete?
The people with the will and vision exist, if anyone can build EuroCloud they would be wealthy beyond the dreams of mortals and it really just needs to be an OK AWS clone. The unfortunate part is that anyone who sees that far can also see how the story will end when the bureaucrats get involved early instead of after they have multiple billions for compliance. It'll cripple any ability the project has to compete.
It doesn't have to compete 1:1 when tariffs and regulations get applied, same principle as in all other industries ie cars.
Its not the best approach but it can work, we have half a billion rich domestic market and some subsidies to kick start things would be rewarded in long run tremendously.
The reason why Europe won't have a cloud is because you can't have managed decline and nice things. This is the same reason Europe will forever be depdent on someone else for defence and why it will always be the meek but somehow petulant follower of whoever is the world hegemon. This is why China and Russia cares so little about them also, they know that Europe will fall in line once their current patron no longer can or want to sustain them.
A few years ago I would have maybe said the same but we're now entering a new era where the US capital will likely go towards the EU which is seen as more stable and future oriented.
The US tech companies where somewhat insulated in Europe thanks to the US defense but now that is out of the picture, I don't see much else to protect them.
The issue is not lack of european providers. The issue is almost natural lobby from the likes of Microsoft and Amazon.
All the unis are infected with former corpo programmers who push their preferred employer stacks. Students get free credits to use on Azure. Some of them become gov employees or consultants. You end up with all sides agreeing Microsoft Azure is the most trustworthy solution.
Everything gov in my country seems to be built with asp.net to point where its often requirement in the contract.
Lets start with erasing all the proprietary tech taught at universities. Microsoft can do their education. Universities dont have to do it for them for free.
Also Microsoft consultants are everywhere, and management consultants like McKinsey usually work closely with Microsoft, and many corporations are already locked in.
In many ways its kind of really scary how so much of the digital infrastructure even of key enterprises and entire states is owned and operated by a private company.
What is the problem with asp.net? It is open source.
It didnt use to be open-source. Its closely tied to MS which can do whatever about it. It will always have better support on MS and Azure. It still MS stack and if you want to get out of that you have bazilion better more free options.
In practice, it is source-available, not open source.
I didn’t see any mention of EU’s last attempt to build a sovereign cloud: Gaia-X. I think that didn’t really go anywhere due to disagreements over technical direction. Much like Quaero.
Don’t get me wrong, I hope it goes better this time and the author acknowledges the many pitfalls of such a complex undertaking.
Edit: The author previously discussed Gaia-X at length here: https://berthub.eu/articles/posts/gaia-x-is-an-expensive-dis...
Unfortunately Gaia-X has been set up and run by the wrong kind of people.
Look at this recent job ad for a Senior Software Developer / Tech Lead: https://www.linkedin.com/jobs/view/4181090679
AFAIU it's written by the CTO and see this gem:
" Innovation Mindset: Willingness to explore new technologies – especially related to trust frameworks and associated standards –, staying ahead (no kidding!) of main-stream industry trends (This is a real requirement, not ChatGPT – just look into our GitLab repos!)."
I checked their public GitLab and it's puny. Moreover, I suspect based on the use of bold and bullet points that the job description itself went through ChatGPT.
It's more or less covert subsidies for the usual suspects. In Germany that means the like of Fraunhofer-Gesellschaft, Atos, Bosch, Deutsche Telekom, SAP, BMW and Siemens get some millions (and by "some" I mean around 200 million euro). If there's no outcome, like in the GAIA-X case, it's just bad luck for the taxpayer.
If you start with a shitty name like that you have failed before you even start in my opinion.
Even the utterly straightforward eucloud is better. Or make it cloud.eu :)
Disagreements over direction should be solved by making separate orgs/initiatives to go test each approach.
Better than funding “one true initiative” and force people to work together to keep the funding.
The VC funding model succeeds in experimental domains by letting many experiments run at once. Kind of a form of Darwinism. How to do that with funded initiatives is a really interesting challenge.
The impression I have, but I may be wrong, is that investors in Europe want a return on every investment. Failure is not an option. Which kind of choke innovation.
This is not true, the problem is greed and imo a lack of ROI.
At the end it's just a market problem, no alternative can emerge because the US cloud companies are way too large.
Remove them of the picture and the market will solve itself very quickly, most of the companies to replace them are already there in a weaker form.
The cloud platforms are nowhere near as hard to replace as the chips.
Without protection, a EU hyperscaler will never emerge. It's not that any hypothetical future EU hyperscaler would be an inherently less competitive. It's just plain old first mover advantage. Cloud computing is a VERY sticky good, and once folks got on AWS/Azure no matter how good your product is there's no way larger customers are going to migrate off to some "risky" startup (thus keeping scale small).
Source: I worked on a sovereign cloud PaaS in Germany. Admittedly the project was a disaster, hiring talent in the infrastructure space here is really tough. Probably mostly due to a lack of opportunities to acquire experience domestically and noncompetitive wages limiting migration of those with experience.
Gaia-X is a total dumpster fire though, perfect example of what's wrong with the EU. All the money was spent paying bureaucrats, consults, and standards committee folks to architecture astronaut and design by committee sovereign cloud "standards" bullshit bingo. None of the money went to the engineers.
We should really copy the Chinese approach here, straight out clone a subset of the best/most popular parts of AWS/Azure/GCP. I'm open to contracting opportunities in this space BTW (email in profile).
I'm expecting that to happen. Realistically, the US tech companies are the very first target in line in a market war.
They are paying very low taxes in the EU, contributing enormously on the US economy and are close to the president. That has to be the biggest target.
And the US can't even use the EU defense card anymore to defend them as Trump wasted it with his Russian appeasement.
I'd rate this probability at 50% at least.
Suffice to say as an EU based engineer in the infrastructure space, I'm rather optimistic for my future employment opportunities.
This is not their last, there is another one with funding thats building off the learned lessons of Gaia-X.
I don't think bureaucrats can build a successful cloud by enforcing laws, adding even more regulations, making policies and using subsidies.
This is not how the market likes to work. There has to be a demand and the free market forces will build a supply. Governments can help by reducing regulation and bureaucracy, reducing taxes.
Policies are often used to create markets or - like in this case is propose - to create that demand. The article makes the argument that free market failed to deliver in this sector in Europe. One of the explanations is that once you lag behind competitors, if policies don't force to value specific parameters that can be fulfilled only by other competitors, no competitors will join the market because nobody is going to choose them. So the author argues for example to impose regulation for public tenders such as "must be subject to only EU laws". This creates a demand, which is not currently matching any offer in the market and creates market incentives for new players to compete. So regulations can absolutely work where "free" market fails (quotes because even the big 3 are/were pumped full of money by government/defense contracts).
> There has to be a demand and the free market forces will build a supply
Not in all cases it doesn't.
If the return on investment isn't there or if companies are unable to fulfil the demand then the market has failed. Recent example of this has been EVs where regulation needed to step in to force the market forward.
>If the return on investment isn't there or if companies are unable to fulfil the demand then the market has failed.
If there's no demand, there's no problem. Also there's no money to be made. If there's demand there will always be companies to supply that demand. Of course, it takes a time.
Politics shouldn't dictate to economy.
>Recent example of this has been EVs where regulation needed to step in to force the market forward
Then why should people be forced to use EVs instead of whatever they like? That is not freedom.
> If there's demand there will always be companies to supply that demand.
I want to walk around town without cameras tracking me. Can companies supply my demand?
I want PFAS and microplastics out of the evironment. Can companies supply my demand?
I want net-zero CO2 emissions in all the products I buy. Can companies supply my demand? How do I trust they are not green-washing? Do I depend on even more companies? How do I trust those companies?
Sure seems simpler to have the government enforce privacy and pollution laws.
> Politics shouldn't dictate to economy.
Which means you think William Wilberforce's most famous work was completely wrong.
> Then why should people be forced to use EVs instead of whatever they like?
You should consider how government-subsidized oil production and distribution has forced you to use fossil fuel cars instead of whatever you like.
You should consider even more how government-subsidized planning forces you to use a car in the first place, and often closed off alternative solutions like walking, bike, and mass transit.
Because we have wider goals as a society that go beyond absolute freedom. However things like EV incentives don't make you less free, you can still buy whatever you want.
Free markets like to make money for capital holders, and tend towards a monopoly or oligopoly as that maximizes profit extraction.
Name a free market where there is an economy of scale and network effect, which does not tend towards monopoly or oligopoly.
More fundamentally, if (say) Germany wants to move all of its government operations off of US-based IT, should it wait until the market has provided a solution? Or should it do like all governments have always done, and change the marketplace?
Nor is this unique to governments. Some of my clients require me to agree to their ethical practices, like "no child labor." They are using their purchasing power to change the marketplace.
Why do the countries with higher taxes and government regulations and policies over health care have better overall health outcomes than those with lower taxes and regulations?
There is a demand for clean beaches. Where is the free market for pollution control?
The topic we're discussing would seem to undermine your claim, as cloud is neither a monopoly nor an oligopoly despite having scale and network effects.
Could you explain then how so many European governments ended up one of a handful of US clouds (Amazon, Microsoft and Google)?
Almost reason I can think of comes down to scaling or network effects.
I say "almost" because I've thought of a couple more:
- it's already a Microsoft shop and it's hard to change? Lock-in effect.
- you never get fired for choosing IBM, err, AWS? I don't have a name for this effect, though it seems related to scaling.
Note that "monopoly" does not mean there is only a single provider. "De Beers controlled 80% to 85% of rough diamond distribution and was considered a monopoly" during the 20th century. https://en.wikipedia.org/wiki/De_Beers
This is because I'm using the legal definition of monopoly, not the economics one. From https://en.wikipedia.org/wiki/Monopoly
> In economics, a monopoly is a single seller. In law, a monopoly is a business entity that has significant market power, that is, the power to charge overly high prices, which is associated with unfair price raises.
Because those companies:
1. Invested massively. In the case of Microsoft and Google, ahead of demand.
2. Have lots of developers already familiar with their stacks, especially Microsoft via Windows and .NET which are mainstays of nearly all European non-tech businesses.
3. Indeed, have lots of "never get fired" effect; the word you're looking for is brand equity. Especially the case with non-technical execs already familiar with their consumer products, and it's often non-technical execs who make the decisions about which brand they're going with. Everyone knows MS Office and likes it, everyone knows Amazon and buys from it, everyone knows Google and searches with it. Their products work well, so their clouds must also work well. In government especially the decisions are always non technical so this matters a lot.
Monopoly does actually mean only a single provider: that's the definition and your Wikipedia link opens by stating exactly that in its first sentence. Sometimes the term is stretched or abused by governments. That happens because everyone across the political spectrum agrees to anti-trust laws in principle, and then the left can use them as a stick to beat companies with if only they can twist the definition of monopoly far enough. This so-called legal definition isn't a definition at all, it's a purely political construct which is why it includes ambiguous terms like significant, unfair, overly high etc.
Even using this non-definition though, cloud providers don't qualify. They're actually famous for not raising prices. Raising prices a lot on existing customers is a typical sign that a company has significant market power.
"Invested massively" because the could get scaling advantages. Both companies built off of existing OS and ad tech monopolies which should have been regulated long ago.
"Lots of developers already familiar with their stacks" is a network effect.
> Monopoly does actually mean only a single provider
Is a tomato a fruit?
I was very clear to give my definition of "monopoly", to include "oligarchy" for those who wanted to be picky, to point to how others describe De Beers as a "monopoly", consistent with my use, and to show how even the Wikipedia page on "monopoly" is not disagreeing with me.
You'll note that same Wikipedia page for "monopoly" lists De Beers and Standard Oil as monopolies, when they were not sole providers of diamonds or petroleum, respectively.
> They're actually famous for not raising prices.
Right. You are using Bork's 1970s influential re-orientation of monopoly power away from general anti-competitive behavior to focus only on "consumer welfare". I agree with Lina Khan (starting with "Amazon's Antitrust Paradox") and others, who argue that Bork's standard is too narrow.
Or, to use your terms, Bork's right-wing viewpoints twisted established law to allow abusive behavior by powerful companies, so long as consumer pricing was not affected. V.P. Vance likes Khan because he wants the right to use the stick to beat companies which don't support the right. This is only possible because all of the big tech companies already abuse their market power, because Bork gave them the freedom to do so.
> This so-called legal definition isn't a definition at all, it's a purely political construct
All laws are political constructs. What is "ownership"? Do you own the water which falls on the land you own? What is a "reasonable search"? What is "cruel and unusual punishment"?
We're in agreement that the clouds were funded by other lines of business:
https://news.ycombinator.com/item?id=43397653
Good law is precise and clear, so everyone knows how to follow it. The meaning of the word "ownership", to take your example, is precisely defined via legislation and case law.
Anti-trust law is neither precise nor clear which is why - and we are both in agreement on this point - that right now both left and right are happy to pick it up and use it to "beat companies" for political reasons. They can't do that with property law because property law is well defined.
The big advantage of the Bork formulation is exactly avoidance of this outcome. By requiring there to be actual evidence that harm was done, it hugely reduces the ability to abuse these laws for political ends and keeps them focused on behaviors that are actually detrimental to the market.
Ownership is not precise and clear. There are always new cases to figure out what it means.
Like, if there are two pieces of public land, which meet at a corner, and two pieces of private land on either side of the corner, then are you trespassing on private land if you step from public land to public land across the corner?
I think all of the big IT companies deserve to be beat, broken up, and strict laws placed in future power. That's different than using the cudgel to only go after the ones which disagree with whoever is in power.
> for political reasons
My view is non-partisan. If Amazon were to become a worker-owned co-op tomorrow, infused by the spirit of Joe Hill, I would still want it broken up.
> They can't do that with property law because property law is well defined.
The US has some very big cudgels already.
The government can pass laws changing what property law means. The government has very broad power to seize land, and exercise eminent domain. The government can nationalize companies.
If the US decides to invade Canada and Microsoft turns off services to the US military, do you have any doubts that the US will nationalize Microsoft? I don't.
> evidence that harm was done
See, fundamentally we have different interpretation of market abuse. Bork's view is simple - only look at the effect on consumers.
That simplicity doesn't make it right. Antitrust laws are meant to encourage competition. The free market does not like competition.
The law is full of edge cases. That's why we have the courts.
Since there isn't a precise and clear definition of "cruel and unusual punishment", should we get rid of it?
The elephant in the room here is pay. You can’t expect people to bend over backwards for 70k/yr.
Start paying 2-300k in Europe and watch shit being built.
Why would people need to bend over backwards to build good things?
Because anyone that ever built anything worthwhile in this field knows you have to give it your all, when starting out especially.
Nobody gives their all for 5k/mo.
I kinda disagree on both accounts. Many many people give their all for less than that. More money will get you better (foreign) talent, but the correlation between pay and effort is much lower.
I work in the EU for an American company. The consensus here is Americans work more, but not harder. We deliver better work in less time.
The edge Americans have is a can do attitude and not being afraid of failure.
If you want to attract the best people in the EU it's the fact, and not even for a discussion, that you have to pay them the good $$$. Those people are found in places where the most complex infrastructure software is being built. And that is for worse or for the better is only found in FAANG or HFT or similar very large-scale software where even saving a few cents per line of code is bringing big profits to the company. How many such EU-based products are out there?
Your perspective of "delivering more work in less time" is cute. Put yourself in a pool of 95th percentile of engineers on the market and you will soon find that your experience is not representative.
It looks like you're agreeing with me that it has nothing to do with bending over backwards?
I do not. Even with those skills you're still looking for a lot of hours to burn in. It only happens that those skills are more or less a prerequisite for more chance of success.
Especially if more than half of that goes into taxes
$200k is €180k, in Germany you'd take home 57% or $112k. In California you'd take home $134k
Not sure of the relative costs of housing in San Francisco vs Berlin, but it doesn't feel like a massive difference. Personally I'd rather live in Berlin on €112k net than San Francisco on $134k net
Oh, absolutely.
Hey, but in return you receive government tokens which will pay back in worse times... if you satisfy a lot of conditions, and wait in many queues, and send many letters.
give it your all is good way to a burnout. If you want to build a "cloud" it's gonna be 5+ yrs, a marathon not a sprint, you need to work sustainably
This misunderstanding about EU/US salaries is so baffling to me, how this opinion can come up still in every EU conversation. Do people really just look at $$$ without a second of thought of how governments/taxation works and what they get for that salary/taxes?
People end up thinking this, but they're just looking at tech roles. However, the difference becomes clear when you compare other jobs. In tech, the US pays way more than Europe, for other jobs, they don't. Take teachers, for example:
- Median Salary in USA: $~58k [1]
- Median Salary in Berlin: 35k EUR [2]
So while there is a difference, it is by far not as huge a gap as in tech where Europe pays ~60-70k (very and some countries, e.g. Spain or Portugal way less) and in the US 200k-300k are not unusual.
The difference is that US companies understand the huge value creation potential of tech jobs (especially over the past 2 decades) whereas most european companies think value creation comes from having middle aged balding managers.
[1]: https://www.salary.com/research/salary/benchmark/public-scho... [2]: https://worldsalaries.com/average-teacher-salary-in-berlin/g...
It is unclear to me if that salary includes pension, employer taxes, etc. Comparing raw values also does not put job value into view: you cannot be fired on the spot. Healthcare cost almost nothing compared to the US due to some of those taxes. You can have sick-leave thay is paid for multiple years or parental leave for a year (in Germany). All of that while your employer keeps paying. This pushes salary down as risk for employers goes up, but life quality and happiness can also increase across the workforce.
All-in-all super hard to compare salaries, even though having a 200k salary in the Netherlands I would not say no to of course
There are 44 countries in Europe. What makes you think that your experience of getting the most of your government/taxation whatever you want to call it matches the experience of any other European country government/system?
If you do a comparison of EU/US I guess everyone naturally thinks of something like San Francisco/Berlin and not Bucharest.
The base logic does match though. The execution is very country-dependent.
Then please correct your comment into Berlin/SF rather than EU/US. Otherwise it sounds like you're talking uninformed nonsense.
The gap is _huge_ and goes way beyond having different government taxation.
This. As someone who works for a US company from Europe, there are no financial alternatives. I need to get my kids through college.
I wonder how other europeans get their kids through collage? Hmm maybe because its super cheap compared to US?
Please do work for US company from Europe thats great. Just lets not put it on college and your kids.
They just glue them together I guess.
Which EU country has comparable tuition fees to the US? Do yor kids go to some UK private university?
Who's in the better position? Me completing a Masters degree in the EU, spending anywhere between 10k and 20k EUR, and landing a job that pays 30k EUR/year and living a pipe dream of getting myself salaried with 100k EUR, and only perhaps after 10-15 years of continous effort (and luck). Or is it someone in the States who needs to spend as much as 5x more for the same level of degree (but not education and opportunities!) - but in the very first year of employment gets 100k per year and has a vast more opportunities to build their careers up?
My perspective: I'm an (Technical) Information Security Officer in the Netherlands and I'm at €120k if I include pension contributions (Which Dutch people almost never include in their 'gross' salary).
And this is with all the employee protections you get when being salaried in the Netherlands.
If you elect to go for a more American style contract with your employer where you have little to no protection you can easily make €200k or more. Two of my colleagues get €20k a month. Downside is that they're the first to go since they're more expensive and the company can just let you go for whatever reason they want essentially any time your contract is up for renewal.
These are called 'independent contractors' in my country but essentially they are just employed 'at will' as Americans call it.
I feel like most people online that complain about €30k salaries in Europe (which is minimum wage here by the way) are just parroting online meme talking points.
And those independent contractors get no WW or WIA, have to pay their own AOV and ZVW and if they are sick for a month their contract is not renewed.
So the difference is not even _that large_.
Do you see any prospects of increasing that €120k? Because that’s one of my biggest issues with Europe.
Well since I'm salaried it grows with (at least) the rate of inflation, thanks to unions.
Also I'm 32 now so yeah I don't plan to stop investing and growing yet.
Perhaps I will also swap to the independant contractor model in the future but for now I'm content with all the extra benefits being salaried brings.
Lol, ok. Yeah, at 32 with €120k salaried I think you are pretty well off. Especially compared to the median in that area.
Yes, all IT people in the Netherlands have benefitted from Brexit and Covid, indirectly.
Companies shifting HQ's and hiring Dutch people remotely at London salaries has pushed up our compensation considerably in the last few years.
> are just parroting online meme talking points
This is a great way of discrediting somebody's else opinion and also incredibly a very rude one.
I earn more than you do but that's besides the point since I'm most likely in 95th percentile of the market. The point that you're missing and that I'm trying to reinforce regardless of the very good position I'm in is that 50th percentile of the market is not anywhere near +100k EUR figure. I worked with 100s engineers across different domains in Europe so the sample size should be pretty representative. I also hired many engineers so I am familiar with the market value as well.
Finally, you're also missing the point that your sample of dutch market isn't representative of all European countries markets.
I'm sorry for the rude remark, but claiming software engineers and other IT professionals are working at the legal minimum wage came across as if you were mocking us.
My apologies.
And yes, perhaps the Dutch market is not representatie for all of the EU. But then again Silicon Valley isn't representative of all of the US but these salaries are usually posted for benchmark.
Still your argument doesn't apply since I wasn't comparing salaries of experienced professionals but out of the University starting salaries.
Considering the whole European market as a whole I think I was even optimistic with 30k figure since there are many professionals with years of experience still not managing to earn much more than that. And I'm talking about building smart stuff, algorithmic design.
100k USD figure OTOH was pessimistic and it didn't include SV starting salaries which would obviously be much larger than that.
Ah, okay. Well my starting salary was €50k in 2016 so it's probably €60k or €70k now.
If we're considering entire continents I'd like to point out that the median salary in Mexico is €30k. :)
I'm talking about the EU countries and not the third-world countries but since you're ignorant and disingenuous there's no purpose of continuing this discussion anymore. More graceful interpretation would be that you're limited in capabilities to understand distributions, percentiles and some fresher year statistics, in which case it should be easy to fix.
Do you think that is just about tuition?
That’s the weirdest reason ever. If I were in Europe I could not work at all and put my kids through college.
You would need to be in a good part of Europe.
Which part of Europe do you need a US engineers’ salary to put kids through college?
Touché. This is usually the problem with the "lack of skilled workforce" on the old continent.
I don't know why they don't just hire the people who built Azure / AWS / GCP etc...
If you offer these people a good salary, tax cut and a EU passport after 4 years I suspect many will be open to move to London / Paris / Amsterdam for a few years.
Subsidise the high speed train between all these cities for 5 years, offer crosscountry startup cash and investment and you might just end up with a functioning IT sector in Europe.
The triangle London / Paris / Randstad really isn't much further apart than many Sillicon Valley HQs.
The assumption that those are the only people that are able to build something like that is simply false.
There's plenty of talent around, it just needs capital.
It's not false. Those are the people who have already proven that they have the necessary experience to build those products so the rate of success is simply much higher rather than seeking the talent elsewhere. Startups and US companies apply this strategy all the time - you want success, poach the good people from competing companies.
I think people outside the domain often underestimate the complexity of Internet-scale infrastructure software and hardware. You don't find such challenges very easily elsewhere so simply paying more money to random engineer doesn't stand more chance than paying the same money to already established engineer in that domain.
> so the rate of success is simply much higher rather than seeking the talent elsewhere
This is true to some extend, but I don’t want the riffraff that built AWS building the EU infrastructure. It works to some extend, but some consistency would be a very welcome addition.
I don't understand what is your concern. What consistency? If you need engineers with storage engines domain expertise at scale, or query execution challenges, where else do you find those people if not in existing businesses that have already proven they have a workforce delivering exactly that type of work?
Why reinvent the wheel?
When US companies chase a competitor they always just poach people directly.
No one is saying those are the only people who could do it.
But they have experience and hiring the senior people directly would greatly speed up the process in Europe I think.
FWIW I've recently seen this exact thing play out, I can't really go into the nitty gritty but someone who had VP/Director positions in well known US public companies was brought in on a big project that's local to Europe and he's failing spectacularly.
So your theory might not work as well in practice, just my 2c from what I've seen in the real world.
Yes well, perhaps. But this kind of defeatism has never brought anyone to victory.
Right. But how about: if he's failing spectacularly here, he's not winning over there. On the net, this is actually a win for Europe, because failure here doesn't make our IT industry much worse than it already is, while winning in the US can be leveraged into more winning.
Not that I condone such "crabs in a bucket" mentality, but put this way, it's no longer defeatist :).
If he is not winning here, he probably did not win there, just got landed in the right place at the right time and had his free ride.
There is a lot of bullshit jobs in US IT companies, FAANG are not an exception.
This is why Bert talks about local technical competence.
The cultures across Europe are anti growth and anti speculation
That facilitate it for other people, but not enough interest culture wide
It’s kind of interesting “why don’t we just make the same amount quarter after quarter” like, wow who knew you could do that? But speculation drives innovation. Permission to fail results in faster selective pressures.
Perhaps you mean "investment" rather than "speculation"?
All mainstream European politicians are in favor of more growth, more investment, more industry, more jobs, and so on, across the continent. The parties who are about "de-growth" are marginal, even as the climate crisis is on full steam.
The actual problem is lack of pro-risk capital. Almost all of Europe's capital is concentrated in 1) conservative industry giants, and 2) huge conservative pension funds.
Basically all capital in Europe, public as well as private, is risk averse. There are very few actual "VC"s. I'm not sure that's unambiguously a bad thing, but it is the difference.
No I didn’t mean investment, I meant speculation. Because I was talking about individuals and risk capital.
We are talking about the same things.
Speculation drives innovation
Well, I'm of the opinion that one of the other blockers of major EU growth is the lack of a singular financial market; instead of one market like the USA, we have 27 markets. Twenty seven different sets regulations that all need to go out the window; exchanges in the double digits, most of which should shut down.
This balkanization will never lead to a growth-oriented modern financial system.
Europe needs to centralize capital.
Is OVH not a European cloud provider? There are also smaller players like Hetzner.
Hetzner is not small, I would say it's several times bigger than OVH in turnover, but this is just a guess based on their perceived markets and market share in those markets.
The cloud offered by Hetzner is the perfect sweet spot for me, but the vast majority of IT crowd and non-IT decision makers in EU want AWS-like.
But all EU alternatives that mimick AWS will be worse than the alternative.
Pretty much like wvery MS Word copycat was worse... until Google Docs shifted the paradygm.
OVH is larger. They occasionally burn down a datacenter or two and are still going.
Hetzner had revenues of €470m in 2022 (last available). It was €866m for OVH over the same period.
Source: S&P CapitalIQ
OVH and Hertzner sell lumber, people want furniture.
The fire they had didn't help their reputation:
https://www.theregister.com/2022/09/13/ovh_sbg5_opens/
OVH (french) is very well known and I like them a lot. Used them for domains a lot, because they are very cheap and their management is nice. I also like very much ScaleWay (french also) for price and quality of service, have used them for years on my startup, can highly recommend. Also heard a lot of good things about Infomaniak (swiss), but never used them myself.
Would love to hear about european cloud providers with comments from users.
From the same author about this very subject: https://berthub.eu/articles/posts/the-european-cloud-ladder/
Ah actually, more from today: https://berthub.eu/articles/posts/dear-hosting-providers-you...
They try hard to brand themselves as a cloud provider but I'd say that they mostly are a VPS provider.
The cloud side isn't polished enough to pretend to be a cloud provider.
I wouldn’t describe Google nor Microsofts products as ”polished”. Humongous maybe?
Yes, especially Azure's success seems largely driven by their generous free tier for startups and the lock-in of the Windows ecosystem.
While I like the user interface, after having used it for more than a year I've successfully stayed away from it ever since.
Meanwhile AWS' popularity was largely driven by EC2 servers which are VPS.
Also they are literally called OVHcloud...
Calling yourself a cloud doesn't make you one. Can you do auto scale groups with dynamically scaling load balancers yet on OVH? AWS has had that for 15 years now.
I dont think you realize how small scale majority of EU gov software has to be.
When you have country with population under 10mil and your gov form is used by 10% of those people a year… Thats 80k submissions a month, split by 20 workdays = 4k submissions a day over 8 work hours = 500 submissions hour or about 8 submissions a minute.
I know these are wrong numbers and there are peaks etc. But many would dare to put this on single server with sqlite. Even if you 10x that.
You do need autoscale groups or self healing architectures. Govs requirements are CRUD apps whos biggest issues are design, accessibility, data permissions etc. Not scale.
While the often advertised point of auto scaling is to handle some imagined large load, in reality it's used as a reliability principle.
For example, it enables seamless rollout strategies for frequent cicd. Any regressions in application performance are automatically handled by scaling up, etc.
> The cloud side isn't polished enough to pretend to be a cloud provider.
I mean, see Azure and Google Cloud a few years back. For quite a while the market was AWS, and also some joke services which nobody who wasn't required to used (notoriously, in 2012 Azure was substantially entirely down for _over a day_ due to a leap day).
China was able to build its own IT technologies thanks to decades of protectionism and smart maneuvers from the party. Europe, being under the US "umbrella", was in an almost impossible position to do so, along with the very poor attitude of European elites, which is one of rent-seeking rather than investment.
Regardless of the attitude, I don't see this happening without some form of protectionism, trump sperging out might be the chance of a century to reboot the European military and IT industries.
If Europe wants cloud protectionism then they need apply the usual methods: tariffs on foreign cloud services and requirements on public funds to be spent on the European cloud providers.
They could apply Chinese policy and require foreign cloud corporations to have locally owned partners that will own the technology produced locally.
See https://www.aboutamazon.eu/news/aws/aws-plans-to-invest-7-8-...
Of course, the question is very much how the courts will see ^
> They could apply Chinese policy and require foreign cloud corporations to have locally owned partners that will own the technology produced locally.
This feels like a misunderstanding of the situation. China used that tactic in cases where the Chinese people didn't know how to how to build or do a thing, and it would have taken a long time to rediscover. Saudi Arabia uses the same tactic for the same reasons.
Europe is full of people who know how to build a big cloud. It's not easy, and there's certainly learning-by-doing required, but as the author observes many of the people who work on the US clouds are from Europe or India or still live there. The knowledge transfer has been happening continuously since the start and in both directions.
The reason Europe hasn't produced a competitor on the scale of AWS is simply because that business model is the exclusive preserve of already large tech companies. AWS, Azure, GCP and Oracle Cloud are all the products of huge firms that already had tens of thousands of software engineers and (with the exception of AWS) pre-existing very profitable businesses. They could afford to sink vast sums of treasure into buildouts well ahead of demand, and subsidize their clouds using the other businesses until they had been able to catch up with Amazon. In addition, they all had large pre-existing tech ecosystems to leverage.
European countries have failed to produce tech companies on that scale for all sorts of well analyzed reasons. Moaning about cloud specifically is of no use, it's just a symptom not a cause. The causes meanwhile are the usual grab bag of uncompetitive compensation (a symptom of being poor, it's a feedback loop), bad laws that retard innovation and yet which never get fixed (due to the prevalence of governments that are explicitly suspicious of companies being successful), and finally the culture of US firms which is very globalist, collaborative and internally open, making them pleasant places to work for the people in Europe who have the right skills.
Look at it this way - would a skilled engineer rather work on AWS, having global impact under a global brand selling to every kind of customer imaginable and justifying its existence through technical innovation, or on an explicitly parochial "Eurocloud" whose highest dream is merely cloning what AWS was ten years ago, and which exists for no better reason than some domestic politicians being unwilling to resolve their differences with one specific US administration?
It's just a fundamentally unappealing proposition even if pay was good, therefore, it won't happen.
Even China has an Azure cloud though. I know because I work in it sometimes.
> have locally owned partners that will own the technology produced locally
This may be the way to go. It seems common in countries like China, India, Brazil, ME / Gulf countries.
Or subsidies or funding R&D?
Tariffs are a trade disabler, so in general a loose-loose proposition.
Even if you manage to build the product (R&D) now you have to incentivize the private sector and governments to switch to the EU product. There's only few options of making this happen and tariffs could be one of them. Natural transition won't happen for many reasons.
Tariffs are not particularly different to subsidies. Effectively they are a subsidy except that consumers are paying for it more directly (being forced to pay more for inferior local products) instead of through other taxes. The source of funding and effect on prices are the most meaningful differences.
> Tariffs are not particularly different to subsidies.
Tariffs are the exact opposite of subsidies; though a tariff applied globally to all imports of a class is, I guess, sonewhat similar to a subsidy to domestic products of that class.
In what way are they so different? Both increase prices for the the consumer it’s just they end up having to pay for the subsidy indirectly through taxes and such instead of having to pay a higher price. From the sellers perspective as long as subsidies are only available for domestic producers they are even more similar to tariffs since they reduce the competitiveness of imported products.
> In what way are they so different?
Subsidies give money to selected market participants.
Tariffs take money from selected market participants.
They are, exactly, opposites.
They are similar in that they are government interventions in the marketplace, which under simplistic ideal-market assumptions have net costs which must be born somewhere, but at that level every possible government action is "the same thing".
> Both increase prices for the the consumer it’s just they end up having to pay for the subsidy indirectly through taxes and such instead of having to pay a higher price.
No, they don't. "The consumer", "current taxpayers", and "future taxpayers" are different-but-overlapping groups of people.
Think further: we need a European DARPA.
There’s plenty of R&D funding in the EU. How would that help get a European cloud?
The recipe is simple but nobody supposedly wants to do it.
1. Find all (e.g. LinkedIn) the EU-based engineers working or having worked for American cloud companies.
2. Offer them a proportional salary to join and start building the EU cloud product.
3. Incentivize the engineers to top-performance by giving them bonuses, stocks, whatever ...
4. Go back to the garage engineering and cut-off the bureaucratic BS with endless PMs, milestones, JIRA boards, scrum masters, chapter masters, architects, and all other similar BS.
But EU doesn't want to do it - they rather spend money in ways so that "EU funds", which are supposedly public, end up running up to their pockets directly. Dozens of examples out there but "AI factories" being the last embodiment of that.
You describe a plan for building it. That is but one piece of the equation. How will you sell it? Why should people pick you over established AWS/Azure/GC? Will you target new projects, or get people to migrate? How does the strategies there differ? What is the timeline to becoming profitable? How will you fund it in the meantime?
I'm aware of that. Please see my other comment that already addressed your concern. The point rather was that EU strategy won't even get us so far because it doesn't have or doesn't want to have a plan.
The EU is ran by lawyers and political science majors. They have absolutely zero idea how to setup proper incentives. Any hope of EU getting its act together is wasted. Much like our tax euros.
Talking about solutions and helping them gain mindshare is the first step to make things happen.
There’s nothing in the EU quite like DARPA: oriented to long term, fundamental, high risk, high reward projects.
Furthermore, this European DARPA would be an organization with its own physical structure - its own labs. Say, with each different discipline in a different EU country.
But of course, the agile funding agency could be decoupled from creating more, broader reaching, European research institutes.
Ok, so very different from DARPA (the original) then.
Look into EU HORIZON programs.
Cloud is highly overrated for a lot of projects. While in theory you could scale up or down based on demand, my experience the need for it was rare. More common was provisioned resources sitting there for months. And S3 encourages you to basically save everything
And there was another aspect, since compute / storage was a consumable. Devops were less shy about spinning up services as if they were candy, rather than being more economical with resources that were fixed in the short term.
Cloud isn't just scaling.
It also takes care of backups, disaster recovery, maintenance, HW replacement and offers many services right out of the box that would require at least a single extra DevOps guy to do on-prem.
Managed databases and managed storage like S3 is where it's at.
Please can Australia join these initiatives. We have the same issues but are too small to do it ourselves.
Australia needs to have a rethinking of its geopolitics strategy. It still seems to think it's part of Europe while at the same time being one of the members of Five Eyes, which makes it an extension of USA and UK's intelligence network (which has become problematic from the EU's perspective). In reality, Australia's biggest trade partners are all Asian (China, Japan, South Korea, India). If anything, Australia should be seeking to join efforts with its Asian partners and neighbours.
I'd argue Aus should really seek a policy of neutrality, and strong Asian trade. Remaining neutral requires strong diversification and the EU is a great independent technology supplier to keep on board.
All true but I would like the option of a local European cloud provider because of the EU privacy rules.
So that you can AUKUS us once again after signing the first waves of contracts? :)
But we do contribute to Eurovision. That'll do as a trade right? :-)
The Eurovision Song Cloud - Europes next cloud provider.
If anyone can build a European cloud surely it must have an Euro-vision! /s
Some here think that we should have stayed with the French contract
Any free country should be able to join, maybe one day US can too
You also have Outscale (outscale.com), they are French, located near Paris. They are owned by Dassault Systèmes, a big French company.
I think the big problem is this: That NLNet thing he mentions gives subsidies of the insanely large amount of €5-50k that’s not even worth getting out of bed over, much less quitting your job on the off chance you are successful in the 3 months that gives you.
This will be harder-when AWS (and the others follow) launch the EU sovereign cloud[0]
Word is that it'll be run similar to the top secret (lol) cloud [1] that they run for the USA 3 letter agencies.
[0] https://aws.amazon.com/compliance/europe-digital-sovereignty... [1] https://aws.amazon.com/federal/top-secret-cloud/
The author addresses this in the text:
> There are also false prophets—we all want quick solutions, and sometimes we get (too) excited about people who promise them. For example, in Germany, some claim they have “made Microsoft’s cloud European” through SAP. However, this turns out to be not quite the case. If the U.S. refuses to cooperate, they (by their own admission) would be down within a few months. Amazon also claims they are going to do something similar, but they’ve been saying that for years, and they conveniently avoid addressing what happens in case of a conflict with the U.S.
An EU sovereign cloud by Amazon though would be built in the EU and staffed by EU employees though.
The effect of the US trying to insist it won't co-operate would be similar to what happened to most Russian based American franchise locations: troublesome but it's not like the facilities, people or inventory just vanished.
> An EU sovereign cloud by Amazon though would be built in the EU and staffed by EU employees though.
At that point, why is Amazon needed? Don't we really have the expertise and capital to do this in Europe without Amazon (or Google, MSFT)? I somehow doubt that very much.
All proprietary software that runs in the cloud would vanish.
Well, that's why code escrow exists as a scheme.
Make all tech knowledge available, spin it into another completely separate company with 0 ties to the parent and then we can actually start talking. Rest is fluff and PR.
Did you know alibaba cloud has a GDPR compliant datacenter in frankfurt [1].
1. https://www.alibabacloud.com/en/press-room/alibaba-cloud-lau...
> top secret (lol) cloud [1]
Not sure what the “lol” is for there. They aren’t saying the cloud is top secret. It’s qualified to work with top secret data.
I am really questioning that initial premise that "we need an European cloud". Here I roughly understand "we need a few big European hyperscaler with a dozen huge datacenters we can migrate to" because this is the model we understand. But you are fundamentally "fighting the last battle", the wrong battle.
The idea of an European cloud has happened several time before and I was once strongly encouraged to select one so I investigated, here are my findings:
About 10 years ago all those countries needed a cloud to put their government and sensitive industries on, so public money and contracts was given to friendly corporations. Once the initial chunk of free money ran out a lot of those companies got rid of the business. In France a lot got sold to Orange, the big telco partially owned by the state.
I considered it because according to some metrics they had the biggest cloud in Europe.
I found out quickly they actually immediately given everything to Huawei. So it was all managed by Huawei but if I remember correctly still hosted on Orange premises.
Strangely information about it is more difficult to find now but you can still read about it on third party websites [0].
By the way all of this is GDPR whatever compliant.
I just did a quick search and it seems they had to shut it down in 2024 because Huawei became such a radioactive topic. Information is hard to find about it in English.
In the end the companies who were forced to select a "sovereign cloud" probably went through hell and moved around several time in the last 10 years. You are told you should/must go on a sovereign cloud, you then end up on a Chinese cloud and you are then kicked out of it.
I am not pushing for the big american hyperscalers but be very careful before listening to politicians.
- [0] https://www.techmonitor.ai/hardware/cloud/orange-introduces-...
You have a French perspective, but EU is big and varied.
Here in Belgium we do have our own Government Cloud which, while hosted using Broadcom (VMWare) and IBM (RedHat Openshift) software is locally owned and controlled.
Now, this "cloud" is a misnomer. It is not as automated as AWS. I would rather call it a government hosting provider, but it's doing its job.
> Now, this "cloud" is a misnomer
Absolutely.
I have seen a lot and usually the smaller the better.
I remember the Flemish gov (7M people so quite small) was doing great in this domain. France where everything needs to be centralised, big and controlled is doing terrible I guess.
This is why starting the discussion with the concept of "EU" and "Cloud" is wrong. We do not need an aircraft carrier, we need a lot of drones.
> we need a lot of drones.
I agree but for a slightly different reason. EU cloud is not about chasing AWS or GCP or Azure scale and complexity. It's about hosting our businesses and most importantly governments locally.
Thus, "cloud" is not an objective but a means to a goal.
>But you are fundamentally "fighting the last battle"
In your opinion, what is the current/future battle, then?
It seems we are pushed into major conflicts and disruptions so we have to decentralise it as much as possible.
The current model is very awkward, we have been pushed into those very centralised, hyperscale solution that most businesses and individual do not really need. Most businesses do not really need much to do their job.
We are probably gonna have major connectivity disruptions on of those days or what happened when your Hertzner or Orange/Huawei datacenter is down? And I know they have "better security" but if we are attacked they are gonna be taken down anyway.
We do not wanna look at what a major cyber disruption would do to our societies. I believe just the consequences on the electricity grid and hospitals would make us wave the white flags in a few weeks.
You wanna own your hardware because if there are disruptions it is gonna be hard to get your hand on some because of supply, logistics or just monetary disruptions.
How hard is it technically to setup something like Ubuntu Microcloud [0] for your own use or local businesses you provide that service for?
From my experience the biggest risk and headache is all those EU regulations like the EU Cyper resilience Act and in some cases energy costs.
So when Ursula tell you Europe needs to prepare for war and authorise/push your country to go into even more debt so her friends can build expensive tanks that will never see a battle it is ridiculous.
They should focus on removing all the obstacle to resilience they setup themselves, find a way to reduce energy costs (pretty hard now...) and maybe encourage moving out of those big tech cloud (starting by doing it themselves).
- [0] https://canonical.com/microcloud
>It seems we are pushed into major conflicts
>so her friends can build expensive tanks that will never see a battle
I am not sure how you consolidate the two statements. It seems to me that there is a very real risk of armed conflict in Europe over the next decade.
Yes there is good chance we are being pushed into one.
What I mean is the tanks, aircraft carriers or F35s probably won't really matter.
You can create enough pain on current western societies by other means such as major cyber disruption so that the classical battlefield won't really matter.
You could have said the same about Ukraine. Assuming rational (even if transactional) moves from dictators seems naïve.
I am not OP but let me give you a Microsoft Office example.
For many years, Open Source and businesses tried to make a Microsoft Office competitor by mimicking Microsoft Office. Naturally, all clones were worse than the original.
Until Google changed the paradigm and rolled out Google Docs that had a unique feature of online collaboration.
Then was the turn of Microsoft to mimic collaboration features of Google Docs in Microsoft Office and be worse at it almost by definition.
Another example is S3.
For years, businesses tried to have POSIX-capable filesystems seamlessly scale in size and in availability. It took Amazon to roll out a simpler alternative that, by having a smaller set of features, enable so much sought properties of distributed file systems in an efficient and commercially viable way.
I think Hetzner Cloud is a sweet spot of a cloud. Instead of reimplementing AWS, EU should standardize of something like it.
The cloud is an anti-sovereign concept. Maybe Europe doesn't need other people's computers, but their own computers instead.
You are so right about computers, except from Raspberry Pi (UK) not much as been done in Europe.
Regarding the eu cloud, it is definitely about sovereignty, specially since the CLOUD Act (H.R. 4943). In the context of the global trade war initiated by the US of A, it also makes sense from a European Union perspective.
A "cloud" is just other people's computers, so a "sovereign cloud" is an oxymoron; renting other people's computers is in no way sovereign.
I’m dreaming of a set of individual providers, all building with similar dev experience, auth, billing functions, quality of service etc.
I own https://mailpace.com and would love to expand from there- reach out if you want to help.
Sorry to be a killjoy, I see these sentiments everywhere across the EU sovereign cloud space. The dream of distributed "cloud providers" made up of individual companies products, all cooperating and communicating together in some utopian way. That's the stuff that Gaia-X was huffing.
Make no mistake, this will absolutely be curb stomped by vertical integration.
But I get the sentiment, we need to figure out a way of collaborating on this. Perhaps airbus is a good example.
Although you are from UK, so not the european union anymore, you could try getting listed on https://european-alternatives.eu/category/transactional-emai...
Anyway yes, we need to better list and advertise the european alternatives for services like yours.
No he couldn't. That site uses a non-geographical definition of Europe that excludes Britain. "EU, EEA, EFTA, or DCFTA member country" also excludes Serbia, Turkey, Belarus and Russia but includes several countries that aren't able to join the EU at all due to corruption or misaligned legal systems.
This really shows the incoherence of the whole Euro project. There's no such thing as European-ness: when the sort of people who wave the blue flag use the word Europe they are imagining an ideological construct subject to inconsistent and ever-shifting definitions. They don't even agree with each other what European means. One minute Britain is in Europe and Ukraine isn't, events happen, and suddenly Ukraine is European via DCFTA and Britain isn't. Switzerland is a similar case: sometimes it's considered to be European by these types, and other times not.
Why should I, a man born in Britain now living in Switzerland who has worked on two different US clouds, want to apply that experience to a Eurocloud given this history? This of thing is why it will never inspire much loyalty, and why Bert instantly gives up on the idea of a European cloud being used because it's actually good. The resort to force of law underpins the entire project because the European identity is a sort of social engineering programme, not something organically developed.
To be fair, the non-geographical definitions that excludes Britain, actually only excludes Britain because Britain excluded itself of the European Union in 2020.
And yet that website doesn't use the EU as a definition of European, so it clearly doesn't matter in this case. That's what I'm getting at: the word European doesn't seem to mean anything because the people who use the word the most are relying on definitions that yield unintuitive and self-defeating outcomes, like deliberately excluding one of the countries in Europe that actually does have a bit of a tech ecosystem.
You are right on that, not including the UK in a European list (whatever the reasons) is a shame.
I've done quite a bit of government contracting for governments in Europe, including various government levels in Belgium and the EU itself. Every single time, I encountered a lot of friction regarding hosting. The applications were always hosted by a partner [1] that won a large public tender, usually by virtue of supposedly being the cheapest. Those hosting partners were invariably awful and incompetent. Simple tasks would take weeks. Complex tasks wouldn't happen at all. There was enough red tape to go to the moon and back.
Every project I worked on, developers (consultants like me) advocated for finding a way of bypassing the hosting partner, and either doing their thing on top of platforms like Hetzner, or switching to AWS/Azure. Not because they particularly love AWS, but because the default hosting solution sucks so much arse.
I can totally see EU cloud infrastructure working for the public sector, but ONLY if they come with the same level of self-service management as the big cloud providers. Without that, it's a non-starter.
[1] One of those is a large French company on the verge of banruptcy.
In my experience this is mix of lack of knowledge sometimes mixed with self-dealing/corruption.
But it can be prevented by writing hosting specs well. If in your proposed solution you specify really detailed hosting envinronment then most of these old incopetent companies wont even apply.
Unfortunately, the way my governments have dealt with the issue is by writing one very large tender, from which a Hosting Framework Agreement is derived. Individual projects have no say in it, and they're certainly not consulted when the agreements are drafted. So basically they're stuck with whatever is on offer, which isn't much.
A choice example was a hosted K8S solution: it was supposed to be highly available, but the framework agreement only specified 2 datacentre locations. The hosting provider thus couldn't build a reliable stretched cluster, and instead built two identical clusters. Forcing every project to figure out how to deploy their applications in a HA fashion on top of two clusters. Much fun. Not a giant waste of resources at all. /s
I have been working for large European utilities and even within AWS/Azure deploying something takes weeks.
Fairly certain this is true for large anythings if they haven’t been explicitly set up to avoid that problem.
Can you give a hint for [1] ?
There are large European data centers run by European telecoms, including Telia and Altice. So there's something to scale up from.
Cloud is easy nowadays.
There is everything:
Servers/Storage/Network/Linux/Kubernetes
Experts are still rare but people can learn.
Hmm one other thing - USA is completely English but EU is quite multi-lingual so to be competitive within EU market services need a lot of more efforts.
If you take the triangle between London - Paris - Randstad (Amsterdam+) then we're talking about 60+ million people all connected via <4 hours of high speed rail where essentially everyone can speak English.
(Okay maybe in Paris only people under 40 can but in IT it's usually okay imo).
You don't need the entire EU, just a chunck of it that can carry the industry like Silicon Valley does for the US.
[dead]
Do onerous EU regulations have to do with the inability to produce a competitive cloud locally?
> The European cloud industry has struggled to take off, partly because existing players lacked the will or vision to compete effectively (among many other reasons).
The EU would have come after them. It is coming after Google, AWS and Azure even though they have diplomatic support from the US proper; how is an EU cloud provider supposed to compete?
The people with the will and vision exist, if anyone can build EuroCloud they would be wealthy beyond the dreams of mortals and it really just needs to be an OK AWS clone. The unfortunate part is that anyone who sees that far can also see how the story will end when the bureaucrats get involved early instead of after they have multiple billions for compliance. It'll cripple any ability the project has to compete.
It doesn't have to compete 1:1 when tariffs and regulations get applied, same principle as in all other industries ie cars.
Its not the best approach but it can work, we have half a billion rich domestic market and some subsidies to kick start things would be rewarded in long run tremendously.
Too true!
The reason why Europe won't have a cloud is because you can't have managed decline and nice things. This is the same reason Europe will forever be depdent on someone else for defence and why it will always be the meek but somehow petulant follower of whoever is the world hegemon. This is why China and Russia cares so little about them also, they know that Europe will fall in line once their current patron no longer can or want to sustain them.
That would be believable 2 years ago, but right now it seems like you're engaging in wishful thinking.
A few years ago I would have maybe said the same but we're now entering a new era where the US capital will likely go towards the EU which is seen as more stable and future oriented.
The US tech companies where somewhat insulated in Europe thanks to the US defense but now that is out of the picture, I don't see much else to protect them.