Aah, earnings day. That quarterly event where we (and by we
, I mean people who spend way too much time thinking about IT infrastructure) get to see just how fast the biggest cloud providers are growing. The answer, of course, continue to be very fast,
especially at Amazon Web Services
If you click on either of those links, they’ll take you to stories in GeekWire (in fact, there are a lot of stories from GW today) that includes graphs charting growth over the past few years. But the headline is that AWS grew 41 percent year over year in the third quarter, with revenue hitting $4.6 billion and profit topping $1.2 billion. Growth at AWS is remarkably constant, and revenue has quadrupled since the same time 3 years ago.
Things at Microsoft are a little trickier. While its Intelligent Cloud division topped $6.9 billion during the company’s fiscal first quarter, both revenue and profit were down from last quarter. That being said, Microsoft does claim that Azure revenue grew 90 percent year-over-year, so that’s definitely something. If my math is correct, that would mean Windows Server and other legacy software products lumped into “Intelligent Cloud” are falling or flat-lining, while Azure is moving into the driver’s seat.
Google parent Alphabet also announced earnings today
, but it doesn’t break out cloud revenue at all. The “Google other revenues” business line into which Google Cloud is lumped grew by just under a billion dollars since last year, to $3.4 billion. Experts suspect cloud computing comprises a large percentage of that number, although Google’s cloud business includes G Suite as well as servers, databases and the like.
So, there’s the recap.
However, I think the more interesting game is predicting where future revenue will come from, especially in areas that don’t mentioned every day. So, yes, serverless, cloud-native and AI workloads will certainly be money-makers in the coming years. But so will, I predict, a lot of workloads that previously spent their lives inside banks, national labs and even some three-letter agencies.
For example, earlier this week Microsoft and Cray announced a partnership
wherein supercomputer users will be able to host Cray systems inside Azure data centers. As I noted back in August when Microsoft acquired Cycle Computing
, high-performance computing was an industry worth somewhere between $20 billion and $36 billion in 2016. Analysts estimate that cloud-based HPC spending was nearly $800 million and was also the fast-growing segment. Cray systems presently hold 3 of the top 10 spots on the Top500 list
of the world’s fastest supercomputers, and 57 systems (or 11.4 percent) overall.
Those workloads are not going away, but they are going to the cloud. So, there are potentially billions of dollars at stake if Microsoft can capitalize on its deals with Cycle and Cray. IBM, which built 27 systems (or 5.4 percent) of the systems on the Top500 list, also has an opportunity here should it choose to step up its business beyond the limited approach it seems to be taking presently.
And speaking of Microsoft and IBM – and Google – there’s also the impending era of commercially available quantum computing, which will almost certainly have to be delivered as a cloud service if it’s going to have any sort of broad adoption. It’s still not clear which applications will excel on quantum systems, but it seems likely that workloads in scientific fields like chemistry and pharmaceuticals, currently chugging along on supercomputers, will be among them.
I’m not predicting supercomputer or quantum computer workloads will be the primary driver for any of the major cloud providers, but I do think there’s an underappreciated amount of money to be made on them. And if the name of the game in cloud is earning the most money and being the biggest, a few billion extra dollars is definitely something to shoot for.