I first wrote about the importance of cloud computing at the end of 2009. Since then I’ve documented its rise in 844 posts published on this site. Of course its services have become a huge component of enterprise IT. However 2023 was a pivotal year in which total sales declined for the first time – dropping 12% from $465b to $410b. Sales picked up again in Q1 2024, so it’s a great time to reappraise these services as part of the overall IT market.
The overall market is divided into three categories (to mis-quote Julius Caesar, ‘omnes nimbi divisa sunt in partes tres’). The largest of the these is Software as a Service (SaaS), which created new companies such as Salesforce.com, rejuvenated office applications (Microsoft’s Office 365 for instance) and helped eliminate the physical shipment of software packages. The two smaller parts are Infrastructure’ and Platform as a Service (IaaS and PaaS – differentiated by the specifics of the subscription, the former of which is more generic (x amount of storage, y amount of virtual workplaces for example) than the virtual balanced system with associated processor, memory, storage and networking of the latter.
The market has been led and driven by a single supplier, Amazon – whose AWS division had a 17.4% share of the total $437b cloud services market in the year to the end of March 2024 (despite not having any of its own SaaS offerings) and 31.7% of the combined $240b IaaS/PaaS market. Its $94b revenues placed it fourth in the overall IT market as well – not bad for a company that wasn’t an IT supplier of any sort until the beginning of the 2000s. We’ll look into market shares of cloud suppliers below.
Cloud computing has always been attractive to companies and individuals who lacked either the interest or resources (or both) to build their own data centers or systems from which to build and run their own applications.
Cloud computing sales have been disproportionately greater in the Americas than in other regions (see my Figure above, where I show the annual spending in each continent on a rolling 4 quarter basis. In the year to the end of March 2024 the ratio was 51:21:27 for the Americas, Asia Pacific and EMEA respectively. For Asia Pacific and EMEA I’ve shown spending in current $US conversions (the solid lines) and conversions based on Q1 2006 levels (the dotted lines). It’s interesting that the positive gap between spending in EMEA and Asia Pacific is eliminated when we move from the current o constant dollar views. In other words the spending levels in each of these regions was the same in the year to the end of March 2024 in local currencies. As in so many areas of the IT market the uptake of cloud was faster in the USA than outside. It looks now that the market will always be significantly stronger there, as we see in the enterprise network market for instance.
Despite the fact that Microsoft is the leader of the overall software market (6.4%) and was an early and full adopter of both Cloud and SaaS, it was second in the cloud market behind Amazon, who is only a IaaS/PaaS supplier by my definitions (see my Figure above). Salesforce was in fourth position in the overall market despite being a player only in the SaaS market. Google was in third position in the whole market, although it is not a SaaS supplier. IBM is both a major software provider and cloud provider, although its hybrid, multi-cloud approach of the last few years has limited its growth rates in comparison with the major public cloud suppliers – Amazon, Google and Microsoft. Alibaba and Baidu were in sixth and eighth position in the total market due to their large IaaS/PaaS revenues from China, while EMEA’s largest cloud suppliers OVH and France Telecom were only placed in 16th and 20th positions of the world market.
My Figure above shows the leading market shares for IaaS and PaaS separately. Cloud ‘infrastructure’ solutions differ subtlety from ‘platform’ ones of course, with the latter being adopted by those who want to duplicate the specific resources they would have had to invest in when deploying applications in on-premise solutions. I have purposefully left HPE out of my cloud computing research, since its Greenlake approach is to charge in a cloud-like way from equipment pre-installed in the customers’ datacenters. Although Greenlake has been successful, I don’t classify it as cloud any more than I would a server customers acquisition of more power through ‘on/off upgrade on demand’ deals.
Now cloud computing has reached its apogee perhaps, we’re in a better position to look at how its three components are likely to affect the rest of the markets they’re part of. by 2023 SaaS offerings generated 12% of total software sales (see 2023 in my Figure above). I believe it will grow to be almost double that (570b, 23%) by 2033. This is good news for its suppliers, who already have the advantage of servicing the fastest growing of the three basic categories of IT. Incorporating Generative AI into SaaS will allow SaaS offerings to deliver increasingly sophisticated functions over time as well.
The largest parts of the IT services market are implementation, hardware and software maintenance (see my Figure below). Despite their rapid growth over the last decade IaaS (5.5%) and PaaS (3.2%) are the two smallest in terms of spending. I don’t believe they will make further inroads into the overall IT services market in my forecast to 2033 (see my Figure above).
So will cloud computing kill the market for on-premise infrastructure? I believe not. My Figure above shows my forecast for cloud infrastructure (IaaS and PaaS combined) compared with on premise infrastructure, which is a combination of all server, storage system and networking hardware, infrastructure software and half of all PCs and peripherals. Although these cloud elements will continue to grow, there are strong reason to expect a renaissance of on-premise hardware – in particular for large and medium sized companies implementing generative AI applications, since for many the cloud will be considered too insecure.
Cloud computing has reached an important level of maturity in 2024. There are few companies across the globe that haven’t experience its use. Many want better integration into their overall IT estate and better security in coming years.
Information Technology and Communications research
