17 comments

[ 3.1 ms ] story [ 51.6 ms ] thread
This is a personal pet peeve: people classifying the build-out of facilities as "CapEx."

You can classify portions of your AWS bill as capital expenditure with a bit of effort, and none of it requires breaking ground on a new facility. It's not entirely clear how much of that type of maneuvering goes into the underlying numbers.

How do you classify portions of your AWS bill as a capital expenditure?

I'm not familiar with this, assuming your context is as a user of AWS?

Yes, that's the right context.

In short form, you have to go with dedicated tenancy on your instances, 3 year RIs, and get your auditors on board with it first.

It's not worth the hassle for most shops until they're looking at a seven to eight figure annual AWS spend.

I know lots of folks doing mid to high 8 figures and I haven't heard of anyone capitalizing the leases like that. But maybe.

Thanks for expounding.

IANACPA, but based on your "In short form" sentence, I guess I see what you mean and how it could be treated as capex. Still, from a layman's point of view, it seems sketchy as hell to me in that the primary goal looks to be making the operating budget look more profitable than it really is if the AWS spend can be amortized over a couple years.
You're not at all wrong. Changing how things reflect on the income statement is a bit of smoke and mirrors; it doesn't change the underlying truth, but does change how they're represented.

It's a weird, weird world.

he's talking about amazon and other cloud providers classifying their money spent on infrastructure as capex, not their customers.
Absolutely-- but internally to these providers, the CapEx model doesn't cover what most people think it does. It's a very, very rough analog for infrastructure spend. GAAP is a fun thing to play with...
It will be interesting to see the followup article where the author will focus on IBM and Oracle which are arguably targeting a different market share to AWS, Microsoft and Google.
Is it just me or the "trillion dollar" figure is completely unrealistic? I mean, even if there are no price wars (and they are inevitable), there isn't any realistic possibility to capture all of IT spend. With price wars in the picture, we'll see a race to the bottom which will erase much of the margin, and make some of the "table stakes" capabilities unprofitable to some of the players, further cutting into the profits.
I'm having trouble reconciling him showing Amazon as the lowest spender by far of the three with their place as largest player by far. Are they just tremendously more efficient? I wouldn't imagine so given they're all essentially renting out a physical resource to their customers. Maybe I misunderstood something about the article?
Google and Microsoft's infrastructure supports search (which is vast) and other workloads beyond cloud infrastructure services.