3 ms·

> Profitability is inevitable. Unless the rise of many viable competitors drives cloud computing down to a commodity service with razor thin margins. There's

> Profitability is inevitable. Unless the rise of many viable competitors drives cloud computing down to a commodity service with razor thin margins. There's no guarantee that the AWS gravy chain is going to materialize for other companies.

paxys 4y ago
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.
hotpotamus 4y ago
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.
crazygringo 4y ago
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.
- peanuty1 4y ago
  There are quite a few players though. In addition to AWS and Azure, there's also Oracle and IBM.
  - crazygringo 4y ago
    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. They're not clouds where Netflix or Spotify are going to spin up a few thousand servers. Unless I've been misinformed? So they don't really seem to be in the same category. I mean is anyone not already in the IBM ecosystem ever choosing between AWS and IBM?
    - teepo 4y ago
      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.
    - IceWreck 4y ago
      Zoom runs on Oracle Cloud. I guess Red Hat's managed stuff now runs on IBM cloud but they bought their way into that.
    - mr_toad 4y ago
      > 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.
  - qchris 4y ago
    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.

paxys 4y ago

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.

hotpotamus 4y ago

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.

crazygringo 4y ago

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.

peanuty1 4y ago
There are quite a few players though. In addition to AWS and Azure, there's also Oracle and IBM.
- crazygringo 4y ago
  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. They're not clouds where Netflix or Spotify are going to spin up a few thousand servers. Unless I've been misinformed? So they don't really seem to be in the same category. I mean is anyone not already in the IBM ecosystem ever choosing between AWS and IBM?
  - teepo 4y ago
    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.
  - IceWreck 4y ago
    Zoom runs on Oracle Cloud. I guess Red Hat's managed stuff now runs on IBM cloud but they bought their way into that.
  - mr_toad 4y ago
    > 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.
- qchris 4y ago
  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.