I find the Europe's relationship with tech to be wired, there is one section that is hardcore-opensource fanatics, they want to host everything by themselves, and want to go through the trouble of keeping things updated, and would not want to use a close source tools even though they are developed by European counterparts.
On the other side there are people who are techy but happy to use US products, and when you pitch something European they would cite some tool that's better and bigger in US.
It's hard to find people who are in the middle who would like to pay and use a EU made tool.
Also processes take forever, and everything has to go through lot of meetings, and bureaucracy and red-tape and no one is willing to take a chance on a small startup.
> I find the Europe's relationship with tech to be wired
I don't think it's weird: almost nobody cares, they just use whatever they know/is free. It turns out it is US tech. It's the exact same situation in the US, except that for them it is not a sovereignty issue.
Now maybe there is a bigger open source community in Europe, but I don't see a problem with that.
You switch from talking about Europe to talking about the EU half way through. The article was about Europe (excluding Russia and a few others).
> there is one section that is hardcore-opensource fanatics, they want to host everything by themselves, and want to go through the trouble of keeping things updated
Using Cloudflare, AWS etc. does not mean you do not have to keep things updated. Using an SaaS does. The numbers in the article count both.
There are plenty of people who use FOSS only and non-US hosting, and still use Cloudflare.
> On the other side there are people who are techy but happy to use US products, and when you pitch something European they would cite some tool that's better and bigger in US.
A preference for what they already know (maybe reinforced by marketing). Its not that they prefer American products, but American dominance means it is what everyone already knows.
Well it's because few people have "European-ness" as a strong personal value. Some people have strong values around open-source, or even around the specific country, but the sense of being European and valuing European things is just not very widespread, so in absence of a specific personal value, they pick the cheapest/biggest/most-known option which is usually American.
This is quickly changing though: my subjective take is that the US antagonism is pushing people away from American product AND making the European identity stronger.
Computer software is so incredibly cheap in a business setting (that includes public sector) compared to all other tools and expenses, that it always makes sense to pay for and use the most feature complete software you can get.
Wait to see what they are using for emails and for most of their internal docs (containing any kind of secrets)!
I know companies that will tell you "I'm not gonna put any of my data in cloud, especially not American ones" but they are perfectly fine using any major cloud based office suite (mail, docs, chat/video apps, ecc ecc) where they voluntarily and deliberately load any kind of data.
Good article because it clearly exposes the methodology and the shortcomings of the measurements (mostly the front CDN of a ~20k number of old continent entities of apex/www domain).
> The practical point is not that every European company should leave US infrastructure tomorrow. The point is that sovereignty discussions often start too late in the stack. Before organisations debate cloud regions, subprocessors, or contractual controls, they should know which vendors already sit in front of their public web estate.
This seems like the wrong takeaway and I'd advice (European) companies to do the opposite: Don't look at what your marketing/landing page does first, care first about where your actual user/company data lives, what processors are touching it and so on. Then once you have your internal house in order, then do the easy surface-level stuff like what vendor sits in front of your marketing websites.
I don't understand why they'd advise people to do things in that suggested order, seems really backwards and like they're more interested in patching over the problem rather than actually solving it.
> For European infrastructure vendors, this is the market map. For policymakers, it is the base rate. For buyers, it is the inventory problem.
> Dammit, fell for another AI slop article AGAIN...
Is it really "slop" if you keep falling for AI written articles again and again? Doesn't that actually mean the opposite - the AI seemingly convinced you (or almost) that it was written by a realy human?
They pointed out that the article is suggesting a completely wrong solution.
If the article was written by a human, it would be valuable feedback for the author. Because the article was written by a LLM, it was just pointing out that slop is slop.
LLM text is like an optical illusion for the language part of your brain, only instead of the payoff being "Oh, cool! The dots aren't actually moving!" it's "Oh, 'cool'. You weren't actually trying to tell me anything worth my time."
> Don't look at what your marketing/landing page does first
True. Also: the landing page and all the other corporate boasting are often run by a marketeer or similar, who knows nothing about the technical part. They just have a design, and outsource implementation. So it just ends up somewhere.
This tries to capture Europe as a single coherent market, which it is not and by far.
It's comparing countries with vastly different socio-economical landscapes and sizes.
Is "using Cloudflare as a CDN but hosting everything at, say, Hetzner using generic systems/opensource components" the same as "having built a complete ecosystem with Amazon specific software"?
Getting out of the walled garden of AWS, GCP or Azure is notoriously difficult. Some european cloud providers made this one of their key selling point, advocating for openess and "multicloud". This had, to my knowledge, next to no effect.
Vendor lockin is real. Dependency to a vendor located outside of your generic law system is, indeed, a risk. But this article probably isn't the way to measure it (and it's a tough job to do)
As an Italian solo-founder, I have to admit that the US vendor dependency is really strong, but when you look at what you need to build a serious product, what can you actually use from European vendors that is even close to US products?
Take Stripe as an example: is there a real alternative that covers what they do?
Not to talk about Cloud and Edge Computing vendors: GCP, AWS, Cloudflare... does anyone even get close to these products / companies and what they offer? Managed environments, automatic scaling, serverless architectures that just work and cover all your needs?
I'm a big fan of Hetzner, which has great prices, a great managed environment and lot of features that give you a reliable structure to work on, but I don't actually want to manage everything by myself.
I also use Bunny.net for my products, but the services are still limited and contained to very specific stuff.
Just take a look at Neon Postgres as an example: where do you find a product like this in Europe?
I believe that the problem is mainly structural and cultural. When a new technology comes out, it's usually from US researchers and companies. So how does Europe even stand a (real) chance at giving the world (or the continent) the best packaged services?
Asking for a complete replacement is how people get locked in. The way governments end up with m365 by putting out an RFP with the exact feature list of m365. Of course then you limit yourself to only one option.
I think this will start moving a lot now that people are really aware of American dependency and also ethically are opposed to it.
It's not about complete replacement. People don't buy a Golf because they googled "Toyota European version" and decided the limitations are acceptable, or because it has all the same buttons as a Hyundai. They do it because Volkswagen and a number of other European manufacturers make genuinely good, competitive cars.
There isn't going to be "EU tech" as long as US can access the EU market as freely because for the mainstream anything it doesn't make sense to have a duplication effort for the EU. USA is/was easier to start business and get funding, you can access Europe just as well from there so why bother?
After Trump now there is a reason and EU is trying hard to clean up the field with things like EU-inc, digital Euro, common markets etc but its not happening fast enough to make a difference today.
Maybe if all goes well and Trump can finish his term or even invades Greenland then EU can have its "tech", but for now its happening slowly because its primarily driven by the hypothetical risks that are convincingly real but but costly to act on.
> USA is/was easier to start business and get funding
I think this would highly depend on the country. With a solid business plan, I could easily get funding via banks and literally start a company with the press of a button in web portal, in Sweden. Similarly, Estonia seems to have made it ridiculously easy as well. In Spain it's slightly harder, I have to fill out some forms, but with stable income, very easy to get a bank loan even for business ideas that probably shouldn't.
Sure, you won't attract multi-trillion VCs that route, but is that exclusively what you're talking about? How much easier can it be to start the company than the press of the button, since you seem sure it's much easier in the US than all the countries in the EU?
Not automatically. If the loan is in your name, or you personally guaranteed it, you still have to pay even if the business fails. If it's only in the company's name, usually the company is liable, not you personally. There's also insolvency/"second chance" options, but that’s a legal process, not automatic.
Basically, depends on how you setup the company in regards to liabilities (which I'm guessing is true in the US too), and the exact terms of the loan.
Totally agree. As you said, the gap is significant and closing it requires a lot of effort and investments to cover a risk that could eventually cool down if the general situation stops being this critical.
The problem is that many of europe's top talent moves to USA, and also that American funding for European startups is huge, and many are made to relocate/IPO in USA too.
For cyber security examples there are real movement in europe - ciphercue is trying to create a directory of alternatives which is gaining traction - https://ciphercue.com/directory/eu
Maybe going beyond cyber alternatives is whats needed
For payments I've used Mollie, which is EU based and has good DX. It also has something critical that Stripe does not: proper customer service. Adyen is also good.
You're a bit generic on other fronts but Scaleway is an excellent EU-based cloud provider. Haven't missed anything from AWS so far.
> Just take a look at Neon Postgres as an example: where do you find a product like this in Europe?
Why would you even want such a product? Managed postgres for cheap in the low 10s of Euros will scale to lots of users.
If your problem is scaling you have the best problem in the world, and one that most cloud vendors offer solutions for out of the box.
Neon is a great solution..which fits a handful of use cases.
US definitely enjoys an apex position in cloud services, but there's little to no core irreplaceable products beyond leading edge AI in European offerings.
Cloudflare might be an exception if latency, ddos protection and global reach are important.
There are plenty of alternatives to Stripe. Just search for card processors. If you’re a high-risk merchant, or your product falls into a high-risk category, Stripe will probably reject you. That’s why there’s a big market for other processors to step in.
Adding for somewhat completeness sake those are 13 anycast IP addresses with upwards of 2000 servers responding to them. The US has about 500 to 700 of those servers. The EU about 600 servers and the rest of the world 600 to 900 but I do not have exact numbers. One of the root DNS server admins would have to chime in for more accurate numbers. I am fairly certain that the EU servers belong to companies in the EU not that it entirely relates to the overall management hierarchy but I don't know their political structure.
Africa and the middle east have the least, something like ~100 or so.
The anycast is for the /24 those IP addresses share (first.three.octects.0/24)
But if you look those /24 are Arin.
Dig +trace shows the recursive lookup (forwarders) used for your NS.
They almost always end up Arin and arpa. I've troubleshot some connection issues between different data centers and transit providers and found root hint oddities I just escalate to my supervisor
That part makes sense to me. If one has ever tried to announce an ARIN IP in the EU or a RIPE IP in the US they will send warnings that the allocations will be pulled back if that is not fixed. Been there, done that. The root servers were created in the US. To have RIPE IP addresses I would imagine would require ARIN to set up some partnership but there again that gets into management and politics in their organization that I am unaware of. Every time I start to read any of their minutes they lose me after the first paragraph.
If the concern is that the US would somehow break the root servers and disrupt many trillions of dollars of trade I guess that is technically possible but probably unlikely given the amount of trade, tariffs, tax revenue that would impact would end anyone's political career and things would be reverted quickly in my opinion.
> If one has ever tried to announce an ARIN IP in the EU or a RIPE IP in the US they will send warnings that the allocations will be pulled back if that is not fixed.
ARIN, RIPE, APNIC, AFRINIC, LACNIC. If you have announced addresses from one in the others region and have not received warnings it's just a matter of time. ARIN is the most vigilant. There was a time they used to do ping sweeps and nag people about not utilizing their addresses which led to my putting a 1U server on a /17, /19 and some smaller networks with Labrea responding to pings for all of it. They stopped doing that which led to people hoarding IPv4 space and now putting a monetary value on it which is dumb dumb dumb.
Surely that not the case. Check any public BGP looking glass: Arelion/Telia uses RIPE space in the USA, Lumen/Level uses APNIC space in Europe and so on.
ARIN are the most aggressive about it. Maybe others just don't care any more. They certainly used to. I would never hear the end of it. Perhaps others pushed back enough to stop that practice elsewhere.
TLDR: Yes, ofc we use Microsoft, Amazon (AWS), Cloudflare and Cisco...
There is even mainstream press articles about it here in Sweden. "dependance on microsoft ooh so bad" etc.
I find it laughable.
Unless you have a time machine to 2005 (EC2 came out in 2006 that should have been the signal) there is no way to compete now. That train has left the platform.
Second, Nokia and Ericsson dominate mobile infra in the west, but that is good I guess as they are EU? What does USA think about that?
Third, let us say you get rid of MS. Now you have no MS but all network infra for broadband is Cisco, Huawei, Juniper etc. Good luck ripping that out. And for what?
Same with AI. Mistral was amazing at first, Le Chat. Almost as good, generous free limits, good docs. Now? Just plain bad. Deepseek is better (I dislike china so I avoid it). EU should have gone in 500% the moment Mistral showed promise.
But lately we let USA and China take the lead on everything and EU can write a strongly worded letter after about how bad it is.
People will "care" when EU starts making good stuff again.
And lastly lol, people do know everything ends in Taiwan in the end right?
Let the yellow man put some more tariffs on imports from the EU. Let him ban EU citizens from Visa/Apple/Microsoft when they disagree with MAGA. Maybe give east Europe to Russia, the west half can be the 51st, 52nd and 53rd states.
As long as there are no options we have no freedom.
I did a similar study but analyzing actual API subdomains, and ignoring those fronted by Cloudflare, Akamai, etc and the conclusion was the opposite: European companied are more likely to be using OVH and Hetzner.
AWS / GCP can invoice with EU reverse charge so I don’t even pay VAT. With Hetzner I have to claim it back at the end of the year.
Hetzner is just way cheaper and pricing is more predictable. I don’t need any advanced cloud offerings. And yes, being outside the US is another advantage rn.
It's not "the company" that makes such decisions, but some senior managers or directors who do. They don't see a penny extra in their paycheck for savings money, but they are on the hook if anything goes wrong with less mainstream tech choices.
Given current prices I doubt anyone is itching at the bone to spec out hardware.
Idk where you would even buy registered DDR5 dims right now. Newegg shows $Request a quote money and that its built to order ... for a 64 GB stick lol.
I'm amazed any company pays AWS for servers. I can get if you're paying for a service like media transcoding, but to put it in perspective, AWS normal prices are now about on par with Hetzner's 5x-increased-due-to-the-crisis new price for new servers
-most companies's are not global, and so most of their customers are going to be in the same country or nearby so makes sense to keep near the serves close.
This doesn't say anything about which providers are the majority in Europe, which is what OP's study is about. Your findings (non-US companies are "more likely") and OPs findings (US companies are the majority) are both compatible.
Here is a hint: quite many companies need a certificated provider who not only is certified but also has - 99,999% overlook this fact because of ignorance - enough insurance and can theoretically be held responsible financially for services being not reachable.
EU regulation requires such settings.
While I highly value Hetzner and Strato, they don't want to deal with such companies, which is reasonable.
Also what you see is just the visible part, not the internal APIs or due to failure safety different APIs serving in a failover scenario.
Internal networks are huge. And masked or hidden behind quite some intrusion detection as well as layers of protection exactly for this reason.
In other words: you did an interesting posting however it is meaningless without knowing why these what you called churn occurs.
Usually you don't simply migrate from one cloud to another.
Accenture for example had a partnership with Amazon, and use their services. So maybe during the development phase or whatever there appears to be a spike.
In other words: it was a planend. Times series need to be observed for many years.
But nevertheless a nice posting.
It is just that sometimes "facts" from the outside lead to speculations, an insider can only chuckle about.
Before I joined a huge global bank, I thought any startup would eat them alive, think N24. Remember N24? No? Well...
I was part of the senior management, dbCORE as a hint.
Maybe repeat the study, make it run over years, or even better: observe something you know for a fact and see how things change, not the other way around. You would have needed to conduct interviews etc.
By design there is a paradigm called security through obscurity. That's why torture for example seldomly helpful. Is the poor soul lying or not? You need to verify first, in any case.
Outside observation is just that: Platon's allegory of the cave. Useful or not, you never know. That's why I laughed a bit about your disclaimer ("Limitations").
This is just incorrect with a way too small set of websites. Their estimates are more than double.
However, that is if you take all websites into account. If you only take the most popular websites/biggest companies, their estimates are closer to reality.
Europe right now is the ultimate US vasall. Germany is the leader here; France and Netherlands are much more self-conscious but also way too dependent on US corporations. The worst part is that in Germany with Merz in charge, this will not change. He is a good puppy for the USA.
115 comments
[ 3.4 ms ] story [ 49.0 ms ] threadOn the other side there are people who are techy but happy to use US products, and when you pitch something European they would cite some tool that's better and bigger in US.
It's hard to find people who are in the middle who would like to pay and use a EU made tool.
Also processes take forever, and everything has to go through lot of meetings, and bureaucracy and red-tape and no one is willing to take a chance on a small startup.
I don't think it's weird: almost nobody cares, they just use whatever they know/is free. It turns out it is US tech. It's the exact same situation in the US, except that for them it is not a sovereignty issue.
Now maybe there is a bigger open source community in Europe, but I don't see a problem with that.
> there is one section that is hardcore-opensource fanatics, they want to host everything by themselves, and want to go through the trouble of keeping things updated
Using Cloudflare, AWS etc. does not mean you do not have to keep things updated. Using an SaaS does. The numbers in the article count both.
There are plenty of people who use FOSS only and non-US hosting, and still use Cloudflare.
> On the other side there are people who are techy but happy to use US products, and when you pitch something European they would cite some tool that's better and bigger in US.
A preference for what they already know (maybe reinforced by marketing). Its not that they prefer American products, but American dominance means it is what everyone already knows.
This is quickly changing though: my subjective take is that the US antagonism is pushing people away from American product AND making the European identity stronger.
Self-hosting mail infra is a forgotten skill, indeed
I know companies that will tell you "I'm not gonna put any of my data in cloud, especially not American ones" but they are perfectly fine using any major cloud based office suite (mail, docs, chat/video apps, ecc ecc) where they voluntarily and deliberately load any kind of data.
This seems like the wrong takeaway and I'd advice (European) companies to do the opposite: Don't look at what your marketing/landing page does first, care first about where your actual user/company data lives, what processors are touching it and so on. Then once you have your internal house in order, then do the easy surface-level stuff like what vendor sits in front of your marketing websites.
I don't understand why they'd advise people to do things in that suggested order, seems really backwards and like they're more interested in patching over the problem rather than actually solving it.
> For European infrastructure vendors, this is the market map. For policymakers, it is the base rate. For buyers, it is the inventory problem.
Dammit, fell for another AI slop article AGAIN...
Is it really "slop" if you keep falling for AI written articles again and again? Doesn't that actually mean the opposite - the AI seemingly convinced you (or almost) that it was written by a realy human?
If the article was written by a human, it would be valuable feedback for the author. Because the article was written by a LLM, it was just pointing out that slop is slop.
True. Also: the landing page and all the other corporate boasting are often run by a marketeer or similar, who knows nothing about the technical part. They just have a design, and outsource implementation. So it just ends up somewhere.
There aren't many completely european solutions, but there are more than zero.
https://xkcd.com/932/
It's comparing countries with vastly different socio-economical landscapes and sizes.
Is "using Cloudflare as a CDN but hosting everything at, say, Hetzner using generic systems/opensource components" the same as "having built a complete ecosystem with Amazon specific software"?
Getting out of the walled garden of AWS, GCP or Azure is notoriously difficult. Some european cloud providers made this one of their key selling point, advocating for openess and "multicloud". This had, to my knowledge, next to no effect.
Vendor lockin is real. Dependency to a vendor located outside of your generic law system is, indeed, a risk. But this article probably isn't the way to measure it (and it's a tough job to do)
Right on the front page...
Take Stripe as an example: is there a real alternative that covers what they do? Not to talk about Cloud and Edge Computing vendors: GCP, AWS, Cloudflare... does anyone even get close to these products / companies and what they offer? Managed environments, automatic scaling, serverless architectures that just work and cover all your needs?
I'm a big fan of Hetzner, which has great prices, a great managed environment and lot of features that give you a reliable structure to work on, but I don't actually want to manage everything by myself.
I also use Bunny.net for my products, but the services are still limited and contained to very specific stuff.
Just take a look at Neon Postgres as an example: where do you find a product like this in Europe?
I believe that the problem is mainly structural and cultural. When a new technology comes out, it's usually from US researchers and companies. So how does Europe even stand a (real) chance at giving the world (or the continent) the best packaged services?
I think this will start moving a lot now that people are really aware of American dependency and also ethically are opposed to it.
And you know their infra generally works, you know when it's not on fire.
And when it is, well you are not going to be the only one down.
After Trump now there is a reason and EU is trying hard to clean up the field with things like EU-inc, digital Euro, common markets etc but its not happening fast enough to make a difference today.
Maybe if all goes well and Trump can finish his term or even invades Greenland then EU can have its "tech", but for now its happening slowly because its primarily driven by the hypothetical risks that are convincingly real but but costly to act on.
I think this would highly depend on the country. With a solid business plan, I could easily get funding via banks and literally start a company with the press of a button in web portal, in Sweden. Similarly, Estonia seems to have made it ridiculously easy as well. In Spain it's slightly harder, I have to fill out some forms, but with stable income, very easy to get a bank loan even for business ideas that probably shouldn't.
Sure, you won't attract multi-trillion VCs that route, but is that exclusively what you're talking about? How much easier can it be to start the company than the press of the button, since you seem sure it's much easier in the US than all the countries in the EU?
Basically, depends on how you setup the company in regards to liabilities (which I'm guessing is true in the US too), and the exact terms of the loan.
USA population: 348 million
The problem is that many of europe's top talent moves to USA, and also that American funding for European startups is huge, and many are made to relocate/IPO in USA too.
Companies also move to the US (e.g. Stripe) or are bought big American companies (e.g. Deepmind)
Maybe going beyond cyber alternatives is whats needed
For payments I've used Mollie, which is EU based and has good DX. It also has something critical that Stripe does not: proper customer service. Adyen is also good.
You're a bit generic on other fronts but Scaleway is an excellent EU-based cloud provider. Haven't missed anything from AWS so far.
> Just take a look at Neon Postgres as an example: where do you find a product like this in Europe?
Why would you even want such a product? Managed postgres for cheap in the low 10s of Euros will scale to lots of users.
If your problem is scaling you have the best problem in the world, and one that most cloud vendors offer solutions for out of the box.
Neon is a great solution..which fits a handful of use cases.
US definitely enjoys an apex position in cloud services, but there's little to no core irreplaceable products beyond leading edge AI in European offerings.
Cloudflare might be an exception if latency, ddos protection and global reach are important.
Africa and the middle east have the least, something like ~100 or so.
Dig +trace shows the recursive lookup (forwarders) used for your NS.
They almost always end up Arin and arpa. I've troubleshot some connection issues between different data centers and transit providers and found root hint oddities I just escalate to my supervisor
If the concern is that the US would somehow break the root servers and disrupt many trillions of dollars of trade I guess that is technically possible but probably unlikely given the amount of trade, tariffs, tax revenue that would impact would end anyone's political career and things would be reverted quickly in my opinion.
Who are "they" in this case?
TLDR: Yes, ofc we use Microsoft, Amazon (AWS), Cloudflare and Cisco...
There is even mainstream press articles about it here in Sweden. "dependance on microsoft ooh so bad" etc.
I find it laughable.
Unless you have a time machine to 2005 (EC2 came out in 2006 that should have been the signal) there is no way to compete now. That train has left the platform.
Second, Nokia and Ericsson dominate mobile infra in the west, but that is good I guess as they are EU? What does USA think about that?
Third, let us say you get rid of MS. Now you have no MS but all network infra for broadband is Cisco, Huawei, Juniper etc. Good luck ripping that out. And for what?
Same with AI. Mistral was amazing at first, Le Chat. Almost as good, generous free limits, good docs. Now? Just plain bad. Deepseek is better (I dislike china so I avoid it). EU should have gone in 500% the moment Mistral showed promise.
But lately we let USA and China take the lead on everything and EU can write a strongly worded letter after about how bad it is.
People will "care" when EU starts making good stuff again.
And lastly lol, people do know everything ends in Taiwan in the end right?
Let the yellow man put some more tariffs on imports from the EU. Let him ban EU citizens from Visa/Apple/Microsoft when they disagree with MAGA. Maybe give east Europe to Russia, the west half can be the 51st, 52nd and 53rd states.
As long as there are no options we have no freedom.
But then okay, say you want to do more.
Then do more. Not some letter from EU.
There is a focus here on office and javascript... why?
Okay if you want an EU M365, good luck.
https://bloomberry.com/blog/we-analyzed-50k-apis-heres-which...
Hetzner is just way cheaper and pricing is more predictable. I don’t need any advanced cloud offerings. And yes, being outside the US is another advantage rn.
Hetzner can do the same as long as you have a valid VAT code.
Idk where you would even buy registered DDR5 dims right now. Newegg shows $Request a quote money and that its built to order ... for a 64 GB stick lol.
-data is kept in EU,
-lower latency for SSH if nearer
-most companies's are not global, and so most of their customers are going to be in the same country or nearby so makes sense to keep near the serves close.
What exactly were you looking for? APIs?
Here is a hint: quite many companies need a certificated provider who not only is certified but also has - 99,999% overlook this fact because of ignorance - enough insurance and can theoretically be held responsible financially for services being not reachable.
EU regulation requires such settings.
While I highly value Hetzner and Strato, they don't want to deal with such companies, which is reasonable.
Also what you see is just the visible part, not the internal APIs or due to failure safety different APIs serving in a failover scenario.
Internal networks are huge. And masked or hidden behind quite some intrusion detection as well as layers of protection exactly for this reason.
In other words: you did an interesting posting however it is meaningless without knowing why these what you called churn occurs.
Usually you don't simply migrate from one cloud to another.
Accenture for example had a partnership with Amazon, and use their services. So maybe during the development phase or whatever there appears to be a spike.
In other words: it was a planend. Times series need to be observed for many years.
But nevertheless a nice posting.
It is just that sometimes "facts" from the outside lead to speculations, an insider can only chuckle about.
Before I joined a huge global bank, I thought any startup would eat them alive, think N24. Remember N24? No? Well...
I was part of the senior management, dbCORE as a hint.
Maybe repeat the study, make it run over years, or even better: observe something you know for a fact and see how things change, not the other way around. You would have needed to conduct interviews etc.
By design there is a paradigm called security through obscurity. That's why torture for example seldomly helpful. Is the poor soul lying or not? You need to verify first, in any case.
Outside observation is just that: Platon's allegory of the cave. Useful or not, you never know. That's why I laughed a bit about your disclaimer ("Limitations").
However, that is if you take all websites into account. If you only take the most popular websites/biggest companies, their estimates are closer to reality.
Source: I have access to better data.