Oof. Curious to see Microsoft and Apple. I guess if you’re a Google investor you’ll be happy with cloud continuing to grow. Profitability is inevitable.
Cloud computing profit is largely driven by companies that don't make profit themselves.
If we end up with a major tech crunch, which I suspect we will, and a bunch of these big, non-profitable new tech startups/IPOs start to fold, we'll see a massive and rapid contraction in the cloud space.
The last 3 companies I worked at invested very heavily in cloud infrastructure but none of them has made a profit. There's a pretty good chance if you're on HN the same is true for your company.
But we're not seeing that happen, because the initial capital outlay to build a cloud platform is absurdly large.
Google has been at this for a decade and is only now at the cusp of profitability.
With a barrier to entry this high, and vendor lock-in so strong, you can be sure all big 3 cloud services are going to be minting money for many years to come.
I'd also put companies like Digital Ocean into that space as well, plus other VPS providers, including specialty providers like Vast that focus on deep learning hardware, potentially Backblaze for backups, etc. Even if they're not direct competitors, they're eating into services that might otherwise provide margin (or even just a funnel) for the larger cloud providers.
> Does anybody really use those for generic cloud projects though?
My impression was that those were geared towards firms who have already bought heavily into the Oracle/IBM consulting/licensing.
You could say the same about Azure. I see a lot of click-next-collect-cheque being done in Azure, only with a browser instead of a desktop wizard.
WhatsApp, Fitbit, tumblr among many others ran on IBM. Of course none were what I’d consider “generic cloud projects”, they chose IBM bc of specific capabilities not easy to achieve on other hyperscalers. -unless you run your own hyperscaler.
The game seems to be to keep IaaS pricing high and tell people that use it that they're doing it wrong, they need to use our cloud-native offerings which will indeed lower their spend but increase lock-in. I suppose the counter to that is Kubernetes and other cloud-native-like abstractions on top of commodity IaaS so at least you can somewhat easily move between cloud providers and seek better pricing.
From what I can tell it's a cat and mouse game of who can lock-in and who can abstract away, and I assume it's been going on for longer than I've been at this.
Commodities can still be expensive and profitable. Yes it is easier than ever for a company to get into cloud computing, but despite the increased competition retail prices are going up the entire sector is making more money than ever.
The read I have is more around PC sales declining. Cloud still has good margins across the big 5. GCP is a quarter or so away from profitability after the massive build outs and Azure has probably the largest GP right now of the hyperscalers.
