TL;DR: IBM’s “cloud” should always have quotation marks around it.
Once again, I have been triggered by IBM claims to be a leader in cloud computing.
IBM says it is jettisoning its single biggest business – Managed Infrastructure Services – to “accelerate its hybrid cloud growth strategy”. Managed Infrastructure Services are the core of the IT outsourcing business IBM has built up over the last 25 years (and until today, one IBM proudly if misleadingly called “Infrastructure and Cloud Services”). This divestiture is more financial engineering from the incredible shrinking IBM, but dressed up as a bold cloud computing move (even if they’re just emulating a maneuver HPE made years ago). But what particularly chafes is the implication they have a material cloud business to double down on.
IBM trumpets revenue and growth numbers for its “cloud” business every quarter. Here are IBM’s Q2 2020 results, including the portion of each of their four business segments they attribute as “cloud” revenue:
IBM tells us they did $6.3 billion in “cloud” revenue in Q2. When we look at the company that puts them in, it is a pretty impressive number. In the IaaS space, AWS did $10.8 billion in Q2, while Azure and GCP did over $6 billion and $2 billion, respectively. In the world of SaaS, Salesforce did $5.15 billion in their Q2. Microsoft’s Commercial Cloud business all-up (IaaS + PaaS + SaaS) was $14.3 billion.
Staggeringly, $6.3 billion of “cloud” is 35% of IBM’s total revenue. If this dinosaur is getting over a third of its revenue from the cloud, why are we not hailing this company as a turnaround for the ages and inducting current management into the business hall of fame? Perhaps we are wrong to think of IBM as a legacy mainframe company surrounded by low margin consulting that has missed every technology wave of the 21st century?
The question is whether IBM’s “cloud” is actually cloud in the sense of what the rest of the industry calls cloud, or a manic, multi-billion dollar case of cloud-washing? Most would consider cloud computing to be the delivery of computing services over the Internet, whether IaaS, PaaS, or SaaS. Their (unaudited) “cloud” numbers imply IBM must have some very big cloud services. Lets go through their segments and try to figure out where IBM keeps its “cloud”.
Cloud and Cognitive Software – this segment includes Cloud and Data Platforms ($2.8 billion; 30% growth in Q2 but that reflects the addition of Red Hat as opposed to sustained double-digit growth), Cognitive Apps ($1.2 billion; -8% growth, alas Watson), and Transaction Processing ($1.7 billion; -14% growth). IBM claims $1.6 billion of this — over a quarter of their entire software business — is “cloud”. There are high margin software assets in this segment, but next to none of this is cloud services. They probably put OpenShift PaaS services here but they’re infinitesimal (thought I’m sure IBM would tell us the fact they are cutting prices is a sign of immense customer demand and that the open hybrid cloud word salad diet is all the rage).
Global Business Services (GBS) – this segment includes Consulting ($1.9 billion; -3% growth), Global Process Services ($0.2 billion; -12% growth), and Application Management ($1.7 billion; -8% growth). Maybe there is some SaaS in Global Process Services but even if that is entirely SaaS, we still need to find $1.2 billion in cloud revenue across consulting or applications management to get to the claimed $1.4 billion. Dubious…
Global Technology Services (GTS) – this segment has two businesses. The larger was recently called Infrastructure and Cloud Services ($4.8 billion, -5% growth) but now has been renamed Managed Infrastructure Services as it gets spun out. NewCo is emphatically the non-cloud, legacy business. The other business is Technology Support Services ($1.5 billion, -6% growth), which IBM seems to be keeping. This segment supposedly was $2.4 billion in “cloud” revenue in Q2 but NewCo appears to have no cloud revenue. So whatever else is going on, IBM’s “cloud” business should shrink by up to 38% when they split up the company (and I’ll wager they’ll have moved onto a new financial reporting shell game by the time the divestiture closes).
Systems – this is the hardware ($1.5 billion; 13% growth on the back of 68% mainframe growth) and operating system business ($0.4 billion; -13% growth). A remarkable $0.8 billion, 42% of the total hardware business, was classified as cloud revenue. Not only is this is the hardest (yuk yuk) to believe, it is also the segment with the most of its revenue claimed as “cloud”. Is there a mainframe cloud somewhere we missed?
IBM has been caught before playing games with what they count as “cloud” and what isn’t. The best definition seems to be “things that are growing this quarter” are “cloud” (or sometimes “strategic imperatives”), as they desperately try to highlight whatever green shoots they can find across the desert wasteland of their withering business.
It is, ah, “fascinating” that over 60% of IBM’s claimed “cloud” business ($3.8 billion) comes from their consulting segments, not the product segments. What kind of a cloud company doesn’t think of cloud services as products? (perhaps we’ve found the root cause for their malaise?) Add in the hardware segment, and we get to almost three-quarters of IBM’s “cloud” business is in the last places we’d expect to find real cloud businesses.
But the biggest question is where is IBM Cloud? This is their public cloud offering that ostensibly competes with the likes of AWS, Azure and GCP, and which they go to great lengths to blur with all the other “non-cloud” things they sell to achieve some semblance of a cloudy patina. I have always assumed it was under GTS, specifically in that “Infrastructure and Cloud Services” business. The brutal answer is IBM Cloud is so small it doesn’t move the dial for any of these numbers. I’d quote Gartner’s estimate of IBM Cloud’s revenue, but they downgraded IBM in 2019 into the Other bucket and don’t estimate its standalone revenue any more.
But, you may say, Q2 was the first full quarter of the pandemic and perhaps those numbers don’t accurately reflect IBM’s longer term situation? I think IBM’s dismal longer term trend is pretty clear (to link to that post for a third time), but nothing shows that IBM isn’t a cloud company more than the performance of their stock during the pandemic. While the real cloud companies saw their stocks soar, IBM underperformed the market overall and badly underperformed technology stocks. That simple fact, plus probably disconcerting forward visibility, likely precipitated the belated decision to cast out IBM’s single biggest business as they finally realize the ship is sinking. And, as part of that announcement, IBM preannounced their Q3 earnings and confirmed yet another revenue decline.
It is hard not to conclude that IBM’s “cloud” numbers are somewhere between optimistically delusional, intentionally deceptive, and outright hallucinogenic. When IBM talks about “cloud”, they’re just not talking about what the rest of the industry is talking about. But it will be fun to see them double down on their cloud-washing business. I’d ask for more disclosure to prove me wrong, but I’m sure that will only accelerate the day when they through all the cards up into the air again and provide yet another set of reporting segments. Those new segments no doubt will have cloud in the name. I just request they put “cloud” in quotes.
And to keep this short(er), I won’t even get into the question of how do you create a successful hybrid cloud business when you’re nowhere in the public cloud business. I explore that in some detail here.