I start more blog posts than I finish, due to no shortage of other distractions and at times a near-crippling inability to blog short. When IBM announced they were acquiring Red Hat in October 2018 for $34 billion, I embarked on a deeply intertwined discourse about the continued decline and fall of IBM, the two companies’ mutual irrelevance in the cloud era, the true essence of hybrid cloud, and why people who invest in technology through the lens of financial statements are doomed. After amassing eleven pages but still incomplete, I put the nascent post aside in favor of parody song lyrics about the acquisition (my worst performing blog post of all time, no less). In an attempt to salvage some of this work now that the deal has closed, here are some abridged thoughts on the acquisition.
I valiantly resisted issuing a hot take on IBM’s proposed acquisition of Red Hat, leaving it to others to unpack and analyze. But I have not seen anything that gets to the core of this deal and critically assesses the hyperbolic claims accompanying it. Hopefully takes, like revenge, are best served cold…
IBM says (over and over again) buying Red Hat “changes everything about the cloud market” and means “IBM will become the world’s #1 hybrid cloud provider”. Those are bold claims from a company that has failed to deliver on every big promise it has made in the 21st century. The IBM-Red Hat deal hinges on the notion of “hybrid cloud”. Unfortunately, that term was already a muddled miasma of marketing even before it got the IBM bear hug (they’re now calling it “next-generation hybrid multicloud” because they love that word salad).
There has been little scrutiny of the mechanism by which this combination vaults them from mutual irrelevance to game-changing dominance in the cloud, what they even mean by hybrid cloud, how that relates to the actual seismic shift of computing to the public cloud, and the applicability of their respective and combined product portfolios. Never mind the crazy price tag.
WHY IS THIS DEAL HAPPENING? (TL;DR MUTUAL DESPERATION)
The bar is very high for pithy descriptions of Hail Mary technology combinations, ranging from “two garbage trucks backing into each other in slow motion” (Scott McNealy on the merger of Compaq and HP) to “two turkeys do not make an eagle” (my former and still unindicted co-conspirator Vic Gundotra on Microsoft buying Nokia’s handset business). So I will not overreach for a metaphor and just characterize it as a hookup between two of the biggest losers of the cloud era.
Simply put, IBM as a platform company has failed to make the transition to cloud computing. The company’s streak of successfully traversing multiple generations of computing is in serious jeopardy. They now face an existential threat that not even their cash cow mainframe can salvage (though as we’ll see they’re of course playing the mainframe card). Some predicted this fate (that post missed top-ticking IBM’s all-time stock high by 16 days). The last couple years of playing “hey, look over there” with tangential initiatives like Watson (curing cancer in retrospect might have been a slight overpromise) and blockchain (solving the great mashed potato provenance crisis!) only underscores IBM’s irrelevance on the main stage of cloud computing (and because CAPEX doesn’t lie). Despite the then much-hyped acquisition of SoftLayer in 2013, IBM can no longer even see the public cloud leaders with the help of an orbital telescope and a star map. Meanwhile the cloud continues to relentlessly devour IBM’s business. After shrinking by almost $28 billion in revenue over the last six years, and the current CEO having underperformed the S&P500 by over 15,000 basis points at the time the acquisition was announced, IBM needed to do “something”.
Red Hat has its own challenges (and at the acquisition price, has found a wonderful resolution that leaves IBM and its shareholders holding that bag). Red Hat may look like a gem to IBM (anything that isn’t shrinking would), but they too have a cloud relevance problem. The fact Red Hat is the poster child for commercial open source is an orthogonal irrelevance. Red Hat faces a very traditional technology industry problem: generational obsolescence. The bulk of their revenues come from “infrastructure-related offerings”, namely the Red Hat Enterprise Linux (RHEL) server operating system. As computing shifts from customer data centers to the public cloud, RHEL is not moving along with it. You may have heard that the cloud runs on Linux. It does, it just doesn’t run on RHEL. AWS, Azure and Google don’t pay Red Hat for Linux (they do let customers run RHEL as a guest operating system if desired, but the case for paying grows ever more tenuous – if the hyper-scale clouds don’t need it, why do you?). This shrinking TAM finally started to bite in 2018 as Red Hat’s core growth slowed and they missed Wall Street estimates for two straight quarters, which is considered problematic for a growth stock, as evidenced by a third of their valuation disappearing. Those misses combined with their visibility going forward are likely the catalyst for Red Hat deciding it was finally time to give IBM a call.
WHAT IS HYBRID CLOUD? (TL;DR SOFTWARE)
So what is this hybrid cloud that changes “everything about the cloud market”? Don’t expect much enlightenment from IBM’s announcement. The release mentions cloud at least 19 times, hybrid cloud 12 times, and multi-cloud 7 times, but doesn’t trifle with definitions. If one were cynical about IBM, one might conclude they were trying to conflate some vague notion of hybrid cloud with the cloud cloud, where they have performed so abysmally.
An industry definition of the admittedly fuzzy concept of hybrid cloud would entail integrating on-premises infrastructure (sometimes and often inaccurately referred to as private cloud) with one or more public clouds so applications and/or data can span a common pool of resources. To manage, orchestrate, and automate this pool as a single entity, you need a common software model that spans both customer data center(s) and the public cloud(s). As is usually the case, software is the critical element (a typically insurmountable challenge for metal-bending legacy hardware companies).
There are two general approaches to a software architecture that provides a management, operations and development model for hybrid cloud. The public cloud providers not surprisingly are approaching it cloud-first and treat hybrid cloud as an on-ramp to the public cloud. They are extending the rich and modern architecture of the cloud back into the legacy customer environment. In the absence of an even modestly competitive cloud offering, IBM and Red Hat are both starting from legacy software running on-premises, and somehow intend to architecturally span and control the vast and ever-expanding array of software that powers the hyper-scale public clouds (a task for which IBM might consider adopting the motto “Benediximus Cum Eo!”, which Google Translate says is Latin for “good luck with that!”).
AzureStack is the most advanced example of hybrid cloud from the public cloud vendors. Because of its vast presence in customer data centers, Microsoft started down the hybrid path early and did it initially with a truly hybrid architecture that bridged on-premises and public cloud but was neither fish (its existing Windows Server architecture) nor fowl (Azure). Recognizing the futility of this approach, they hit refresh (yuck yuck) and embarked upon a cloud-first approach that extended Azure back into the customer data center.
Amazon Web Services, after years of dismissing any form of on-premises computing as “whacky”, “ill-advised”, and “archaic”, has recently gotten hybrid cloud religion. As is often the case with new converts, they’re especially enthusiastic, to the point where they plan to offer not just one hybrid cloud software architecture but two. Their recently announced Outposts product carries the AWS architecture into the customer datacenter. But Outposts hardware can also run a second software stack from partner VMware, which extends vSphere from the customer premises out to a special bare-metal neighborhood (ghetto?) of the AWS cloud running VMware software. AWS is also putting at least some of its higher-level public cloud services (RDS on VMware) on top of that VMware stack as well, providing a second way to get at least some AWS services inside the customer data center. How you decide which stack to run presumably depends primarily to which company’s salesforce you are speaking.
Google also is starting to bring their software on-premises with GKE On-Prem (which I suspect was named by a Microsoft alum…), subsequently rebranded Anthos (not sure who thought naming it after a Greek tragedy was an improvement). Google is the least mature commercially of the big public clouds and have some huge hurdles to overcome in supporting enterprise customers in the cloud, never mind deploying and supporting software running in customer data centers. Google’s hybrid model is all about allowing containers to run on-premises or in the cloud, which we’ll revisit later, as opposed to bringing their broader set of higher-level cloud services (which are the most compelling thing about Google Cloud) into customer data centers.
WHAT DOES IBM THINK HYBRID CLOUD IS? (TL;DR THE KITCHEN SINK)
While IBM is vague and imprecise about what they think hybrid cloud is, they are very specific about its size: $1 trillion. This big and coincidentally very round number takes IBM’s traditional and oft-mocked playbook of investing precisely $1 billion dollars in new initiatives to a whole new four comma level.
Barron’s inquired as to the source and composition of this magical $1 trillion number and received a surprisingly frank answer:
“When I asked IBM about the $1 trillion opportunity, the company said the projection came from the Boston Consulting Group and McKinsey. And it includes private clouds, public clouds, and virtual private cloud spending. The IBM spokesperson said the hybrid definition includes not just servers, but also software, business process, and services.”
So hybrid cloud for IBM is the kitchen sink (and note hardware comes first on the list). It is anything and everything they already sell (or already resell in the case of Red Hat), albeit in ever diminishing quantities in recent years. A catalyst that “changes everything” is notably absent.
The entire IT market was about $3.65 trillion in 2018 per Gartner, so this is simply reclassifying vast swathes of the existing market to be “hybrid cloud”. IBM is belatedly catching up to its legacy peers with the latest incarnation of cloud-washing (savor the irony of fellow-cloud-also-ran HPE defining cloud-washing in that link), whereby vendors dub their Same-Old-On-Premises-Shit (SOOPS) to be cloudy because it is so much easier than actually building a product portfolio for the cloud era.
And what kind of an aspiring industry leader has to go out of house to figure out the size of their market opportunity? McKinsey should be your go-to for propping up authoritarian regimes, CEO insider trading, and pioneering the Enron business model, but not for forecasting technology markets.
IBM has a long and glorious history of playing fast and loose with their cloud numbers. First they classified a big chunk of their existing revenue as cloud (i.e. they were hybrid cloud before it was cool!), but were then forced to break out what was and wasn’t subscription revenue, which resulted in IBM’s ridiculous cloud and not-cloud cloud revenue lines. Then, they revised their reporting segments and tried to draw attention away from across the board revenue declines with their so-called “strategic initiatives”. Cloud was one of their strategic initiatives, but it turns out they were stuffing mainframe sales into it, at least until it was no longer beneficial to do so (if you click on just one link in this post, it should be the previous one). Bottom line: numbers from IBM are not to be believed.
WHAT DOES RED HAT BRING TO IBM’S HYBRID PARTY? (TL;DR OPENSHIFT? LOL)
While it is disappointing that this deal appears to be a legacy-first, cloud-washing exercise to sell SOOPS, lets try a thought exercise to prolong this post. Pretend IBM actually understands software and further presume they realize they need a coherent software platform spanning both on-premises and the public clouds to have any real hybrid cloud strategy or relevance. What about Red Hat, combined with IBM, “changes everything?”
Red Hat is a nearly $3 billion dollar revenue stream and still growing modestly (though slowing). They grew 15% in FY19 which ended in February (in contrast, AWS and Azure grew roughly three and five times as fast respectively on larger revenue bases in the same time frame). RHEL and other infrastructure products aren’t cloud assets although Red Hat does assert with a seemingly straight face that just running the RHEL operating system makes you hybrid cloud (by that standard, Windows must be a game-changing cloud asset too!). The infrastructure products were 72% of revenue and grew by less than 10 percent. Beyond cloud irrelevance, RHEL also suffers from an aged distribution model for its bits, doing waterfall distributions that roll up a forked set of patches. The rest of the software industry has moved past this to continuous delivery and/or SaaS. RHEL does give IBM some revenue (and software that is certainly more modern than anything from IBM), though how much incremental opportunity it provides is unclear as IBM has been reselling RHEL since 2002.
Beyond RHEL, the other 28% of Red Hat’s revenue is in their grandiosely named “Application Development-related and other emerging technology offerings” other bucket. They have been pushing this business hard as a source of growth as the core RHEL business has slowed. This is an amalgam of acquisitions including “Red Hat JBoss Middleware, Red Hat OpenShift, Red Hat Cloud Infrastructure, Red Hat OpenStack Platform, Red Hat Ansible Automation, Red Hat CloudForms, Red Hat Storage technologies and Red Hat Mobile Application Platform.” Some of this is outright legacy (JBOSS) and some more properly infrastructure (OpenStack, Storage), but these products include what Red Hat believes are its best claims to cloud relevance.
The problem is they are small, undifferentiated, incomplete, non-strategic, and overlap significantly with IBM products based on the same underlying open source products. OpenStack remains a punchline when it comes to public cloud. Ansible is a configuration tool that competes with Chef or Puppet, but is hardly a decisive cloud asset. OpenShift is a container platform and, if the narrative is to be believed, is the hybrid cloud asset, core rationale for this deal, and IBM’s salvation.
After more pivots than I can count from the early days of Platform-as-a-Service (PaaS), OpenShift has settled as a container platform that “brings Docker and Kubernetes to the enterprise”. The challenge is this is a very crowded space, OpenShift is far from the clear leader, and just being able to run containers on-premises or in the cloud isn’t remotely competitive with the full range of services offered by the cloud vendors. The Cloud Native Computing Foundation has certified over 80 Kubernetes distributions as of this writing, including OpenShift and at least two from IBM (as well as from offerings from AWS, Google and Microsoft). Buying OpenShift is yet another ringing endorsement for IBM’s software development capabilities, or lack thereof. At best, OpenShift is in a three-horse race with Docker Enterprise Edition and Pivotal Cloud Foundry (never mind the dozens of other Kubernetes distributions). OpenShift is closer to just being a Kubernetes distribution while Cloud Foundry and Docker offer a broader range of functionality. Forrester puts Docker EE ahead of OpenShift. And my backchannel data suggests both Cloud Foundry and Docker have more paid customers than OpenShift.
But the biggest problem for IBM and Red Hat’s hybrid cloud software ambitions is not Cloud Foundry and Docker, but cloud hyperpowers AWS, Azure and GCP. Containers are fun and irresistible in the Lego-like metaphors they evoke, but they provide no higher-level services, and many will argue are just a brief layover on the way to serverless computing as the native programming model for the cloud. The hyper-scale clouds can sling containers back and forth just as well and probably better, plus they also have an incredibly rich and every-expanding array of higher-level services (e.g. things like databases which many customers find useful). IBM Cloud has struggled both to keep up with the real clouds’ portfolio of services and to just keep IBM Cloud’s minimalist services up and running. Gartner cautions: “IBM has, throughout its history in the cloud IaaS business, repeatedly encountered engineering challenges that have negatively impacted its time to market.” Red Hat brings no services DNA nor higher level services to help with this problem. OpenShift doesn’t “change anything” about IBM’s ability to compete with the hyper-scale public clouds.
But, you may say, as long as we’re setting aside disbelief, why couldn’t OpenShift really take flight under IBM? (Ok, set aside an extra dollop or two of disbelief given they haven’t been able to make this happen for their own comparable container products AND that they are competing with the real clouds every step of the way). Imagine IBM takes some open source cloud middleware and goes big with it? Shades of Linux in the late 20th century! Oh wait, remember BlueMix? We’ve seen this playbook before. It was IBM’s attempt to take the open source Cloud Foundry bits and go big with them in the same space. We’ve seen that playbook and it didn’t work. What “changes everything” this time around?
DOES THIS DEAL MAKE ANY FINANCIAL SENSE? (TL;DR NO)
$34 billion is real money (the total price is actually more than that as IBM originally was going to stop their stock buyback for two years to help finance the deal, but had to reinstate it after their stock dropped to nearly decade lows after the deal was announced due to what can only be interpreted as a lack of enthusiasm for the transaction). The deal also saddles them with $20 billion in new debt (a sum perhaps better spent on cloud infrastructure if you are a member of the CAPEX über alles club).
Beyond the financial outlay, acquisitions are hard. Most fail. Big acquisitions are really hard. Execution and integration risk go up with the size of the deal. This is the biggest acquisition ever in technology software. IBM has done 182 acquisitions in the 21st century and have very little to show for their $52 billion. IBM’s CEO crows that her experience buying and integrating PWC for $3.5 billion mitigates the risk (she neglects to mention her experience buying and vaporizing SoftLayer for $2 billion). Conflating an order-of-magnitude smaller professional services acquisition with a product and technology acquisition might be the deepest insight into what IBM has become (not a product company).
But if the strategic logic is to “change everything” thereby “resetting the cloud landscape”, they’re really paying over $34 billion for OpenShift. Which is a vanilla container runtime. That is open source. That IBM already had at least one of (as does everyone else in the industry, including the hyper-scale clouds). And one that isn’t a dominant brand or implementation. IBM could have bought both Docker and Pivotal at a nice premium for a quarter the price of Red Hat and gotten better assets if that is the strategy.
DOES THIS DEAL “CHANGE EVERYTHING”? (TL;DR NOT EVEN REMOTELY)
IBM knows they can’t compete with the hyper-scale cloud providers, so their current aspiration is to become some kind of gatekeeper whose busloads of offshore consultants in inconvenient time zones will intermediate enterprise customer interactions with the public clouds which will somehow give IBM both strategic leverage and pricing power. Red Hat, despite the lack of any relevant or dominant cloud software assets for achieving this fantasy, is evidently the best option they could come up with (or, perhaps in another post, we can speculate on even more cynical reasons for this deal…). Benediximus Cum Eo!
In reality, this is a strategy much more akin to becoming an MVNO for the cloud (and to be clear, MVNOs are terrible businesses) and hope to stall the cloud transition for as long as they can to sell the same old legacy stuff for a few more years (or at least let the current CEO exit stage right). Meanwhile, IBM’s whole business is up for grabs. Their remaining customers are the disrupted, not the disruptors. Any company in a market where technology matters that is dependent on IBM has a immense, self-inflicted liability (shorting IBM’s top ten customers seems like a strong investment thesis).
And if you actually read this far, please click here. I’m trying to see if anyone actually reads my posts to the end.