Opinion: Armonk's $34 billion acquisition of the Raleigh, N.C.-based open source, Linux, and Cloud vendor may be its last chance to return the company to a leadership position in the technology industry.
I tell a lot of people that, although I have been out of the company almost seven years, had a five-year stint at Microsoft, and have since moved on to other things, I still identify myself as an ex-IBMer.
There is a certain conditioning that occurs when one spends any length of time at some companies. Microsoft certainly rubbed off on me, and it was an experience I will never forget, but it was at IBM that I fully accepted and embraced my profession.
Soft spot for Big Blue
At IBM, I truly learned to become a trusted advisor. Sure, I had been in independent consulting for quite a few years, and I had also spent some time at Unisys, but it was learning IBM's consultative methodology at Global Technology Services and working on some very large datacenter consolidations and business continuity engagements that turned me into a systems architect-type.
Going to work at IBM was a childhood dream. Twenty-somethings and perhaps even 30-somethings may not understand this today, but there was a time when the letters "IBM" were synonymous with computers. Computers and IBM were one and the same.
From the 1960s up to most of the 1980s, in the business world you didn't say "computer" you said IBM. Its closest competitors -- Unisys, Hewlett-Packard, Control Data, Wang, and Digital Equipment Corporation -- were far, far smaller and less impactful.
During that period, IBM had more weight in the industry than Apple, Microsoft, Google, and Amazon combined wield today. It was essentially a monopoly. Every large corporation used IBM systems -- as did the whole of the US Government and any entity worth mentioning.
It sounds ludicrous today, but it's true. Every other kid wanted to be a professional baseball player, a doctor, or a lawyer. I wanted to go to work for IBM.
So, I have a very soft spot for Big Blue. And I have watched it mature and evolve and face paradigm shifts that have caused many of its competitors to go extinct.
IBM knows how to pivot
During the 1990s, IBM faced an existential crisis similar to the one it is facing now. In July 1993, the company laid off approximately 60,000 workers, because its mainframe business at the time was shrinking and the PC industry it created in 1981 was dominated by Intel x86 clones from companies like Compaq, Dell, and many, many white boxers.
There was a real possibility that IBM's history might have ended in the mid-1990s. But instead, it pivoted. It moved into software and services with strategic acquisitions. It embraced Linux and open source. It got out of desktop PCs and moved into laptops. It became a boutique contract semiconductor manufacturer.
It also turned the clunky mainframe and the RS/6000 into premium high-performance computing and transactional server brands, System z, and POWER, although its overall hardware business relative to software and services had become relatively small, closer to 20 percent of its annual revenue. But it wasn't shabby.
Under the leadership of Lou Gerstner and Sam Palmisano, IBM did, in fact, turn itself around. It stomped on its competitors in services, strategic outsourcing, enterprise software, and big iron. Sun Microsystems? Toast. DEC? History. HP Superdome compared to IBM POWER and System z market share? Please.
Anyone who bought IBM stock in 1993 at the low and kept it through 2011 when the company's market cap hit its peak would have been very happy indeed.
IBM is again at a crossroads
Now, we are at the end of 2018, and IBM is again at a crossroads. The software and services businesses that held up the company and saved it from extinction in the last round have now dried up.
Still, things are not as dire as they were in 1993. The company has a decent amount of cash reserves, retains a lot of patents, undertakes a lot of important research, and has made key advances in quantum and optical computing, memory, and most notably, analytics and natural language processing. In terms of emerging technologies, IBM still has a lot of life left.
But IBM has too many aging employees, in particular, middle management types that do not significantly contribute to the company's bottom line and are essentially waiting for their stock to vest before they retire. The company has too many unused real estate assets that cannot easily be repurposed. And, it has already sold off all of its large, poorly performing business units, such as its x86 PC and Server businesses (both went to Lenovo) and its semiconductor plants that make its POWER processors that go into its big iron systems.
So, the company has more bloodletting to do, and it has engaged in quite a bit already in the last year or so. However, it's going to have to become a much smaller and much more agile company in order to get healthy. After that bloodletting is done, it needs to make some important choices.
Big Blue's is desperate to make itself relevant to customers that are increasingly looking to Amazon Web Services and Microsoft Azure as green fields of choice for new cloud-based applications and also business continuity services, which is a big part of IBM's datacenter bottom line.
Those big strategic outsourcing deals and big hardware maintenance contracts the company has traditionally depended on are not growing on trees like they used to. It needs new sources of revenue, badly.
Take IBM's mainframe and big iron UNIX systems business, which has become clobbered by client-server systems in the last 20 years. This liability is actually an asset.
Sure, a lot of new application development went into open source and to Wintel, but there are still tons of legacy workloads running on the mainframe and AIX. Those workloads aren't going to leave where they are sitting or without being completely or partially re-written, and that is a monumental task. These are "sticky" workloads.
Microsoft certainly doesn't want that work. Oracle doesn't want to do it (frankly, it wouldn't even if it knew how). HP won't/can't do it. And small and medium-sized SIs staffed with ex-IBMers who know that stuff most certainly won't do it. It's too risky and entangled.
Only IBM can do it. The problem is nobody wants to buy new mainframes or big iron UNIX boxes. Revenue continues to decline in that business year after year -- it was down more than 17 percent last year, and this year, it will be worse.
Heck, nobody wants to make capital investments in anything if they can get away with it, because -- finally -- people are coming to the realization that cloud is an expense and not CAPEX. CAPEX is stupid.
Looking toward the future
So, IBM needs to build a Big Iron Cloud and position it as the target destination for aging System z and AIX workloads that would otherwise be too expensive to re-write or re-implement. It's got a ton of unused datacenter space from old hosting businesses in Poughkeepsie and other underutilized sites where it would be relatively easy for them to stand up something like that.
Assisted living for mainframes. That has to be worth something.
But IBM also needs to look toward the future, not just transition legacy workloads to the cloud and squeeze more milk out of the aging System z cow. The company still has a tremendous amount of brain trust in services and the data center, which it can use to move new workloads to the cloud and to the growth businesses.
The investment in Red Hat will help. Certainly, the containerization, automation, and orchestration assets will be useful in building toolsets, intellectual capital as well as subject matter expertise the company can use in order to bolster its software and services business.
The 5,000-seat and below market segment has traditionally been partner territory, and when it comes to reseller partners for its cloud, IBM is way, way behind AWS, Microsoft, Google, or even (gasp) Oracle, which is now offering serious margins to partners that land workloads on the Oracle cloud.
IBM's partner/reseller ecosystem is nowhere near what it was since it owned the PC and Server businesses that Lenovo now owns. And IBM's Softlayer/BlueMix cloud is largely tied to its legacy software business, which, again, is slowing.
IBM needs something partners will want to sell
To capture partner interest, IBM actually needs something partners will want to sell. Their current cloud is not something that a partner can sell. Partners want AWS and Azure, and to a lesser extent, Google. I think Google has a lot of unrealized potential with certain kinds of workloads and hasn't truly stepped in the enterprise cloud game yet.
In addition to treating its employees like human beings again, and returning morale to levels where identifying as an IBMer is again associated with pride and respect, I believe IBM needs to get its best architects, system engineers, and programmers trained on all those new Red Hat and (gasp) Microsoft cloud technologies.
Amazon and Google could also potentially be approached as partners, but ideologically, IBM and Microsoft have far more in common than IBM has with Amazon or Google -- especially now that they own Red Hat. Microsoft and Red Hat have had vastly improved relations in the last five years or so and Red Hat Enterprise Linux is a fully certified solution on Azure.
With the new Red Hat assets, it will almost certainly want to rearchitect the IBM Cloud running on Softlayer and BlueMix assets to use Red Hat's flavor of OpenStack, its Kubernetes-based OpenShift, and Ansible automation.
But I think that it is very unlikely the IBM Cloud, even when juiced on Red Hat steroids, will become anything more ambitious than a boutique business for hybrid workloads when compared with AWS or Azure. Realistically, it has to be the kind of cloud platform that interoperates well with the others or nobody will want it.
A Red Hat-based multi-cloud management layer?
IBM then needs to build out a hybrid cloud management platform for on-premises workloads using Red Hat technologies that are agnostic to the actual datacenter IaaS stack, such as Red Hat's own OpenStack running on KVM, or even Microsoft's Azure Stack or VMware's vSphere, which partners can then resell.
I envision IBM building a Red Hat-based multi-cloud management layer, although I personally feel the very idea of a typical enterprise embracing "multi-cloud" beyond just integrating identity management between SaaS applications and a single public cloud platform of choice (AWS, Azure, GCP, Oracle) to be a fantasy.
It's not my experience to see large shops spread their resources among multiple public clouds -- because of the complexity involved as it relates to networking and security, and consistency of APIs.
A new catalog of services
IBM needs a new catalog of services that are uniquely IBM and Red Hat, including their software platforms, which would allow the use of hybridized workloads. This includes the legacy big iron stuff they need to get off customer premises and into the popular public cloud platforms, which they would eventually need to land on once they are re-written, extended, transformed or re-platformed.
IBM also has a lot of experience in business continuity, and it works heavily with regulated industries that other public cloud providers have challenges working with, and despite Azure's or AWS's suitability to those workloads, there are many enterprises that would love to have IBMs ability to act as impartial trusted advisor and access to its heterogeneous skill sets.
These companies want private clouds, not public ones, but they don't want to host them or build them on their own. And, yes, they also want IBM's unique and growing products and services like Watson, predictive analytics with Cognos and AI -- and now Red Hat. And partners would love to sell something like that, especially if IBM could use Azure as its main public cloud target while using its own boutique cloud for orchestration purposes.
Not only that, but Microsoft really wants (needs) large hosting provider partners to build Azure-compliant private and public clouds. It sees huge value in Azure becoming viral and the overall ecosystem growing beyond its own datacenters.
Believe me, I know this. IBM with newly infused Red Hat blood could be that partner.
IBM needs to bury the hatchet with Microsoft
For all of this stuff to happen, IBM needs to bury the hatchet with Microsoft. The two companies have been in an effective detente for two decades. However, the old guard at both companies is now gone. So, a rapprochement is long overdue.
That being said, I am not sure Armonk is ready to bury the hatchet with Redmond just yet, even though Red Hat did a long time ago.
In order for this to happen, however, there needs to be a regime change at IBM. Just as Microsoft underwent a painful transformation that upended its long-standing senior leadership and moved toward a cloud-first strategy led by a much younger, tech-savvy CEO, IBM now has to do the exact same thing.
Taking on the personality of Red Hat
It needs to have a technologist lead the company and it needs to disrupt internally. I believe that IBM has to more than just absorb Red Hat, it has to take on the personality of Red Hat. Just as Apple took on the personality of NeXT and Steve Jobs when it underwent its own transformation in 1996, IBM needs to select its future leadership and ongoing strategy from Raleigh.
That is a decision its stockholders and board of directors need to realize. And it needs to make these sweeping changes now while the company still has the potential to rebound. The Red Hat purchase is the company's last chance to become relevant -- it's the tech industry equivalent to a Hail Mary pass.