The 2025 enterprise computing forecast – cloud wins hands-down


The sales of cloud computing (IaaS and PaaS) is growing at such a rate (see the Figure which shows sales v growth in the year to the end of June) that it is going to overtake both ‘traditional’ outsourcing/managed service and data center equipment m markets – but when? I’ve been analysing this subject in depth since 2008 and cloud computing has been the subject of 340 of the 760 posts on this site. You’ll want to read more about how much of the enterprise computing market cloud will take over and the consequences.

Table – IaaS/PaaS market shares ($USb) – year to end of June 2016 and 2017

<Q2 2016 Share % <Q2 2017 Share % Growth %
Amazon $9.9 19.5% $14.5 22.0% 46.1%
IBM $10.0 19.6% $14.4 21.8% 43.7%
NTT $2.5 4.9% $3.3 5.1% 32.8%
Microsoft $2.4 4.6% $2.7 4.0% 12.4%
Google $2.1 4.1% $2.6 4.0% 23.6%
Fujitsu $2.2 4.4% $2.4 3.6% 8.4%
AT&T $1.8 3.5% $2.2 3.3% 20.7%
Dell* $1.3 2.5% $2.1 3.2% 65.3%
Rackspace $1.5 2.8% $1.8 2.7% 20.7%
Oracle $1.1 2.2% $1.5 2.3% 33.8%
Other $16.2 31.7% $18.5 28.0% 14.3%
Total $51.1 100.0% $66.0 100.0% 29.3%

Notes: Dell* includes Virtustream
Source: ITCandor, 2017

The Table gives the annual market shares of the top 7 cloud services providers worldwide for the year to the end of June in 2016 and 2017. It demonstrates the strong leadership positions taken by Amazon with its AWS and IBM with its SoftLayer and the absence or relatively poor showing of Cisco, HP and Dell who are the current leaders of the enterprise systems market. More astonishing is the current growth rate, which is over 10 times that of the traditional equipment and IT services areas.

The Figure shows the annual evolution of spending on the 3 generic enterprise computing areas and my forecast to 2025. It shows that:

  • Infrastructure software grew rapidly between 2005 and 2014, but has levelled off since then; I believe it will begin to grow strongly once more; while currently linked more to enterprise hardware than they are to the cloud, formalisation of multi-cloud and hybrid cloud solutions will see more of these tools deployed in the cloud itself, typically bought as SaaS.
  • Sales of traditional enterprise hardware have been sluggish for a number of years; I believe sales will pick up in future as systems suppliers win multiple contracts with local and smaller cloud service providers, although its likely that the number of ODM systems will grow as a proportion of the market.
  • Outsourcing and managed services involve suppliers taking responsibility for part or all of a users computing – they predate cloud services by decades and involve (often inflexible) multi-year contracts. I believe spending here will flow away to aggregators using cloud and applications services as the only way of achieving ‘digital transformation’. Paradoxically the new owners of Dell’s Perot and HP’s EDS businesses will probably become major cloud services companies faster than their ex-parents.
  • IaaS and PaaS cloud services will continue to grow at double digit rates as there areas still many nations and regions of the world catching up on the adoption rates of the USA, UK and Australia and because they will be chosen in preference to building your own data center and acquiring staff with the relevant technical skills.

I can’t see any major issues which would disrupt this picture dramatically; in fact continued poor world economic performance and the business disruption of new nationalism at a country level are likely to accelerating the adoption of cloud services.

In the Figure I show the forecast in more detail by separating out servers, networking and storage systems from enterprise hardware, managed services and outsourcing, as well as IaaS and PaaS.
Cloud computing (akin to hyper-converged infrastructure solutions) allows an organisation that lacks (or doesn’t want to acquire) traditional technical skills and/or the budget to build a new data center to build modern, scalable applications by relying on the output of a supplier’s resources. It’s not just the technical solution but also the way you pay for cloud services which is increasingly attractive, allowing users to opt out of the tradition of paying up front for hardware and moving to more of a ‘pay-as-you-grow’ model.
Skills shortages are also driving this style of centralisation, resulting in a smaller number of technical experts being responsible for much larger numbers of servers, routers and storage arrays. Every application in the world has to be running on hardware systems somewhere, which is why I laughed the first time I heard of ‘serverless’ applications; however – and unfortunately – for the traditional systems vendors, the largest cloud service providers are Original Design Manufacturers (ODMs) with technical knowledge to build their own servers, storage, networking and other data center components and enough influence to get great – sometimes even zero – prices for components directly  from suppliers such as Intel, Broadcom, Western Digital, Seagate and Micron.

My advice to previously-reluctant systems suppliers is to launch your own cloud services as quickly as you can, or you’ll be disrupted as the canal builders were by the coming of the railways in the nineteenth century.
Please contact me if you’d like to sponsor this research.

Leave your details below if you’re interested in purchasing ITCandor’s Cloud Computing subscription service:

[email-download download_id=”12581″ contact_form_id=”12535″]