IBM’s Red Hat Acquisition Will Pay For Itself By Early Next Year

Big Blue has reported its financial results for the third quarter of 2024, and the thing that stands out most is not the System z16 mainframes and Power10 midrange and big iron are getting a little long in the tooth ahead of new product cycles expected to start in 2025. Revenue slumps in these two core platforms from IBM are expected at this point in the cycle, and in fact, both are holding up better than past generations of systems did this late in the cycle.

No, the interesting thing is that IBM is nearly done getting back the $34 billion that it spent in July 2019 to close the deal to acquire open source software supporter Red Hat that it announced in October 2018.

That Red Hat deal has turned out to be the most significant acquisition Big Blue has ever done in terms of making the company relevant and injecting some new life and different perspectives into the world’s oldest IT company.

The five years of new stories it could tell for hybrid cloud and AI are essentially priceless, but you could say that they compelled Wall Street to think about International Business Machines in a different and more hopeful way.

There is little question that IBM’s market capitalization has grown from $104.9 billion at the end of October 2018 to more than $200 billion, growth largely been powered by the Red Hat acquisition. Assuming market capitalization grows at the pace it now has along with the Red Hat business for the next three quarters, and assuming the “Telum II” System z17 mainframes come out in early 2025 and the Power11 systems come out in May or June, there is reason to believe that IBM’s market capitalization will grow to $233 billion if current trends persist, and the incremental gain in market cap since Red Hat merged into IBM will be 3.8X the cost of buying Red Hat. So, in essence, Red Hat will have paid back the cash for its cost in revenues and nearly four times that acquisition cost over the same time in market cap gains.

To put some numbers on it, the Red Hat products in the aggregate grew revenues by 14 percent in the third quarter ended in September, reaching $1.87 billion in our estimation. We think that somewhere around 70 percent of that revenue is for core system platforms, with the remaining 30 percent being middleware, storage, database, and development software that sits above Red Hat Enterprise Linux, OpenStack, and OpenShift Kubernetes. If you add up all of the Red Hat revenues up from Q3 2019, when the Red Hat deal closed, until now, that is $29.78 billion that IBM would not have had. There is another $4 billion or so to come between now and the end of the second quarter of 2025, and then all the bait IBM used to catch Red Hat will have come back.

When the Red Hat acquisition was completed, the Red Hat systems software stack represented around 9.2 percent of 2019 overall “real” systems revenue at IBM, which is our estimate of sales of hardware, systems software, tech support, and financing relating to its System z and Power Systems lines. In the third quarter of 2024, we think that Red Hat contribution of overall systems revenues from Big Blue rose to 17.5 percent.

It is safe to see that without Red Hat, IBM’s systems business would have continued to dwindle. To an extent, IBM’s combination of the Linux operating system and big NUMA Power Systems iron with fat gobs of main memory that made its platforms good for very large SAP HANA in-memory database workloads, and it is not a coincidence that SAP HANA has been a big driver in the resurgence of the Power Systems line. Linux flat out saved the mainframe business when it was ported to Big Blue’s venerable card-walloper systems back in 2001. It has been a long time since we saw any data, but Linux compute engines on System z mainframes are priced at around a quarter of those running IBM’s proprietary MVS and z/OS operating systems, and the Linux software is also cheaper than IBM’s homegrown software. It did not take long for Linux to be the main driver of new System z capacity revenues.

IBM’s support of Linux breathed new life into its System z business, and to a lesser extent, its Power Systems business, and the acquisition of Red Hat gave IBM a way to get a revenue stream for Linux software and as well as to meet desire of most customers to keep their legacy platforms but also to add Linux partitions to their systems for running new data analytics and artificial intelligence workloads.

Arvind Krishna, IBM’s chief executive officer and the architect of the Red Hat deal, said on a call with Wall Street analysts going over the Q3 2024 numbers that Red Hat had doubled in size since the acquisition, with an annual run rate of $6.5 billion and delivering a compound annual growth rate “in the mid teens” over those five years.  OpenShift, Red Hat’s variation on the Kubernetes container controller theme and the Ansible IT runbook automation tool both grew at more than 20 percent year on year and the core RHEL Linux business grew “in the double digits,” according to IBM chief financial officer James Kavanaugh.

Yes, Red Hat was growing at 20 percent when IBM bought it, and that growth rate held for a while but went down as growth rates inevitably always do when companies get larger and start approaching their limits.

Perhaps even more significantly, OpenShift had an annualized run rate of a mere $100 million when IBM took control and is now running at $1.3 billion, a factor of 13X increase over the five years. There is no reason why OpenShift cannot grow much larger, but it is hard to envision that OpenShift will generate anywhere near the revenues that the WebSphere Java application middleware did over two and a half decades. We shall see. The addition of AI frameworks to Red Hat Enterprise Linux and OpenShift certainly helps.

For now, though, IBM’s generative AI play seems to be in consulting services more than in hardware or software. Krishna said on the call that the cumulative book of business for generative AI projects broke through $3 billion in the third quarter, which is a $1 billion increase from Q2 2024. As with the last quarter, about four-fifths of the bookings are for consulting engagements and one-fifth is for software, which includes the Watson.x framework and analytics stack for GenAI and the RHEL.ai and OpenStack.AI variants of Linux and Kubernetes with AI bundled in.

We wonder if part of the reason that IBM’s Consulting group was under pressure in the quarter was that companies are saving up money to devote to AI consulting engagements. We certainly have seen this recessional effect on general purpose servers and storage when GenAI entered the picture two years ago.

There was some hint of this from Kavanaugh, who said this on the call: “We are operating in a challenging macroeconomic environment and see no change in client buying behavior. At the same time, clients are reprioritizing their IT budgets to prepare for generative AI. While demand for large digital transformations remain solid, our overall signings declined for the second consecutive quarter as we wrapped on record third quarter signings from last year. Despite the weak current demand environment, we are well positioned to capture growth from generative AI.”

Perhaps the biggest opportunity it has is for modernizing the code bases of RPG programs on Power Systems and COBOL programs on mainframes using code assistants that it is developing using its “Granite” homegrown AI models and application code from its own labs and donations from customers and experts. The idea is to run these code assist models natively on mainframe and Power servers, accelerated with its “Spyre” AI accelerators if the on-chip matrix math engines on its systems do not have enough oomph to handle the load.

In the quarter, IBM’s overall revenues grew by 1.5 percent to $14.99 billion. Gross profits were up by 5 percent to $8.42 billion. But thanks to higher overhead costs, higher research and development costs related to the z17 and Power11 platforms coming in 2025, and a $2.75 billion pension settlement charge, even with a $485 million tax benefit, IBM reported a net loss of $330 million in the quarter compared to a $1.7 billion gain in the year ago period.

That stings.

IBM exited the quarter with $13.7 billion in cash and securities, which gives it room to keep taking pension hits as well as to do a modest amount of acquisitions.

IBM’s Infrastructure group, which sells servers and storage as well as operating systems, providing tech support for its machines, generates revenues from Power and z capacity on the IBM Cloud, and peddles secondhand equipment to customers when they buy used gear, had sales of $3.04 billion, down 7 percent in the quarter. Gross profits for Infrastructure were off 4.4 percent to $1.67 billion. Server and storage sales were down 9.8 percent to $1.72 billion, and support was off 3.1 percent to $1.32 billion.

IBM does not provide specific revenue figures for its System z, Power Systems, and storage lines, but IBM did say that System z sales were off 19 percent and “distributed infrastructure,” which is a nonsensical amalgam of Power Systems and storage, fell by 3 percent. It added that storage sales were up, which implies that Power Systems sales were off more than this. It is no surprise that the z16 and Power10 systems, which debuted in 2018, are seeing revenue declines ahead of the delivery of z17 and Power11 machines next year.

The neat thing about the mainframe and the high-end Power Systems sales happening now is that they are very profitable. Big Blue ships mainframe and Power E1050 and E1080 nodes with all of their processor cores and memory cards installed, and customers typically activate only a portion of this when they buy a machine. Later, when these customers need more capacity, they get activation keys from IBM and turn it on. These “upgrades” cost nothing to do, and yield a lot of revenue. That said, the initial sale of the big iron is inherently less profitable because that latent capacity is just sitting there and IBM had to put capital in the box to cover that future sale.

As you can see from the chart above, IBM’s overall revenues took a step function down when it spun out its services businesses into Kyndryl three years ago, but they have stabilized. IBM’s overall “real” systems sales have trended slightly upwards, thanks to the Power10 and z16 server cycles in 2018 through 2020 and Red Hat from 2019 to now.

As best we can figure from our model, IBM’s core “real” systems business brought in $6.16 billion in Q3, up six-tenths of a point, and its pre-tax income was $3.2 billion, up 1.6 percent. That systems business is stronger than Big Blue makes it look the way it cuts itself up to talk to Wall Street.