The issue isn't expanding data centers -- that's "easy". The issue is all of the software tooling, all of the gigantically massive platforms and services they have to not only build, but catch up to AWS/Google/Azure, all of which are moving targets.
AWS launched in 2002. Google Cloud (GCE) basically started in 2012 and only recently has finally mostly caught up to become a fully-fledged viable competitor, which took an entire decade. Meta would be too late to the party.
Sure they could launch something much smaller like a Digital Ocean competitor -- not a fully-featured cloud, mostly just servers and storage. But why would they? The margins are low and there's no strategic reason to do so.
Keep in mind that a hosting partner has physical access to your precious hardware, something that you yourself probably don't even have. As such they are ideally placed to violate your trust six ways from Sunday and if they don't have a clean track record in that respect and you go into business with them anyway you only have yourself to blame.
Apple on the other hand will be looking at the cloud market and be thinking how can they do it differently? Not just compete selling a commodity, be create a step change in the market. I could see them utilising their silicone expertises and some creative thinking to launch a cloud system unlike what is currently available.
As an aside, I would not trust data security on any solution Meta comes up with, their business model and ethos is user data for sale.
There's a lot of automation, but it's mostly for the team running the service itself. If you're a user of a service, you're likely to be quite reliant on knowing a member of that team to get new allocations or resolve problems. This is fine for developers within a single org, but building out the self-service and multitenancy necessary to make it palatable to outside users would be a major undertaking even if all the oxygen weren't being sucked out of the rest of the company by metaverse stuff.
Lastly, that capacity is already being actively used and they're barely keeping up with building out enough new capacity. Where are the resources going to come from to support outside users?
It just doesn't make sense for them to try, technically or logistically or culturally. The last thing they need is more distractions.
I've never worked at Amazon but my understanding was that AWS was born from bezos deciding one day that the whole company would adopt a microservices architecture and the basic technology for that became AWS. This is quintessential eat your own dogfood. They were first to market and committed a lot to it so they became dominant.
Google (IMHO) only ever really entered the cloud space because they were afraid of AWS becoming so large that Google would cease to be a price-maker in the hardware space and would be a price-taker. I think I read at one point Google was purchasing 1 in 10 hard disks produced globally. That gives you a lot of power to set prices and secure supply if you're smart. Google didn't want to lose that position to Amazon so Cloud became a means of creating demand for more hardware.
Internally, Google services run on what can only be described as a data center OS (ie Borg). You set resource constraints like how much memory and how many CPUs your processes need and there's a scheduler that makes sure you only share resources on a server with other processes within total constraints. There's more to this like flex resource scheduling but it'll do for this example. IIRC Linux's cgroups, which form the basis of this, were contributed to Linux by Google.
Even so, my understand was that Google Cloud was relatively inefficient through all the networking and cloud VM/resource layers. This was years ago so may no longer be true and I have no idea how this compares to Amazon. But Google is prepared to get through that for other reasons. Still, Google doesn't use Cloud internally (unlike Amazon).
Facebook does some things very well, particularly produce deployment. You really can push a diff to master and it just gets auto-deployed to prod 1-4 hours later. What's more you have a bunch of UI frameworks such that developing a fully-functional UI that fits the FB style is amazingly quick.
But on the infra side, everything I saw was a disaster. There were still different specs of servers and each server AFAIK only runs one thing. There's none of the resource constraining stuff Google has. So how would you deploy Cloud to such infra? The answer is you wouldn't because you'd have to rebuild it from scratch and/or create completely different infra to what the rest of the company uses.
FB's approach has always beeen to build things quickly and not worry about efficiency. That's not the recipe for success in a low-margin business like cloud.
https://www.theinformation.com/articles/why-openai-spent-bar...
"When Microsoft in 2019 said it had invested $1 billion in OpenAI, the high-profile artificial intelligence startup agreed to develop its products exclusively using Microsoft’s Azure cloud servers."
But since Facebook has been trying to diversify I too am curious why they haven't tried their own cloud products yet.
Not saying they can’t do it if they really committed, but the premise of the question that “they already have technical expertise” is flawed.
Btw, this is the similar learning curve that Google went through (and arguably still going through) when they decided that they could expose their awesome infra to external customers.
However I do think that from a business standpoint it would feel a bit out of place, or even desperate, for Facebook to get into that space. A more logical approach might be for them to do something more like acquire Wix and allow businesses to have pages that were more dynamic than just a Facebook page. Allow for more direct e-commerce, etc. (and take a cut). If that was successful, they could grow into more datacenter type services as a logical extension.
Realistically, Meta doesn't exactly have a great track record with user data and privacy. So I doubt a lot of companies would be willing to join.
OVH was created in 1999 and I would say is less well known than AWS probs meta have a similar client to IBM Cloud or OVH.
If the goal was to make money, Meta may be interested in cloud computing. Some of the biggest open source projects powering the internet are out of Meta. They have some understanding of developers and know how to interact with them and what they want / need. But alas, cloud computing isn't making the world a more connected place or whatever Zuckerberg thinks he's doing, and until he steps down, the focus required for this type of shift would be difficult to pull off.
No, I do not think Meta could compete with AWS.