Because Google is playing catch up and needs to pay higher salaries to poach from MSFT and AMZN and also pay much more for infra than what Amazon paid for theirs initially.
What does Google still need to catch up on in cloud? Building products/services? Sales? AFAIK, they have comparable cloud services to most AWS/Azure offerings.
Mostly I think it’s just acquiring customers. The big cloud providers go to great lengths to lock their customers in, so convincing them to switch is hard. Requires a much better product which seems pretty much impossible against aws and azure or a much cheaper price.
From someone who does enterprise technical sales for a B2B company: Identity Management is a major area of concern for GCP / GSuite, at least in the enterprise space, especially for the low to moderately technical folks that I sell to. The same could be said for AWS but they have first mover advantage in many ways.
I don't think shutting down products applies to GCP services. Even if it did, it doesn't explain the $858M loss. They are spending big on something to "catch up", I'm just not sure what it is.
They also offer insanely cheap deals to nonprofits and education for both GSuite (or whatever they call it these days) and GCP. Totally makes sense why they do it but I have to imagine its a huge loss on paper.
They also might be building more actual datacenters and doing so in a high end / environmentally friendly way whereas perhaps Amazon has already done a lot of those investments in previous years?
I have heard people claim that Amazon pays more and by that they mean that if you hold your Amazon stock until vesting, it appreciates so much that it ends up being worth more than whatever the Google grant would have been worth. But then you have to ask if the relative performance of the stocks will remain stable over time despite having changed recently, and factor in differences in employee churn and so on.
They pay about 100,000 less. Its widely known, not sure why you are arguing this. Amazon also raised their paybands massively, so a blanket comparison doesnt reflect reality. Look at individual data over the last year or so. Google also downlevels, so comparing levels makes less sense, compare years of experience instead
that $350k base limit is top of the band, not everyone walking inside the door is going to get that. A friend of mind got an offer from AWS about a month ago, his base was offered $180k
From what I hear they paw software engineers less, but find it harder to attract sales personnel, as it’s easier to sell the industry leader (IBM, Oracle vs ???). No sources, just some posts a few months ago.
They don't need to "catch up" on infra. They need to catch up on their product offering.