Jim Whitehurst

President | IBM

Columbus, Georgia

Raleigh, North Carolina

Masters of business administration, Harvard Business School
Bachelor of arts, computer science and economics, Rice University

First job
Building a CRM system for a stockbroker. I was 16.

Too many to list. Chief among them, former Chairman of the Board at Red Hat, General H. Hugh Shelton. General Shelton is one of the most accomplished people I’ve ever had the privilege to work with.




New World Towers: Jim Whitehurst on the Rebalancing of Innovation and Social Impact

Editor’s note: Today and yesterday, CSQ is publishing this two-part conversation with Jim Whitehurst, president of IBM.

If you missed part 1, you can read it HERE.

“The strongest principle of growth lies in the human choice.”
–George Eliot

[The year is 1998.] IBM is At the time, we’re fighting it out with a company called SUSE Linux, which still exists, and Oracle had just launched on Linux. Over the next few years, we consolidated a position as the lead in Linux. People would articulate that we’re like the Kleenex brand of Linux, you think of Red Hat. From that, we then made the giant leap to being a multi-product company with JBoss.

More recently, which got IBM interested, was the move to a cloud-native infrastructure. So, we were an early, early adopter, committer to investors around cloud-native Kubernetes containers. All that’s Linux oriented, and containers are a Linux construct. Microsoft ultimately ported the same construct to Windows, but it’s mainly been a Linux thing.

So, we were in a pole position there, building a leading position around cloud-native infrastructure; things are going great. The only issue, though, is with the birth of public clouds, which are massive partners, because the whole idea is we built this infrastructure that could run bare metal, run on VMware, and run on any of the clouds, which is a great platform.

The big public cloud providers had so much more scale and ability to invest in the services needed to sit on top, which then has value to reinforce the platform. Then we started thinking, well, it would be nice to be part of a bigger company that had more, both sales reach and the ability to invest on top of the platform. Serendipitously, Ginni Rometty (former chair and CEO of IBM) called and said, “Hey, we’d be interested.” There was a clear strategic fit there; we had been thinking about the problem of scale. But we didn’t want to join a cloud company that had a single cloud because our whole model was hybrid, horizontal: How do we cut across platforms? That was the beauty of Red Hat Linux.

People would go in there, and we would sell based on cost. Over time, we found the real value came from when you separated the application from the infrastructure. It allowed so much more flexibility over time for enterprises, which ultimately led to a faster innovation pace. We were then, and we believe entirely now, and so does IBM, that having a horizontal infrastructure allows people to innovate faster. Having that choice, that flexibility over time, and with an unknown future, is critical.

We thought by ourselves; it was going to be hard for us to land the platform. To be in the realm of the container world, we needed a more significant partner. IBM came along, and also shared that vision of, “This thing’s got to run everywhere.” And so, ultimately, that’s how we got to the acquisition. I did that for about nine months, running Red Hat as an independent unit of IBM. Red Hat is still independent, but the board asked me to move to be president of IBM. So, I moved to that role in April 2020. And now the Red Hat CEO is kind of my number two, who’s now running Red Hat as an independent. Therefore, I have a whole new set of challenges to think about as we think about IBM and strategy and what we’re doing as we advance.

Boston Consulting grew out of the military. Everyone has to be in synergy and working together; it’s a system. How did that color that part of your career?

I think that the militaristic background of “sometimes there are checklists, sometimes there are binders, sometimes there are things that one does” is a fascinating thread running through this again, as we’re talking about meeting new workloads and then assigning value and assigning the correct tool based on the gravity of the workload.

I think the question is, we’re talking about data loads, data, gravity, and CEOs and leaders, in particular, assigning value to the workload, yes, but also the outcome of that workload. How do you see those sorts of metrics rolling up into the average boardroom? For IBM, you can talk about meeting the compute needs, the big deals with Bank of America, and so on as sort of proof points. But if I’m running a CPG or running a hospitality business, if there are such things next year as we knock on wood and get out of this pandemic, how do I look at those data priorities? How would you report those things to an ESG committee? Beyond governance, beyond compliance, which we all have to do, where do you see that leadership coming from? And how do I know I’m doing it?

 The analogy I often use for different workloads, and I often use to talk about different kinds of management problems, is “I need to execute this specific known task” versus “I need to go create something new,” as leadership models are different. That’s also true in terms of technology. So, the example I always use is, at Delta Airlines you do not want an agile team experimenting with the safety procedures before your flight? Right? Those are locked down PhDs who study them for years. Before we make a change, there is a massive amount of research in collaboration with the FAA and the rest of the industry.

I was talking to a CEO—and I don’t want to quote him because I hadn’t asked him—of a huge company you’ve heard of, and he said, “You know what, we used to execute every decade, and once a decade we would make some innovation.” He said, “Now it’s more like every year we have to have some degree of innovation while we’re executing. And so, it’s much more of a balance.” And I do think that’s true. Frankly, most enterprise tech teams, and most enterprise leaders, came up in a generation where it was 90/10, where 90% of value creation happened by excellent execution.

We are definitely in a world where half to two-thirds of value creation comes from enabling teams to innovate or create or do things differently. We’re automating a lot of the other pieces, and there’s the pithy example I use. That is, making a car 5% cheaper is a massive and very noble effort. And that’s great to do. But we’re so far down the scale curves; 5% is hard. Getting a car to be used for 95 minutes a day versus 90 minutes a day has more value in many ways. But that’s an innovation. It’s, you know, how do you think about information? How do you think about context?

So, the nature of value creation is changing, and therefore, how we lead needs to change; how we think about data needs to change. That’s been one of the I think most challenging problems, back to your question on data, that I see out there. I always say regarding open source is if a problem is relevant to a Web 2.0 company, there’s probably some great open-source tools out there to help solve the problem, because user-driven innovation, they’re significant IT users.

Suppose it’s not relevant to a Web 2.0 company, not as much as most large enterprises don’t have a whole lot of people involved in open source. We’re still trying to convince general counsels that giving away IP can be a good thing, not a bad thing in general. So, the problem you see in technology is that there are incredible data tools out there coming out of open source. However, they’re mainly intended for companies born and that built their IP when we were smart enough to realize you separate the data from the application logic.

Jim Whitehurst with Robert Brennan Hart, and Chris C. Kemp, former Chief Technology Officer of NASA and Co-Founder of OpenStack. Photo courtesy the author, and taken at an event he hosted where Whitehurst delivered the opening keynote as CEO of Red Hat.

Sounds obvious now, but most enterprises are in a different world where it’s all tangled together. So much of what we are doing is helping people untangle that. Ironically, one of the most significant use cases that we see now with what we call OpenShift, the Red Hat container platform, runs on the mainframe now. So, enterprises are saying, “Wait a minute, I have all these Python machine-learning algorithms that I’ve been running on Amazon on my data. I cannot run them on my mainframe at night?”

The correct answer is, how overtime do you re-architect so your data is separate from your application? And so obviously, we’re doing a ton of work there. But you also didn’t get, because of the nature of that work, how do you think about data cleaning? How do you think about data governance? All of those are difficult.

And I was then taking that all the way forward. This has been one of the fun things I’ll say: I think I speak for most Red Hat engineers when I say this is working with IBM to look way far off into the future and execute against it. So I love Red Hat. My heart and soul are there with what we accomplished. We were taking technologies and iterating on them. That iterative innovation that we talked about in open source and modularity is extraordinary to drive the pace of innovation. But then it’s also kind of nice to say something like, “Wow, we need to think about bias in our AI systems,” and be able to put 25 PhDs against it, to think 10 years out, and then write against it. IBM is one of the few companies with a research component left, whether it’s bias in AI, or quantum, or quantum-safe cryptography.

We had a couple of significant client engagements where they wanted to move applications from these on-premise monolithics to public clouds—any of the same researchers who did what we call “Project Debater” within IBM, the follow-on to beating Jeopardy. To be the champion debater means you need to understand the context around language. Well, guess what? If you can understand the context around human language, you can understand many contexts around source code.

We take the same engineers and technologies, and we can apply it to looking at source code to help migrate those things over so that a lot of what we are doing is “How do you get from here to there, given the complexity of where and how the data resides?” Then importantly, as senior leaders of companies, you will be held accountable for your actions today based on what we know, 10 years from now. I’m going to use a horrible example because I can’t think of another one off the top of my head. The tobacco industry, when did you have a sense? We all know now but 40 years ago was when the surgeon general’s report first came out, right?

We know there’s bias in AI models, but we don’t know enough about it. I think most CEOs wouldn’t even know what questions to ask. In five years, some of those CEOs will be in front of a Senate special committee investigating this stuff. And will get asked, “Didn’t you know there was bias?” To which they’ll say, “There was an article in the Wall Street Journal once, so yeah, I guess?” So, we’re trying to invest in getting ahead of that and have ways to measure model and model drift and governance around that now. It is fun to be the company where we can invest in problems we know will exist and kind of work our way back into it. That’s been one of the exciting things.

We have a considerable effort around anti-bias and AI and how you manage around that. The other one that we do a lot of is auditability. In government-regulated industries, like financial services, you can’t oversimplify this. You can’t make a loan based on a machine-learning algorithm because it’s not auditable, right? It’s a black box. So, we’ve invested in all kinds of technologies that allow you to leverage AI to make decisions, but in a way that’s auditable so that we can check bias. We need to think about some remarkable technologies as we think about the future, recognizing some of the ethical and moral challenges we’ll face in the future and already starting on them. Now. It’s been fascinating to be a part of.

What has IBM done to continue the push away from innovation for innovation’s sake and embrace more efficient decision-making styles?

That’s one of the things I found fascinating about being a part of IBM. I joined officially in April of last year. I’ve formally been an employee for about 9–10 months now in the middle of a pandemic, which is a little crazy. An observation of IBM, which has made it so fascinating, is that most companies are defined around a product or a service or a set collection of products and services they offer. They look and say, “Well, how can we go apply that?” I would say that is true of Red Hat. Open-source software companies, we look for use cases. IBM has been around for over 100 years, and one of the reasons is that it starts with a very, I would argue simple premise. It’s how we translate technology into impact to the enterprise. We’re willing to fundamentally change who we are and what we do. So, if you went back 30 years ago, we were in networking; we were in all kinds of things we sold off. The simple way I would say just one generation, I won’t go all the way through, in the ’90s, in the early 2000s, if you thought about the technology context, at the time, the simplified explanation was client server on the hardware side, and best-of-breed applications on the functionality side. The problem with that is you had, therefore, heterogeneity of hardware and heterogeneity of software. And you had to plug all that together to be able to implement an RP system ultimately. So, on the technology side, IBM built the most extensive middleware portfolio at the time, the whole WebSphere portfolio, to plug all of that stuff together and then build this comprehensive services organization to implement it. Then, many CEOs wanted to hand the keys to somebody to run it. So, we created this big managed outsourcing business to say, we will build it for you, we’ll handle all the complexity for you, hand us the keys, we’ll run it for you, and you’ll get an outcome.

If we roll forward to where we are today, everybody’s trying to eradicate the data center’s heterogeneity or complexity. The whole point is, how do I have a homogenous cloud-like infrastructure where whatever’s underneath doesn’t matter? Because it’s been commoditized and abstracted. It’s all about how do I accelerate the pace of innovation? Now, combined with that is an explosion of sources of innovation. It’s no longer three or four people in any given category of software that are IT vendors. You still have IT vendors. You have a massively greater startup community because there’s so much money floating around in VC. You have open source, which has exploded in many, many ways. You have software as a service as a whole different kind of model instead of vendors. And you have cloud providers that are exposing real fantastic functionality via APIs as services. So, suddenly, you say I have this homogenous infrastructure that I’m going to consume innovation that can come from myriad different places. And I got to do all of that and drive it to build innovation myself as an enterprise.

And on the enterprise side, it’s called a strategic competitive advantage. I mean, every CEO cares about this. I’m amazed how many CEOs can at least say the word Kubernetes.

Suddenly, on the interface with the enterprise side, it’s no longer about we’ll do it for you, it’s we’ll do it with you, and we’ll co-create, we’ll partner. A lot of the journey that IBM is going through around the spin-off of the managed service business, or the acquisition of Red Hat, really does come back to the idea that we’re not going to do everything anymore. We’re not a holistic, full solution provider. But we have to have this hybrid cloud platform. So, an enterprise can consume innovation from wherever they want it.

If the point of technology is not as much around automating rote tasks, it’s more about innovation. You’re going to source components of that innovation from so many places. How do you think about that architecture? That’s where we’ve been going and what we’re trying to deliver. That gets into data and how you’re virtualizing data away from applications. It’s where and how it resides. That set of problems are the areas where we’re focused.

Jim Whitehurst was President and CEO of Red Hat before its acquisition by IBM. Photo: Sharaf Maksumov / Shutterstock

Cloud has gone from the buzzword of the century to constituting much of the fabric of our daily lives. How has IBM, arguably the first cloud provider, kept up with that transformation and managed to stay so relevant?

I remember my very early days at Red Hat. IBM was a significant partner. They’d said they’re going to spend a billion dollars on Linux way back in the day. I used to see Steve Mills and these wood-paneled offices and all of that; it’s nothing like that now, right? I mean, Ginni spent a decade trying to work to change the culture. So, all of that piece of it is different. That’s been gratifying. It’s a whole different place than I think most people remember. So, I’ve felt very much at home on that piece.

They are related to how we think about the market. One of the critical lessons for all of us to recognize is that I’ll say user-driven innovation is extraordinarily powerful. And so, people like Amazon, really good at running e-commerce and websites and all of that. So, guess what? They have unique technologies to be able to do that. They’ve done a great job of expanding beyond that and the other big partners with Red Hat and now IBM, so I have much respect for what they have done and are doing.

But there’s a whole set of workloads where I would argue it’s crazy to try to compete with someone like Amazon or like Google in areas adjacent to their wheelhouse. Now, that said, none of them run the world’s largest financial systems, and these other, you know, mission-critical, transaction-based workloads that currently run in and around the mainframe that we’ve been doing for a long, long time. Or the power systems, the Unix is that the workloads that are there because of a set of characteristics, either instead of performance characteristics or the nature of the workloads themselves.

I’m not trying to defend those things; we’re making changes and evolving. Similarly, Amazon around the set of web-based workloads is a logical place to run web-based workloads. If you’re in financial services, and you are regulated, and you get to show up at regulators and talk about your 500 controls in your control framework around that, we are your best partner to do those types of things.

So, as we thought about this hybrid cloud, the first thing that we have recognized, and that was part of the acquisition of Red Hat, is people have websites they’re going to want to run. And you know what, Amazon’s good at that. And picking on dev in particular, or a specific type of data, things were Google’s good. But if you want to have one common environment or very few familiar environments, if you’re going to have one common security model, if you ever want to think about, “Hey, I love this algorithm today on Google, but what if there’s a new one on Amazon next month that I may want to avail myself of?” That data has gravity and having a common architecture makes much sense.

So, what we’ve done is said, how can we have an architecture that runs from the mainframe to power to any primary cloud provider, runs Nvidia GPUs, runs on ARM so that you can run it all the way out to the edge? That has much value, both in the sense of it makes your ability to run a hybrid environment much less costly and much more secure in a whole bunch of regards. For our public cloud, our ability to take the capabilities we have, especially around regulated industries, and around workloads with a set of performance characteristics, especially around transaction processing, we’re the best there is there.

So, our ability to take those adjacent sets of workloads, we feel terrific about, and I think we’re demonstrating that with many clients now and making much progress. Those aren’t the workloads people want to go and scream loud about on your new website. But those are big essential workloads and, frankly, higher priced and higher margin. So, it’s a place that makes sense for IBM. In the same way, we’re deeply involved in computing, but we shed the x86 server business over time. We’re deeply engaged in the cloud, and we can undoubtedly run hyper-scale workloads. I don’t want to say we can’t; we do that a lot for clients. I think our real competitive differentiation is around regulated workloads and areas like that. So that’s our focus hybrid because you don’t have an architecture like that. You’re never going to be able to innovate as fast as you want. And then, particularly in our workloads, thinking about more similar areas to areas that you would expect IBM to be strong around, mainframe mission critical, hyper-secure.


It is clear to most people that the economy is changed so that it doesn’t seem reversible. Will the slow migration back into office towers take the old way of doing things with it? Or can this moment be capitalized on as the single most significant opportunity to rebalance innovation and progress with the realities of finite resources and disparate social interests? Many Fortune 500 executives on down are already taking their corporations on a path towards that great rebalancing. What remains to be seen is if the forces for change will recognize those they used to deem enemies are now their greatest potential allies. – Robert Brennan Hart


This article was transcribed from a podcast conversation between Jim Whitehurst and Michelle Dennedy on the April 10, 2021, episode of Smarter Markets by Abaxx Technologies (abaxx.tech). To listen to the conversation in its entirety, visit www.smartermarketspod.com