Last week saw the big three hyperscale cloud companies (Amazon, Microsoft, Google, or, as I refer to them, AMG) report their numbers as part of their parent company announcements. And those numbers were fantastic.
Turning to the industry flagship AWS first, it fell just short of a $10 billion quarter ($9.9 billion, to be exact). The key number — and the one I regard as the single most important number in the technology industry — is its growth rate, which continues at a torrid 34%.
This is, to be fair, down a bit, but this small drop must be taken in perspective. AWS is now a $40 billion plus technology provider, and growth at its rate at its size is unprecedented in the technology industry, or any industry ever, for that matter.
Most of us are so accustomed to AWS quarterly results that we fail to really consider just how impressive they are — but consider them for a moment. Forty billion dollars of revenue. A 34% growth rate. That means next year AWS is likely to achieve on the order of $52 billion of revenue. That is incredible. And it’s all organic. Most large technology vendors have grown by acquisition. Not AWS. It’s all based on services the company itself rolled out.
The MG part of the big three did pretty well, too. As in very, very well. Google broke its cloud numbers out for the first time and announced that it had achieved $2.61 billion in revenue. And Microsoft announced its cloud numbers as $12.5 billion.
It’s difficult to directly compare them to AWS as both bundle other services (Office 365 in the case of Microsoft and Gsuite for Google). I have taken a stab at estimating their numbers in the chart below, which shows Azure at $4 billion in revenues, and GCP at $1 billion. I’ve tried to be conservative, so these numbers might be as much as a billion dollars too low.
Turning to their growth rate, Microsoft said Azure grew 62%. Google did not announce a growth rate, but one can assume it’s probably something like 50%.
In summary, AMG are at something like $57 billion in revenue per year, growing at a blended rate of perhaps 50%.
Clearly there is a massive trend going on in which application deployments are increasingly being directed toward AMG. And these aren’t just new (aka greenfield) cloud-native applications; there’s no way the infrastructure spend for these applications adds up to $57 billion. It’s clear that AMG are the beneficiary of enterprises shifting existing (aka brownfield) applications from their original deployment environments into hyperscale environments.
“It was the best of times, it was the worst of times.”
While AMG revenues are skyrocketing, it’s a different, less positive story for the erstwhile technology industry leaders. Unlike AMG, for companies like IBM and HPE, flat is the new up, and flat is not always achievable.
The chart below compares AMG revenue growth with three former enterprise stalwarts: IBM, HPE, and Cisco. One can see that for these companies, growth has essentially disappeared. Today they grimly attempt to keep their heads above water; one can only fear for what the future holds for them.
It’s not just money, either. It’s very clear that technology innovation has decisively shifted to AMG and companies that live in their ecosystem. And that’s deadly for the incumbents because it foretells an accelerating phenomenon.
The IT industry is notoriously fast-paced with improvements tumbling out of the vendor community and users attuned to seeking out competitive advantage by adopting emerging technology. When innovation shifts to one set of vendors, users focus on them to set future direction, consigning the previous industry victors to the sidelines. It’s not out of the question for flat revenues to turn steeply down, with unpleasant consequences.
AMG are on a growth tear. The entire IT industry looks to them and their ecosystem partners for direction and innovation. Even on a $57 billion dollar run rate, they’re still experiencing blistering growth, with no signs of a slowdown.
So the obvious question is how much room do they have to roam? How big can they grow?
From discussions with AWS, they believe they can grow revenues to the multi-hundred billion dollar range. Andy Jassy once said Jeff Bezos believes that AWS could come to be larger than Amazon’s ecommerce business (Amazon’s ecommerce business was on the order of $70 billion for the quarter, so that gives a measure of how much opportunity he sees for AWS).
Another way of looking at the question of how big is to estimate how much more application deployment opportunity exists. In terms of what percentage of total application portfolios that have moved to AMG, I’ve seen estimates ranging from 4% to 20%. Frankly, 20% seems high, while 4% seems awfully low. I’d put the percentage as high single digits, perhaps 7% or 8%. That would imply a total addressable market of $600 billion or more. Of course, not every workload is going to move to AMG, but it’s obvious there’s lots of blue sky above them.
And that’s for existing workloads. But it’s likely that AMG can actually increase the total size of application portfolios, which is to say, they can grow the size of the pie. The reason for this is something called Jevons Paradox, referring to a study a 19th century British economist conducted on coal consumption. As coal mining became more efficient due to the application of steam technology, coal prices dropped. Jevons expected people would take the money saved by using cheaper coal and purchase other goods; paradoxically, they chose to consume more coal. In fact, for coal total spend increased, as the larger volume consumed outweighed the lower cost of an individual unit of coal.
Given how much less expensive cloud computing is than on-prem options (and, yes, I’m aware of those who proclaim that on-prem is cheaper; they’re wrong), one would expect that enterprises would ‘consume’ more applications, applying information technology to problems previously uneconomic to address.
Jevons Paradox does not address a second factor that is likely to grow the total size of industry application portfolios even more: friction. AMG offer computing resources in mere minutes in contrast to the weeks or months typical of on-prem environments. That ease of access reduces the friction of consumption and, in effect, reduces the labor cost of obtaining resources.
Nearly every cloud-using organization has experienced growth in total computing resource use as developers and operators are able to access resources quickly and easily. In my experience, reduced friction is a powerful engine of cloud use, and one unaccounted for by Jevons Paradox.
The combination of Jevons Paradox and reduced friction means that the total addressable market for AMG may be significantly larger than the revenue stream associated with existing application portfolios; the increased addressable market made possible by these two factors might be as much as two or three times as large.
So the answer to the question, how high is the sky, is way high. Today, everyone is convinced that cloud computing is a viable deployment option. In the future, we’ll understand just how large the cloud market can be.