IBM to split – spinning off its managed service business

IBM intends to spin off its Managed Infrastructure Services business by the end of the year into a new company yet to be named (see my Figure above) and dependent on certification by the usual regulators. Accounting for a third of its current business, this is a drastic, yet logical, move in line with both IBM’s business and long-term market developments (see my Figure below).
The new company will be based on IBM’s current Global Technology Services (GTS) group minus its Technology Support Services (TSS), which will remain in the ongoing IBM company. TSS provides hardware and software support activities including maintenance of other vendors’ kit which it terms Multi-vendor Technology. IBM will shift approximately 90k of its staff into the new venture, which has 4.6k customers currently. The ongoing IBM will contain most of the staff brought in through recent major acquisitions such as Softlayer and Red Hat and will almost certainly have a younger age profile than the split-off one.

In the Figure I show worldwide annual spending on:

• Enterprise hardware (inclusive of servers, storage systems and networks),
• Outsourcing (defined as the acquisition of customer data centers in their entirety along with the staff running them),
• Managed services (similar, but of parts of a customer’s data center operations) and
• IaaS/PaaS cloud services;

Complete with my forecast to 2026. I have not included the consulting services designed to help customers change the way they run or maintain their systems, or to redesign them as part of their journey to the cloud. What’s apparent is the extent to which the spending on older outsourcing, managed services and enterprise hardware areas is slowing, while that on cloud services is growing radiply.
The other main Cloud Services Providers (CSPs) such as AWS, Microsoft Azure, Google Cloud, Alibaba and Baidu make their enterprise revenues almost exclusively from IaaS/PaaS. Very occasionally these companies may build a dedicated data center for a single customer (such as the US Federal government); however even here the servers and storage systems are often proprietary systems designed and built by the CSP. They provide a rich infrastructure accessed via APIs and open source software tools, but they don’t typically acquire customer data centers and the staff of those running them, or provide application software. Microsoft is an exception, although there is a significant difference between Azure and its other businesses.
Of the CSPs IBM has the richest set of other hardware, software and service applications; splitting the company will allow it to separate its newer cloud, AI/ML, blockchain, quantum computing and other innovative offerings from its older outsourcing, managed services and other people-based ones. It is the leading supplier in embracing a hybrid multi-cloud strategy; once the split is completed the new company will support customers pursuing their ‘journey to the cloud’, while the ongoing IBM will provide the software, cloud platforms and innovation for the new things they want to do.

New Figure above shows the proportion of its revenues IBM has reported non cloud areas, showing that it reached about a third of its business in 2020. The ongoing IBM after the split will have a higher proportion of cloud revenues, although the split-off company will also contain some.

On a long-term basis IBM’s revenues have shifted (see my Figure) away from hardware to software, while service has maintained its proportion. Both software and hardware will stay in the ongoing company, while service will be split between those helping organizations to run/modify their existing systems and those providing an end-point new innovative computing.

IBM suggests that the two companies will maintain a positive relationship after the split, which will be necessary; not least because its Financial Services will reach down into many contracts the new company executes. Time will see whether the simplification of its activities leads to greater success. Historically ‘one and one doesn’t equal two’ when you compare the revenues of a merged IT company with the addition the two companies beforehand; in a couple of years we should discover whether ‘one and one equals more than one’ when they split.