What The New Top Brass At Big Blue Means For IBM i
February 3, 2020 Timothy Prickett Morgan
We have been expecting such an announcement for many years, but now all of the pieces are in place and Ginni Rometty can retire to chairman of the board and let other executives steer International Business Machines for the next decade or so.
Sam Palmisano, who was an assistant to former IBM president, chief executive officer, and chairman Louis Gerstner, without question saved IBM from disaster in the early 1990s, tapped Rometty to be president back in October 2011. It is the tradition that IBM’s president becomes the next CEO and eventually chairman, so succession is not a question and everybody else can settle down into the hierarchy and let the top brass call the plays. This was true with the exception of Jack Kuehler, who was named president in May 1989 under CEO and chairman John Akers, which was just as The Four Hundred was founded and this intrepid reporter was about two months away from landing this job of following the AS/400 and its progeny as well as the rest of the IT industry. Palmisano stuck around as chairman for a while and then Rometty had full control. Rometty, who is 62, will step down as president and CEO in April but like her two previous predecessors, will stick around as chairman, probably for a year or two or maybe three.
What is unique about this particular changing of the guard is that IBM’s board of directors is naming a new CEO and a new president at the same time, and they are also not the same person. But just like before, naming the president let’s all of IBM, its partners, its customers, and its investors know who is calling the shots in the next decade or so.
Taking on the role of CEO is Arvind Krishna, who we have run across a couple of times through the decades. Krishna, who is 57, has an engineering degree from the Indian Institute of Technology, Kanpur, and got his PhD in electrical engineering from the University of Illinois. Krishna joined IBM in 1990 and held a number of key technical roles at IBM Research and what would become Software Group – we don’t know what they were because we didn’t know him then and IBM’s bio and his LinkedIn profile are not specific. What we do know is that Krishna created IBM’s security software business (at least in its current incarnation and not including the RACF stuff that has been on mainframes for many decades) and eventually was put in charge of strategy at Software Group under Steve Mills, who ran Software Group for many years and who was a top exec under Gerstner and Palmisano. Krishina eventually took over the Information Management database and data integration software business, which is where we first heard about him. Then he jumped across to hardware and was put in charge of development and manufacturing for Systems and Technology Group under Tom Rosamilia, who is still running Systems Group as well as IBM’s supply chain these days and will soon report to Krishna. After that, in the January 2015 reorganization from Rometty, which broke IBM Research free from Systems and Technology Group after the selloff to GlobalFoundries, Krishna was put in charge of the R&D division. More recently, Krishna was put in charge of its Cloud and Cognitive Software business, setting the stage for the Red Hat acquisition that he proposed and drove to completion.
An aside: Rosamilia was the assistant to Rometty who did all the homework to sell off the System x server business to Lenovo and the IBM Microelectronics division to GlobalFoundries, and based on that alone, he was in the running as the next CEO. (Palmisano was the special assistant to Gerstner. But too much time has passed and Rosamilia, who is 58 and a year older than Krishna, is not young enough to do a long stint at the helm (at least by IBM’s previous standard). So that is why he isn’t getting the job. They have found, as we suspected, a younger person who can take the helm a few years from now when Krishna approaches retirement age: Jim Whitehurst, who has been chief executive officer of Red Hat since 2007 and who is only 52 years old. More on him in a minute, but first, let’s talk about systems.
Here is an aerial view of IBM’s headquarters, built in 1997, on Old Orchard Road in Armonk, New York. IBM moved out of its longtime headquarters at 590 Madison Avenue in New York City in 1964, supposedly because of fears of nuclear war but probably because IBMers wanted a better life in the countryside in New York and Connecticut. The building is a mirror image of the original plan, which was shaped like a Z but it ended up as a kind of lightning bolt S. Which is appropriate for a company that has built systems of one kind or another – starting with custom tabulating machines for the 1890 census in the United States – for 13 decades.
And here is Krishna taking his CEO photo in the boardroom that is part of the glass bay window that is in the lower curve of the S on the right hand of the picture:
Krishna represents everything IBM has done in the past three decades to try to transform itself since what even IBMers call their “near death experience” from 1991 through 1993, when IBM under John Akers went aground and nearly sank as various technology waves tried to capsize and submerge it. Akers took the hit for massive layoffs and writeoffs – eventually half the company was let go, and it took a decade for the company to grow back to that size again under Gerstner and Palmisano. Rometty inherited an IBM that was again missing some of the big waves of change, and selling off unprofitable businesses and buying SoftLayer to become a credible fourth ranked public cloud provider and buying Red Hat to become the hybrid cloud platform of choice in the enterprise were the key moves that Rometty has made to keep Big Blue relevant. Here is Rometty and Whitehurst having a laugh at Red Hat Summit last year when talking about what a good deal IBM got paying only $34 billion for the open source powerhouse:
Whitehurst has not just run an IT supplier, but has been on the other side of the table as the customer and at the end of the table as a consultant sitting between suppliers and customers, which gives him a unique, 3D view of the datacenter. Whitehurst received his bachelor’s degrees in Computer Science and Economics from Rice University, and he also studied abroad at the University of Erlangen-Nuremberg in Germany and at the London School of Economics; he then went to Harvard Business School to get his MBA. After that, Whitehurst worked at Boston Consulting Group for a dozen years, and eventually went to Delta Air Lines in the fall of 2001. Both IBM and Delta have very hierarchical management structures, where everything rolls downhill and a lot doesn’t travel very far back up. So IBM will be very familiar to Whitehurst, and not just because IBM and Red Hat have been partners for more than two decades. After doing a few different jobs at Delta, Whitehurst was tapped to be chief operating officer of the airline, which was having its share of troubles after 9/11 messed up the airline industry and the global economy. Whitehurst was put in charge of dealing with a hostile takeover by rival US Airways and steered Delta out of bankruptcy, and without question when he left Delta in 2007, it was a company that was better off than when he found it.
Red Hat commercializes open source software as it emerges from the various places it does, and the real job is to tame a lot of that chaos so enterprises are more comfortable using it in production. Enterprises can’t have all the smart people – the hyperscalers and cloud builders and IT suppliers who make systems software and applications get a lot of them these days, so they are expensive in their scarcity. Red Hat probably had something on the order of $3.85 billion in sales in calendar 2019, which is an order of magnitude better than any open source company has done and considering that Red Hat is really only selling support, that is like having something like a $25 billion software company that sells perpetual licenses and support as IBM has done.
There is no question that $34 billion is a lot of money to pay for Red Hat, but if IBM wants to have a future, it has little choice but to do this. System z and Power Systems machinery will be around for a long time still, but the revenues and profits from these lines are, generally speaking, trending downwards even as customers consume record amounts of compute capacity on these platforms.
We have said this many times, and we will say this again: IBM needs to build the true successor to the System/360 and the AS/400, embodying their ideals, both together, but on a new platform that can be more broadly adopted. IBM needs to do for enterprises what the hyperscalers and cloud builders have done for themselves – and also build its own public cloud the same way. We can preserve these venerable things and still use them, but IBM has to build the future, with a clean slate, so new companies will deploy that. IBM did that with the System/360 and it did it again with the System/38 and the AS/400. We have some ideas on how this might be done, as we expressed a year before the Red Hat deal was announced.
In the meantime, we don’t think the IBM i or System z platforms are under any kind of threat. IBM is still making money here, and the Power10 chip is largely done and Power11 is being mapped out. We still stand by what we said when IBM announced the OpenPower Consortium shortly after Rometty took over Big Blue, and that is that IBM needed to provide a complete stack and should buy Nvidia for GPUs, Mellanox Technologies for networking, and Xilinx for FPGAs. IBM needs to control compute, storage, and networking, and if anything demonstrates how much profit there is in that control, just look at the $30 billion or so in datacenter product sales that Intel had in 2019 and the $10 billion or so in operating profits it attained even with low or no margins on some of its products. Perhaps getting out of components was not the answer, but rather controlling the whole stack is. Red Hat is a start. IBM will have to do a lot of mergers – not acquisitions – to get back to being International Business Machines. But it could be done, and we could end up with the successors to the System/360 and the AS/400. You could even get clever and call it the Purple System, slash and number not included unless you really need it. And the Purple System/500 would do just fine if that makes you feel better.