Some personal thoughts on IBM acquiring Red Hat
Sometimes Monday mornings aren’t the best part of the week, although having three children under 4 years old, it can sometimes be quite nice having an escape route to the office after a busy weekend. Monday 29th October can definitely be filed in the ‘unexpected’ category. It’s definitely not often that you wake up to find your employer is the subject of the largest acquisition of a software company in history.
I figure I’ll note down some of my thoughts and impressions here, so I can look back in 1, 5, and 10 years and laugh at how wrong I was.
It seems that IBM have decided it’s worth paying approximately US$34bn to acquire Red Hat. That’s quite a lot of money by any measure. IBM’s market capitalisation is apparently ‘only’ US$110bn.
Looking at the numbers
34 is just under one third of 110 (those early years of mental maths were not wasted on me!). However, if you look at revenues, profits, headcounts etc., Red Hat is considerably smaller than IBM. Starting with the number of employees – the numbers I’ve seen listed are that Red Hat has approximately 12,000 employees compared to 380,000 listed for IBM. Allowing for a bit of a rounding error, that’s about 30 times the number of employees at IBM.
Red Hat’s latest quarterly revenue was in the region of US$800m – putting it on a rough annual revenue of US$3bn. I can afford to be ‘rough’ and ‘in the region’ here, because IBM’s figures for the most recent quarter are revenue of US$20bn and a recent annual revenue of US$79.1bn. Whichever way you look at it, IBM is about 27-30 times the size of Red Hat. Of course, there are other metrics you could look at which might paint a slightly different picture, however this is very much a much larger company acquiring a smaller one. The interesting numbers I find are the market capitalisations, which unless I’m mistaken would be the cost to buy every share on the market at current prices. Pre-announcement, Red Hat was US$20.53bn (down from a recent high in June of US$30.9bn). Pre-announcement, IBM was US$113.4bn – which then took a bit of a drop to US$104.9bn (hmmm, announce your intention to spend US$34bn on a US$20bn company and watch your share price drop US$8.5bn). You can really disappear down the rabbit-hole looking at numbers (I guess this is why there is such a large banking industry). Why is IBM worth ‘a bit more’ than their annual revenue, whereas Red Hat is worth multiples of annual revenue?
The pessimist in me
Everyone is prone to pessimism. I’ve been through mergers and acquisitions before, notably when I worked for Novell. Novell ‘acquired’ Cambridge Technology Partners (CTP), Ximian, and SuSE. I was still working for Novell (technically I was in the SUSE division though) when they themselves were acquired by Attachmate – who in turn were acquired by MicroFocus. Although I had left by then. Given that MicroFocus have since sold SUSE for US$2.5bn, is it possible they will also be a target now?
Some of the above mergers/acquisitions undoubtedly went better than others. The acquisition of CTP ended up being more of a senior level ‘acqui-hire‘ with the CEO position of Novell changing from Jack Messman to Ron Hovsepian. I honestly don’t remember the software being much of a acquisition. The SilverStream AppServer never set the world alight, and I have a recollection that some part of the CTP portfolio made it into the Novell DirXML (latterly Identity Manager) product. Otherwise it felt like the Novell senior management had run out of ideas and just wanted to pass the reins on to someone (anyone?) else.
The acquisition of Ximian, with hindsight, was a weird one. Novell must have realised that open source software was the future and needed to get on that train. I remember a rumour at the time was that they had looked at Red Hat but been put off by the price, and that SuSE was also mentioned at this time (2002/3 timeframe). The flagship eDirectory product ran on both RHEL and SLES at this time, so it makes sense Novell were looking. In the end, almost as a toe in the water, Novell acquired Ximian. Best known for distributing a highly customised (and polished) GNOME desktop for Red Hat Linux and others. They also had a Linux desktop management product called ‘Red Carpet’, which I guess the senior management at Novell thought would complement the ZEN product they had for managing Windows desktops. I can’t honestly remember the timeline of events – but not too long after Ximian, Novell announced they had acquired SuSE. Clearly, the Ximian employees had arrived first and, again with hindsight, it seems were able to steer the boat. SuSE was eventually rebranded SUSE, there would be a focus on an enterprise desktop product (SLED), and updates to SLES and SLED could be managed using a ZEN branded product that had evolved from Red Carpet that was written in their new ‘mono’ language which itself was a Linux compatible port/copy of Microsoft .Net.
A bit of a car crash
This did not turn out well. There was a shift that SLED would default to GNOME, whereas SUSE was traditionally a KDE based distribution. GNOME in SLED 10 introduced ‘wobbly windows’ and a customised menu, aping the Windows look at the time. I’m not sure exactly who SLED was aimed at – a big emphasis was put on Evolution’s support for Microsoft Exchange and OpenOffice’s support for Microsoft Office documents. The ‘year of Linux on the desktop’ was much talked about and Novell undoubtedly tried to beat the drum as much as possible – but I’m really not sure who was supposed to buy and deploy SLED.
The roll-out of ZENworks Management Daemon (zmd) was an unmitigated disaster as well. Not only was it manage updates for SLED it also managed updates for the newly SLES 10. The dependency on mono brought a lot of resistance from traditional Linux users, and the fact that it then ran incredibly slowly and was prone to memory leaks just compounded the issue. The ZEN brand, that Novell had cultivated for about a decade, was damaged as a result. Meanwhile, Red Hat was focused on the enterprise server market having spun out Fedora as a separate community maintained project (which SUSE would follow suit with openSUSE).
In the background, forcing the hard core open source SuSE employees to use the Siebel CRM system that only seemed to work on Windows and Internet Explorer 6 also caused frictions – the bottled wine version that was produced to run on the Linux desktop users never ran well. Likewise migrating SuSE from OpeneXchange to GroupWise was not popular either, with the slow/buggy Java GroupWise client on Linux.
As hopefully demonstrated above, I’ve seen first-hand how an acquisition can go well (it went well for Ximian) and badly (it did not go well for SuSE, although they seem to have recovered now – it just took them a while). It’s therefore not an impossibility that IBM acquiring Red Hat could go either way.
As per the heading, the pessimist in me would see IBM coming in an ‘managing’ Red Hat as they think it should be run, ignoring the sensibilities and traditions of Red Hat. There are undoubtedly some areas of overlap – what, for example, will happen with JBoss and the Red Hat middleware solutions that compete with IBM’s WebSphere? Red Hat, famously, boasts that it open sources it’s products and that it is the world’s largest open source software company. Will IBM change this attitude? The acquisition has been described (by IBM) as helping them become the #1 hybrid cloud provider in the world. I’m always wary of self-descriptions of the ‘leading provider of …’ or ‘the number one provider of …’. What defines the hybrid cloud here? The Red Hat definition was a dual managed ‘on-premise’ private cloud and public cloud where services and applications could be seamlessly moved between the two and a single point of management, using tools like OpenShift and CloudForms. To be honest, I’m not sure what IBM’s definition is (they have their IBM Cloud Private product). Recently I was at Google Cloud Summit and they define hybrid cloud as (my simplistic summary) GKE on-site and on GCP with GCP as a single management interface.
A possible concern here is that the acquisition represents two waning companies trying desperately to remain relevant in the today’s changed marketplace. IBM have been around for 100+ years and have undoubtedly managed to move with the times. But, have they done that recently? Red Hat still show quarterly growth (something like 63 consecutive quarters), but is the growth running out of steam? Enterprises are surely going to reduce their growth of on-premise solutions requiring RHEL. Is OpenShift compelling enough to off-set that? Has the peak for both of these companies been reached and newer solutions based on AWS, Azure, GCP, and AliCloud going to take the new growth business?
I’d like to hope that IBM management having paid handsomely for Red Hat have the common sense to manage it sensibly and not treat it as a division of IBM that has to be brought into line with existing corporate standards. However, the mantra of ‘prepare for the worst, but hope for the best’ seems sensible here.
The optimist in me
Believe it or not, I can be optimistic at times (some friends might doubt this). IBM have paid handsomely for Red Hat, they also have a strong history of support for Linux, and understanding open source communities. There is no way they will want to knowingly destroy a US$34bn investment. Arguably, they need to make this acquisition work for IBM to remain relevant and strong player in the technology market. There will undoubtedly be synergies (buzzword bingo!) between the respective customers that can generate additional revenue. IBM simply have greater access to potential customers than Red Hat alone would have done, so further growth can be found there. Products that I expect (and hope) that can benefit from this greater exposure would be OpenShift and Ansible.
The words coming from IBM are that Red Hat will be a separate business unit within IBM (Red Hat, an IBM company) and that they will operate as part of the hybrid cloud division (which is showing strong revenue growth). This suggests that IBM see Red Hat as a driver of cloud revenue – so whilst that might make JBoss an oddity, it should mean OpenShift, CloudForms, Ansible, and OpenStack (sigh) are considered significant products (and as a result RHEL).
Leaving Red Hat much as it is but with the extra exposure of being part of IBM brings should be the aim for IBM. Changes will undoubtedly be made – hopefully those changes are small and subtle to start. If they get it wrong, it won’t just be bad news for Red Hat, it will reflect badly on IBM too. There’s a lot at stake here.
Some final thoughts
The deal won’t close until latter part of 2019 – the Red Hat shares are currently trading at US$175 – below the US$190 IBM offer price. To me, that indicates that the market is reasonably confident the deal will close, but not certain. It also suggests there is currently no other potential buyer. It had been rumoured (there are always rumours it seems!) that Microsoft and Google had either been approached or were interested in acquiring Red Hat. Clearly both passed at the price being talked about or just weren’t interested at all. Was IBM the last man drinking in the last chance saloon? Or are they the white knight whisking the princess away from the circling sharks (nothing like mixing metaphors)? Time will tell …