IoT Analytics does a pretty good job of looking at the growth potential for cloud computing hyperscalers (public cloud providers) moving forward. They see this market moving from $157 billion in 2022 to approximately $597 billion in five years, to anywhere between a ~$0.6 trillion to ~$10 trillion total addressable market (TAM). The wide range accommodates an aggressive or conservative growth pattern or something in the middle (read the article for more details). IoT Analytics is making these predictions on a 10- to 20-year horizon.
These figures represent the market for hyperscalers only. They don’t include people services or many add-ons needed to make cloud solutions truly work. If you add all that in, I suspect hyperscalers will grow roughly at the same rates they are now.
What is most interesting to me is how far into the future this report looked. I rarely take those leaps, understanding full well that technology tends to move in many random directions and at many random speeds based on the fickle nature of how we consume technology and the changing priorities around new and emerging concepts.
What is true about the cloud computing market, and especially hyperscalers, is that cloud computing is the base of emerging technology moving forward. For instance, if artificial intelligence grows more, so does the platform where it runs, which is the hyperscalers. Same for cloud-native development, devops, and anything else you can think of. They run on hyperscalers. Thus, even as interest in new technologies changes, hyperscalers will remain table stakes and will grow no matter what.
Of course, the pragmatist in me understands that eventually markets run out of runway for growth. Look at legacy technology, for example. (Note: I’m not calling out a specific platform so I won’t get hate mail from those who have built a career around that platform.) Although legacy technologies are still around and are an important part of most enterprises’ computing infrastructure, they are no longer high on the list of ways to modernize enterprise IT.
What keeps the hyperscaler space growing consistently is the barrier to entry. You’d need billions in capital to build a competitive public cloud provider that can scale and has the points of presence needed. Also, it would take so many years to become competitive, you would likely miss the market. This is why so many larger players dropped out of the public cloud space years ago to focus on other areas where they had a better chance of success. A few strong players can grow in a relatively simple market and focus marketing and development on growth and better protecting the market.
Granted, we do see a rise in alternative public cloud providers that go after niche services, such as industry-specific cloud services, cheap cloud storage, or clouds that are specific to countries and cultures. They won’t make up a lion’s share of the market anytime soon, but they will influence things moving forward. So, we may be moving from many hyperscalers 12 years ago, to just a few today, and back to many in a few years. Indeed, we’ll soon define these various types of hyperscalers using their category names.
However, the largest force that will drive growth is dependency. As I mentioned, most new and innovative technologies run on hyperscalers. We’re seeing strong growth in AI, serverless, containers, blockchain, and even edge computing, all depending on a hyperscaler as their hosting platform (you have to be in the edge of something). Most technology that we invent or reinvent in the future will continue to depend on public cloud providers and will drive further growth.
So, no matter where the market goes, and even if the hyperscalers begin to seem more like legacy technology, the dependencies will remain and growth will continue. The hyperscaler market could become more complex and fragmented, but public clouds are the engines that drive growth and innovation.
Will it stop growing at some point? I think there are two concepts to consider: First, cloud computing as a concept. Second, the utility of the technology itself.
Cloud computing is becoming so ubiquitous, it will likely just become computing. If we use mostly cloud-based consumption models, the term loses meaning and is just baked in. I actually called for this in a book I wrote back in 2009. Others have called for this as well, but it’s yet to happen. When it does, my guess is that the cloud computing concept will stop growing, but the technology will continue to provide value. The death of a buzzword.
The utility, which is the most important part, carries on. Cloud computing, at the end of the day, is a much better way to consume technology services. The idea of always owning our own hardware and software, running our own data centers, was never a good one. Too much capital investment and risk, too much carbon-producing electricity, and no ability to scale and change at the “speed of need” for most businesses.
Considering this, I suspect that the base scenario of $2 trillion TAM, as projected by IoT Analytics, will be easy to achieve.