Cloud computing is already extremely profitable, and well above their 10-figure line for a sustainable project.

AppEngine has been around for more than a decade. GCP might not catch up to AWS but it's not going away.

Snapchat was (is?) built on AppEngine and they signed a 5 year 3 billion dollar contract with GCP. I wouldn't be surprised if AppEngine is hanging by a thread outside of a few large users as GCP devs have on multiple occasions signaled new projects with the intention of replacing AppEngine (I forget which ones, feel free to verify that anecdote yourself).

AppEngine is just one product in Google Cloud, and the first they released in 2008. Of course there will be new versions and eventually new products that are recommended like VMs, GKE, Cloud Functions, Cloud Run, and more.

It's still supported and has plenty of usage so I don't see it as "hanging by a thread" unless you have more data than things that people said but apparently you forgot?

So what? I can give you a point-by-point rebuttal of everything on that page.

> The standalone legacy SDK and appcfg tooling

Just use the newer `gcloud` command-line tool.

> Go 1.9 (GA) and Node.js 8 (GA)

Upgrade your language man. This is PaaS. You're expected to keep your app up to date.

> Admin API v1beta4/v1beta5

These are beta versions of the API. Why don't use migrate to the GA versions?

Etc, etc. In fact the large number of deprecations on that page only shows how rapidly new features and runtimes and frameworks are being developed to replace old ones.

While AWS has deprecated runtimes, if you have an old running version of Node for instance on lambda, it will run forever. It will force you to upgrade to a newer version the next time you update the lambda.

As far as I know, AWS has never discontinued a service or turned off an API. They may deprecate a feature or make it unavailable for new accounts or regions.

Heck AWS still supports running an EC2 instance outside of a VPC if your account is old enough,SimpleDB , and using S3 as a BitTorrent seed for old regions.

if it was extremely profitable it would be a separate line in their financial statements (rather than being aggregated with other things)

it isn't, so draw from that what you will

Google earnings report shows Google Cloud and Youtube revenue for the first time: https://techcrunch.com/2020/02/03/alphabet-earnings-show-goo...

And it is still estimated that Youtube is barely profitable between what it pays creators and bandwidth costs.

Barely profitable (assuming that's all it is for the number 1 video destination online) is still profitable.

Also Youtube != Google Cloud.

Yes I realize that but having revenue without profit is kind of meaningless.

Optimizing for growth requires forgoing profit, which is the model in place for GCP (and what everyone seems concerned over) so I don’t see what the problem is.

Is it really that you just don’t feel they’re committed enough for your standards?

AWS and Azure seem to both have growth and profit...

... which includes gsuite, which will be most of it

Is it profitable for Google?

Yes.

Your conclusion is based on?

Like I said, it's double-digit billions product line and has a 10+ year history. And internal metrics show lots of profit margin comparable to AWS, although you'll have to find someone to share that with you in person.

Thanks just bought 100k

So either you have private information (that hopefully you wouldn’t be sharing on HN) or you can provide a public link?

Not just that. Google simply doesn't have it in their DNA to deliver the kind of high-touch support required of remaining #1 in this business.

They've been hiring the pushy sales types to deliver one end of it but not the delivery people required on the other.

Not to mention that Google is so full of “smart people” (tm) that their sales force is made up of people who probably would never stoop low enough to do lift and shifts as phase 1. They would probably want you to move everything to k8s and rewrite your entire stack.

As opposed to MS. If a business has a legacy app that requires a 10 year old version of Sql Server running on a 12 year old version of Windows, they will put that in Azure and probably offer extended support for the OS for free.

Worse...they did this to us not too long ago:

They push you to use a Beta feature as it's the "preferred" way of interacting with their internal APIs. Meanwhile they know definitively that there's a race condition in their code when using this feature that will _take down the entire service_. It took down our entire Production project, they weren't forthcoming with any information to get us a fix for several hours (until we started sending nasty escalation emails about our P1 ticket not being handled) and then they blamed us for using a feature that was still in Beta.

We'd had literal face to face meetings where they'd told us we need to use this feature.

Interestingly enough, we were assigned new reps this week. I wonder if our AM and TAM were laid off.

They've recently brought in a bunch of Oracle people to run GCP so, I'm pretty sure anything is on the table at this point.

They are also hiring lots of MS and AWS sales AMs, SAs, and ProServe people

Bringing in people from Oracle as your sales guys. With the great reputation that Oracle has in most organizations what could possibly go wrong?

Worked at a company that signed a contract with Dyn. Oracle asked to try to sell us additional services so we said sure why not.

Oracle flew 23 people across the entire country, all in expensive suits, to sit in a conference room for a full 10 hour pitch meeting where only 2 people talked.

We didn't buy anything.

Rip Dyn. :(