IBM at Bernstein Conference: Strategic Shift to AI and Hybrid Cloud

On Thursday, 29 May 2025, IBM (NYSE:IBM) presented its strategic direction at the Bernstein 41st Annual Strategic Decisions Conference. IBM’s Chairman and CEO, Arvind Krishna, highlighted the company’s focus on hybrid cloud and AI, noting both the promising growth prospects and the challenges posed by the macroeconomic environment. The company has increased R&D spending and divested non-core assets to bolster innovation, resulting in mid-single-digit growth and an increase in free cash flow.

Key Takeaways

• IBM’s strategic focus is on hybrid cloud and AI, with significant investments in innovation.
• The company has increased R&D spending by 60% over the last five years.
• IBM’s AI business is valued at $6 billion, with a significant portion in consulting.
• Red Hat is a key growth driver, with mid-teens growth targeted for 2025.
• IBM’s diverse business model mitigates macroeconomic risks.

Financial Results

• IBM reported mid-single-digit growth in revenue.
• Free cash flow has exceeded commitments, growing by $750 million annually over the past two years.
• Software revenue increased from the low 20% of total revenue in 2020 to the forties by 2024, with a growth rate of 9%.
• Red Hat grew 13% in Q1 2025, with a commitment to mid-teens growth for the year.
• The AI book of business is $6 billion, with 80% in consulting.

Operational Updates

• AI integration has led to $3.5 billion in cost savings across enterprise processes.
• Significant automation in HR and IT help desk transactions, with 94% and 75% automated, respectively.
• IBM’s software portfolio is focused on hybrid cloud, automation, data & AI, and transaction processing.
• Red Hat’s OpenShift has seen a 13-fold revenue increase over five years.
• IBM has built 75 quantum computers available on the cloud, with over $1 billion in signings.

Future Outlook

• IBM expects to increase growth rates and free cash flow.
• The software portfolio aims for 10% growth, with Red Hat targeting mid-teens growth by 2025.
• Transaction processing and consulting are expected to see low-to-mid single-digit growth.
• Quantum computing is anticipated to reach quantum utility soon, with quantum advantage by the decade’s end.

Q&A Highlights

• The AI pipeline includes $5 billion in signings, with 20% in software and 80% in consulting.
• Krishna emphasized the market’s misunderstanding of IBM’s growth potential and the deep trust from clients built over 114 years.

For more detailed insights, readers are encouraged to refer to the full transcript of the conference call.

Full transcript - Bernstein 41st Annual Strategic Decisions Conference:

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Okay. Good morning. I think we’re gonna get started right around 9AM.

I’m, Mark Neuen, Bernstein’s new US IT Hardware Analyst, and, my predecessor, Tony Secondhagen, sitting in the front here. Nice to be back at Bernstein after a few years away. And as as one of the most enduring companies in technology, IBM doesn’t need much introduction. However, it is nonetheless our pleasure to welcome Arvind Krishna, IBM’s chairman and CEO. Thank you for coming today.

Arvind has been at IBM for thirty four years, and CEO since April 2020 and was previously SVP of cloud and cognitive software, SVP of IBM Research, and GM of the Systems and Tech Group. As head of Cloud and Cognitive Software, Arvind presided over the acquisition of Red Hat. And since taking over as CEO, Arvind has continued to drive the transformation of the company, with the divestiture of Kindrel, increasing R and D spend, organic growth, changes in go to market and ecosystem, and a focus on productivity and cost savings. So thanks very much for joining us today, Aravind. Could you spend maybe just to start off, spend a couple of minutes on your background, the evolution of the business, and your priorities as CEO going forward?

Arvind Krishna, Chairman and CEO, IBM: Mark, first, thank you very much for having us here. It’s always a privilege to be with Bernstein and to be in this room. So let me start off by, by maybe expanding a little bit on what you just talked about. So as I was coming into the role, maybe a couple of years before that, I had the observation that there’s gonna be two technologies that are gonna be fundamental in helping all our clients succeed, hybrid cloud and artificial intelligence. This is well before the current, manic obsession on AI.

This is back in the 2016, ’20 ’17 time frame. So if you say that those are the two areas where there’s gonna be an incredible amount of opportunity, you have to sort of focus and double down on it. As we began that focus and double down, that led to the acquisition of Red Hat, which I would say that over the last five years has become the leading hybrid cloud portfolio in the market. You then couple that with, if you’re going to focus, then you divest. So the Kindle divestiture, and by the way, a few other noncore assets, was along that journey.

We are focused very much on high innovation. So as a consequence of that decision, our R and D spend has increased by probably 60% in this time frame over the last five years. You couple the R and D increase with also very focused M and A strategy, over 30 properties purchased in the last five years that are all helping both the software and the consulting growth on those two main dimensions. You then say, how about a big go to market change? Why don’t we partner up with big players in the market?

And whether it’s AWS, it’s Azure, certainly SAP, Oracle, we were already good partners with, but others, the latest probably last year is Palo Alto Networks. That allows you to play into their markets. These are largely completely complementary to our technology assets, so it actually helps the client and you go in there. And you couple all of those things together with culture, how do you get people to take a lot more risk inside the company, then be risk averse? How do you couple this with a lot more growth?

And so I’ll wrap it up with so all of that led to that we are now growing as opposed to not growing. We are increasing our free cash flow. We had committed to certain things in late twenty twenty one. We said we would grow in the mid single digits. We have done that since then.

We actually said we will grow free cash flow at about £750,000,000 a year. I would tell you the last two years are ahead of that. And then going forward, we’re going to increase our growth rates, and we’re to increase our free cash flow growth as well.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: And given the dynamic environment we’re in, can you talk a bit about what is most top of mind for your clients going forward from a macro geopolitical environment perspective, and and and how is that impacting client behaviors?

Arvind Krishna, Chairman and CEO, IBM: Yeah. I think we should divide this into two parts. One is there is a lot of uncertainty out there. I would also look at you and say, we did not see that play out actually in the first quarter. When I look at our demand pipelines, we don’t yet see it play out in the demand profile.

But we should acknowledge, there is uncertainty. There isn’t a survey you see that doesn’t talk about uncertainty in the business environment and in the CEO confidence indices as they go around. Now when I look at our business, we are pretty diverse. If I look at our geography diversity, we are about half in The Americas, North And South, about a third in Europe, and about 20% in AP. That gives us one diversity.

A second diversity is that we’re pretty diverse across a large set of large clients across many, many sectors and many, many areas. The big thing we got to think about is what are clients looking for? When there’s uncertainty, clients look for productivity, they look for cost efficiency, and they also look for is there an easier and better way to grow their business or to scale their business as opposed to taking on fixed cost. Technology is often an answer to all three of those. So as you play into those themes, I think that is why despite the bit of uncertainty, despite the interest rate environment, which is probably the single most important element of all this, we see that the demand is still strong on technology.

And also, as we have converted our business, along with the change in the portfolio that I talked about, a lot of our business is now a lot more annuity based than it is transactional or focused being that lumpy. That also helps to alleviate the short term risk.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Great. Thanks for that. Moving on, I think, on to AI. Obviously, it continues to be a very hot topic. Can you walk us through what is the dream case for IBM in AI?

So looking forward five to ten years, across all your different businesses, software, consulting, hardware, where do you see IBM making money going forwards in AI in the various businesses? And out of that, what do you think is the most underappreciated by investors?

Arvind Krishna, Chairman and CEO, IBM: Look. AI is gonna infuse every part of the business, and I’m gonna walk through some very specific examples and numbers. But maybe I wanna begin with where I was thinking I would might have closed, which is our book of business on AI is already at $6,000,000,000, about 80% in consulting and 20% in software. We’re not actually counting any of the hardware, and I’ll give you some hardware numbers in here which are pretty interesting. So AI is gonna infuse every part of the business.

If I look at it, we set out to say that the strategy is how do you begin to unlock a lot of value from enterprise data. So we are very focused on enterprise, not at all on consumer. So as you look at enterprises and what is their issue, their question becomes, how do you unlock value from all of the data inside the enterprise? And so we are very, very focused on an open model strategy. We build our own models, the Granite family, and those are lighter and more accurate than many other models, but they are deployed typically inside an enterprise as opposed to in a consumer context.

We also partner with others. We bring in other open models, whether it’s Mistral, LaMa to name two exemplars so that we can give our clients choice. Moving on from the models themselves, we also believe that there is a huge opportunity for what we are calling orchestration in the old age, you might have called it middleware. At our Think conference earlier this month, we announced that a 50 different agents, about half from us and half from our partners, integrate into Orchestrate so that our clients can use that as one place where they can deploy agents inside the enterprise with all of the issues around monitoring and performance that they worry about. And always, I think in the early stages, there’s always going to be a huge amount of consulting, which comes in to help our clients deploy AI because while they have those views, they don’t always have the expertise in house.

In hardware, and we don’t actually put those in our numbers at all, our latest Telum and Spire processes are inside the mainframe. So the mainframe that is coming out in June has AI capabilities built in. And to just one number, it could do 450,000,000,000 inferences a day. So that gives you a sense of the amount of innovation we’re doing as well as the capabilities we’re bringing out. In storage, our fusion range of products.

So in many, many places, it’s going to be there. So if I look at it, we are going to see that there is going to be growth in software. That’s going to build up over time. There’s already a massive book of business on consulting that’s going to play out, and it’s going to infuse every element of the hardware and software portfolio. So I think that the amount of AI and what it’ll drive is actually deeply underappreciated for how much it’ll impact the business.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: You talked you talked about, how how AI agents, would accelerate the ability of many enterprise to to turn generative AI into real value. Can you talk about where clients are in their journey on this?

Arvind Krishna, Chairman and CEO, IBM: Yeah. So we just finished doing a survey of about 2,000 c suite leaders. So the good news, if you look at the next two years, people are expecting to double their investment in AI. The bad news, when they look at their investments over the past few years, only 25% of them have found that they got the ROI that they were looking for originally. So you’ve gotta ask yourself why.

And the reason is that if you allow it to be, I’ve called it a hundred points of light inside an enterprise, there isn’t a platform. There isn’t a way to scale it. So that is what you have to do. I kind of couch it as, there were too many experiments going on. If I look at 2023 and 2024, I think the era of AI experimentation is over.

You’ve got to sit down and focus on pick the two, three, five areas. You can scale it massively. And as you begin to scale it, you’re not gonna start worrying about cost. You’re gonna start worrying about resilience. You’re gonna start worrying about privacy, and that is when the value comes in for the average enterprise on AI.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: At your Think conference, you talked about, IBM as client zero. Can you talk about, how you’re deploying AI internally and what benefits you’re seeing from that, and where are you on that journey?

Arvind Krishna, Chairman and CEO, IBM: Yeah. So if I look at the aspects of all three, there is cost efficiencies, there is productivity, and then there is growth in the business. So we started on our journey about a little over two years ago, two and a half years ago, and the number is interesting. So when we look at all of our enterprise processes, we have so far taken $3,500,000,000 out. That’s a hard number.

We publish it every quarter. So that gives you a sense of how much it has helped us on the productivity and the cost efficiency side. So to make it very, very real, we use AI in over 70 enterprise workflows inside. So you’ll say, what do mean by 70 enterprise workflows? How do you do the IT help desk inside the company?

How do you do your HR transactions inside the company? How do you do revenue forecasting in finance? How do you do procurement? So as you go across these functions, here are some numbers. On the HR help desk, ninety four percent of all transactions and are fully automated by AI.

Only 6% go over to the human side. On the IT help desk, about 75% are completely automated, only leaving 25, to go back in. On procurement, we think that about half of all procurement can become what I’ll call touchless procurement as we begin to go on. They’re talking about the growth side on our coding, on our programming, depending on the area, anywhere from 10% to 30% of code is being written by AI. That is going to help us produce even more products, so that’s actually growth.

That’s not a cost saving. That will allow us to produce even more products, which is going to help on the growth and the revenue side. By the way, it is true that we’re doing it to ourselves, but it is also true that we’re going to be doing it with our clients as well. If I begin to look at the number of clients now, I think we have almost over 400 clients who are engaging us in some sense to say, okay, we’ve heard about your story. Can you come help do this inside our own enterprise?

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: It’s it’s good practice, good innovation in your house for rolling out to your clients. Yeah.

Arvind Krishna, Chairman and CEO, IBM: Oh, that’s our client zero. So we are the zero. There is a first and a second and a hundredth after that.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: And just moving on a bit more to your transformation, back to the transformation. Your portfolio has shifted to software dramatically. It’s now 45% of revenues. Can you talk about your strategy and the portfolio you’ve put together in software? How do the different pieces fit together and support the acceleration to the 10% growth you’ve talked about in your model and laid out at the recent Investor Day?

Arvind Krishna, Chairman and CEO, IBM: Yeah. So just to put some numbers and to draw a five year arc on it. In 2020, we were about in the low 20% in software, and it was growing at 2%. Twenty twenty four, we hit into the forties of total IBM software, and we grew at 9%. So just to sort of paint the picture that we’ve sort of been delivering on what we sort of talked about.

So what are the areas we’re focused on? So number one is hybrid cloud, and Red Hat is clearly the leading part of that part of the portfolio. And there, we are seeing mid teens growth for this year. That is, by the way, up from the high single digits last year. And inside the Red Hat portfolio, a number of really interesting pieces.

I’ll point to our OpenShift is the biggest success probably inside the hybrid cloud portfolio, 13 times expansion in revenue over the last five years. So that gives you a sense of the high growth that is inside that part of the portfolio. Then automation. And I think automation is an underappreciated part of the portfolio. People need a lot of automation to help run all of their IT infrastructure, whether it be in a public cloud or whether it be inside their own data center.

And here is where we’ve done a lot of acquisitions. We brought in Turbonomic, Instana, Aptio, and the latest is Hashi. If you look at HashiCorp, what it does is it helps you deploy infrastructure using a tremendous amount of automation. And why do you need to deploy infrastructure? You may choose to deploy, for example, one of our partners.

MongoDB is its own company. That’s not one of our assets. But a client may choose to deploy Mongo in their own private cloud. They may choose to deploy it in their data center. They may choose to deploy it on a public cloud.

How do you do that in a common way, in a secure way so that you can get all of the benefits of that and how she helps you do that really, really well. So that’s an example of how we bring automation in, and it begins to scale like crazy. By the way, built on open source. It’s not built on proprietary standards. Then data and AI, and our focus is very much how do we unlock the value of enterprise data.

Ninety nine percent of enterprise data has yet to be touched by AI. So how do we bring all of the power of AI to that enterprise data? And probably the fact in there is a lot of the unstructured data as opposed to relational databases has yet to be touched by AI from inside an enterprise. And the last piece in software is our transaction processing, otherwise the mainframe software. And that very much grows along with the capacity of the mainframe that is being deployed.

And absolutely, we have seen a massive capacity increase over the last few years. So a lot of organic R and D, which is why the portfolio is growing a lot more and a lot of M and A into all these pieces of the portfolio as well. To touch on just two, I touched on Hachi already. Neuralink is another great example of us doing M and A in the hybrid cloud or the Red Hat portfolio, which is which is, I think, the leading way to do virtualization for LLMs across all kinds of different GPUs and infrastructure.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: So focusing a bit more on Red Hat, which you already touched on, you committed to mid teens growth for 2025, as you just mentioned. Can you talk a bit more about the drivers of Red Hat? Also elaborate on what you’re seeing related to the virtualization opportunity. Are you on track here and the opportunities are yielding revenue?

Arvind Krishna, Chairman and CEO, IBM: So I believe we are on track. So we finished last year, I think Red Hat was about 8% for the whole year. This year, as we are entering 13% in the first quarter, and I’m very comfortable looking at the demand and the signings that have already happened over the last few quarters at the mid teens for the year. So just to put that out as numerically. So why?

So the biggest part inside Red Hat is Red Hat Linux. And if you look at the need from clients where there is an incredible amount of cyber threats that are coming at them, so they need a secure operating system. They also would like to get to a common operating system across all of their environments, across all of the different infrastructures. So Linux has become the default answer, I think, for most enterprises. And within Linux, for those who are deploying it by themselves, Red Hat Linux, I think, is the obvious answer.

And so both the need for stability, the growing need for overall compute, and the need for security drive the growth in Red Hat Linux. Next is OpenShift, which is the leading container platform. I think as the world has grown over the last few years, the question is, do I want to build my future on containers or on something else? And I think the answer is pretty clear. It’s on containers.

The estimates are there’s gonna be about a billion new applications by 2028 built on top of containers. So that drives the growth of OpenShift, and that’s what I mentioned, well north of a billion five, 13 times growth over the last five years, and we expect that that growth is gonna keep going. So that’s in the mid twenties, and we expect that that growth will keep going. But also automation in the form of Ansible, which is how people do what I’ll call their runbooks or their codebooks for how they manage all of their applications inside, and Ansible has seen significant growth. By the way, the same way as Linux became the default operating system for running your applications and all kinds of infrastructure, As AI applications are gonna come in, you’re gonna find that that’s gonna add a whole layer of growth to both Red Hat Linux and OpenShift because people want a common infrastructure on which to go deploy their AI applications.

AI applications are largely deployed in containers, but there is a growth vector there as well. And all of that gives me confidence. I mentioned virtualization. The question for clients is simple. If I’m building my next generation of applications on containers, do I really want a different infrastructure for virtualization, or should I put it into the same infrastructure?

And if you put into the same, did you use the word capabilities inside OpenShift? And that is what is driving the growth there, and we are seeing a lot of clients now making that decision that they want a common infrastructure for both containers and virtualization.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: By the way, I just wanted to mention to everyone, we’re running through some unprepared questions here, but do feel free to submit your questions to Pigeonhole. I think there’s a few on there that you can vote on, and I’ll try to get through to a few of the audience questions at the end. So I just wanna let you know that.

Arvind Krishna, Chairman and CEO, IBM: You’re not going to let Tony ask questions.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Moving on to transaction processing. Business seems to have been pretty strong recently. Could you help us dimension how much of that strength is coming from a more favorable pricing environment versus growing the installed volume versus other factors? And then within installed base growth, is that mostly new clients or new sign ins new clients?

Arvind Krishna, Chairman and CEO, IBM: Look, to be straightforward, most of transaction processing is growing at existing clients. While there’s a few new clients, that is not the majority. Now you do get new workloads at existing clients, so that is a piece of it. So first, if I look at the last cycle, so the z 17 or the seventeenth generation of mainframe will come out later in June. Z ’16 was the strongest cycle in terms of capacity growth.

So if you look at that, that then tells you that the majority of what we’re seeing increase is in terms of the capacity growth as well as new innovation that we’re delivering there. An example of new innovation, both around AI, We delivered a product that can help our clients modernize a lot of their COBOL environments. So if you look at it, COBOL is a pretty straightforward, but now considered to be a programming language that not that many people know well. So can you use a piece of AI that can read COBOL, understand COBOL, look at 50,000 lines, and give you a sense of what the logic is? Because in the end, you’re capturing business logic as well as document it as well as help you upgrade it.

All of those capabilities are built in. That’s a brand new innovation we brought to market in late twenty three, and that is one example. The second innovation that we’re bringing in in our transaction processing is around how do you do the systems management of the platform itself. So as opposed to having a lot of people look at all the different elements of what’s going on, could you use AI to effectively become a digital assistant to the humans, which makes the humans a lot more productive, and you’re gonna find that being applied. So there’s a lot of innovation that is coming in there, and that helps growth because that’s brand new revenue in both those examples that I used.

The biggest piece of it is absolutely the capacity growth on the mainframe. Each time you get more spikes in the market, you’re going to get a lot more volatility. Well, that’s going to consume a lot more capacity, and that helps drive growth across that part of the portfolio. Pricing, we typically stop inflation on pricing. So pricing is not the biggest part of all of that.

It is mostly from capacity growth as well as new innovation. By way, if I look at the long arc on this, I think five or six years ago, it may have been with your predecessor that I said, I expect this to be low single digits, maybe low a negative mid single digit growth, but that was back in 2019. Today, I can comfortably say, I see this in definitely the low single digit growth going forward, maybe mid single digits as we deliver on our innovation going forward. So that’s a pretty massive swing in this portfolio over five years from kind of negative mid to maybe, positive mid.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Got it. Got it. That’s that’s very good. Moving on to consulting, can you talk a bit about, key drivers and long term growth trajectory for that business in consulting?

Arvind Krishna, Chairman and CEO, IBM: Yeah. So let me first just acknowledge, there is always when there is uncertainty, as you talked about the macro in the beginning, the first, area which will get impacted is always discretionary consulting. I mean, it just is a fact. It goes back over at least my career in IT watching it always. So let me acknowledge that for the short term.

But if I look actually at the vectors over the midterm and the long term, I’m very comfortable on the consulting vectors leading us towards mid single digit growth or more. Why? People are going to need to leverage both AI and that drives growth in consulting as well as what people are doing around hybrid cloud. And what are some examples there? If I look at hybrid cloud, our signings on just Red Hat, not even counting the hyperscalers, is about 16,000,000,000 since we bought Red Hat.

So over now almost exactly six years, sixteen billion of signings in consulting. That will play itself out. On AI, I talked about the 5,000,000,000 in consulting signings. That will play out over time. What I believe is gonna happen is that while in the short term, there may be some flatness or some not so great growth in consulting, but over the midterm, I see the demand, I see the clients wanting to lean in.

I see them wanting to go deploy this. And the other part of consulting is that we have built really great partnerships with a number of people. Amazon, Microsoft, SAP, Oracle are all billion dollar partnerships. And then if I look at Salesforce, Palo Alto, easily on their way to being there as well. So you put together the hybrid cloud nature, the modernization people need, the mainframe modernization, the AI needs, and the partnerships, and those are pretty strong signals that tell me that this will be a very healthy business in the midterm.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: And then thinking about synergies between consulting and your other businesses, can you talk about that? Now that consulting is building out partnerships with peers to compete with software and infrastructure business, you know, how can we talk about how can you think about synergies?

Arvind Krishna, Chairman and CEO, IBM: I don’t think of it as competition at all. So, take an example. So we build a great relationship with AWS. So when you go deploy AWS at a client and help a client migrate to cloud, that’s going to be maybe for some of their portfolio, maybe for all of their portfolio. It also gives an opportunity to bring in elements of the IBM portfolio that are complementary to what AWS is providing.

And we have talked about many of these examples. So as you do that migration, then it actually becomes a tailwind for the software portfolio as well. By the a lot of these partnerships are with application companies. There is no overlap with SAP, Salesforce, Adobe, ServiceNow, Palo Alto, and IBM. Those are completely complementary.

So having those in the in the partnership bag actually allows us to then attach other IBM capabilities into those. And then if I look at mainframe modernization, something you’ve not talked about, that’s a massive opportunity for consulting as well. As people look to bring AI into the mainframe, can I attach AI inference to every transaction being done so that fraud reduces? You’re gonna need help to go get that done. As people wanna evolve their COBOL perhaps into Java, you’re gonna need a lot of consulting help to get that done.

By the way, all of this consulting is done on top of tools and AI tools that are built in conjunction with the software part of the portfolio, not exclusively. So consulting tools will always include the partnership tools as well, whether from Microsoft or from Lama and others. Those capabilities are in there, but they are built in conjunction with the software team. And so that helps to give a tailwind on both sides of that equation.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: That’s that’s pretty clear. If we can move on to the mainframe business, which you just touched upon there, broadly, it’s been pretty strong, recently. What can you share to give us confidence that this is a fundamental improvement as opposed to a strong z 16 cycle, and or other one off benefits? And can you talk about what you’re seeing so far in terms of client feedback for the z 17, which is coming out in June?

Arvind Krishna, Chairman and CEO, IBM: Yeah. So, June, eighteenth, if I remember, not that I’m counting the days till the new mainframe is available for purchase. Look. As you can imagine, we do a lot of design work with our clients going back over a couple of years, but then a huge amount of work over the last six months to make sure that the capabilities you’re bringing out fit their needs. Number 17% more power efficient in what it can do for the same workloads.

Not that there is any talk about power and wattage and gigawatts or more in the world right now. So we think that that actually plays very, very strongly into the environment. There is overall more capacity. Each time there is news, whether it’s about interest rates or about tariffs or about geopolitics or about trade, it tends to drive spikiness, I’ll use that word, in the financial markets. Given the high correlation between the mainframe and our financial services customers, you can imagine that the workload begins to increase.

As people think about cyber risk and they worry about having a resilient infrastructure that can always stay on, that gives added workload onto the mainframe. Then if you look at what we have done, we are building an AI that goes into the mainframe. We’re building in a lot of cyber capabilities so that people can remain always on and be comfortable around it. We’re building more efficiency so that you’re more efficient in your use of power. And we’re building a lot more capacity that clients can get so that they get that back as sort of price performance.

You put all of that together, and when we look at our early signals, look, this has not yet played out. It’ll play out over the next month and the next year. But when we look at our early signals on demand, I’m very comfortable that we’re going to have a very strong z17 cycle also.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: And then how much benefit from the mainframe cycle should we expect to distributed infrastructure products such as Linklis One and the storage portfolio? And, what are you seeing in those markets more broadly?

Arvind Krishna, Chairman and CEO, IBM: Yeah. So Linux one may be one that the audience doesn’t know a lot about. So Linux Linux one is our Linux capability on the mainframe hardware. The advantage of that is it can sit in the same frame, so it can access the same data, it can access the same IO, it can access the same storage as the z OS side of the mainframe. So that gives a lot of capabilities, but then you can bring all of the Linux capabilities to run on that.

We when we see the last round, more capacity actually got put in terms of compute capacity on Linux one than on zOS. I expect that that is going to continue and keep going. But to talk a little bit about two other pieces that may not be that appreciated. As you roll out mainframe, you’re typically going to need storage. As you need storage, I think d s eight k, IBM storage is becoming more and more the leader in the storage that is attached to the mainframe.

There are others. Hitachi is out there. Dell EMC is out there. But I think that we are taking share because of the innovation that we’re putting on that platform. It tends to drive a lot of tape also because of backup needs, but it also begins to drive in Linux one a lot of AI capability along with Watson X and Red Hat that then gets associated with the Linux one side of it.

So as you look upon that, there is a direct attachment of all of these capabilities that flow from mainframe growth.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Got it. I wonder if we could move on to, quantum computing. It’s a very exciting area. I think you’ve talked about late part of this decade when you’re gonna start to see some impact. I I I think, you know, looking at all the data that’s come from your team presented at your recent Investor Day, you’ve presented a lot of data points which show that you’re in a strong leadership position here.

However, why I’ve just noticed that in the press, it tends to be more companies like Google, Microsoft, perhaps some of the Chinese players that tend to be in the press more recently compared to IBM, which is a little bit inconsistent with what we’re hearing from the company. So, I mean, what what are we missing here? Why is it that, you know, IBM is perhaps in the press less than these other players considering all the strong information you presented at the Investor Day. What are people missing here?

Arvind Krishna, Chairman and CEO, IBM: So I’m gonna give you two analogies and then a bit of data. So first analogy, to succeed in quantum computing, you actually need a quantum computer. All the people can talk about it, and they can put out a lot of press releases, but that to me is words on paper, and they can talk about software simulation. That’s not a quantum computer. So to actually win, to get revenue eventually, you need a real quantum computer.

The fact, we have built 75 of them and put them all out on the cloud so they could be accessed by anybody and measured, are they really a quantum computer? Two, just a fact, I think we’re gonna get to quantum utility maybe soon in a year or two, but probably quantum advantage by the end of the decade. So I don’t wanna get ahead of the reality because you can also get burnt by putting out too much news way too early. Second analogy I’ll give you, let’s think about AI and GPUs for a moment. For those of us who are perhaps more deeply technical than people would appreciate, in 2015, it was obvious to me that GPUs would win the AI war.

I would say that was not at all obvious to investors as you can see it in the value of the companies. When I say obvious, I don’t mean a little bit obvious. I mean, like, 100% obvious. It would just take time to play out in terms of when AI would get to scale, when people would appreciate that you want to run AI on GPUs, and that goes back to work done in the earlier part in 2010 to 2012. As we know in tech, cycles always accelerate.

So to me, it is obvious that quantum computing is going to solve problems just like AI is solving problems. That point is probably the three to four years out. So what is the path from here to there? So now just on where we are. We said at our Investor Day, our total inception to date signings with clients that includes governments, universities, think tanks, as well as commercial clients, is about $1,000,000,000 total signings to date.

That’s not revenue. That is total signings, some of which go over multiple years. So that just gives you a sense of the strength of our portfolio because they wouldn’t be doing all this unless they saw that. Two, the largest quantum computer we built is a little bit over a thousand cubits. That was in late twenty three.

We actually built smaller ones after that in the 400 cubit range because we wanted to improve the error rates and we wanted to improve the coherence and we wanted to march down this path towards fault tolerant quantum computing. So we believe that in the next two years, you’re going to see significant milestones down that path. And in about three to four years, you’re going to find that the mixture of error correction, coherence, improvements in the hardware, and improvements in the software is going to allow us to solve problems that will become truly, truly interesting, including those that have some commercial advantage. That’s why my end of the decade. By the way, I believe the path from here to there is an engineering gated path, not a science gated path.

We’re over the science to get to the kind of milestones I’m talking about. Now to get to maybe 10 times that scale, maybe there’s some more science, but I think the parts ahead of us are engineering gated right now.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: I think what’s interesting about IBM, you’re involved in not just the hardware, but also the software algorithms, application side as well, all the way through the value stack in in Quantum.

Arvind Krishna, Chairman and CEO, IBM: Well, just like any computer, you need to build the equivalent of a compiler and an operating system, so you need to give people that. IBM’s version of that called Quizkit is the most well used. We put it out, by the way, as open source, so it’s actually the most well used. I think the last time we looked, it’s maybe three fourths of all the usage of quantum software out there. On the library and circuit side, I mean, we work with others.

We put out some in terms of how you do chemistry calculations, how you do risk calculations. But a lot of that will be done by partners and by clients also. But we wanna make it easy for them to do that, and that is why our focus on the layer below that, to be done.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Who do you see as your biggest competitor in, quantum computing? Is it one of those aforementioned behemoths, Microsoft, Google, or is it more one of the start ups? One of the one of the Chinese players?

Arvind Krishna, Chairman and CEO, IBM: I think that, number one, you gotta be focused outside The US in terms of where people are gonna make progress. And I think one should not underestimate the effort of nation states, I’ll use that phrase, to want to go after the space because of the advantages. I’ve been pretty public about this. There’s massive commercial advantages, but there’s also natural security implications. So that is the number one.

I always held a lot lot of respect for the start ups in this space. I think that many of the start ups just to throw a couple of names out, I think Quantinuum is worth watching. I think Pasquale is worth watching, but there’s many others also. There’s different underlying technologies that are being looked at. I’m actually much more of a fan of superconducting qubits because it’s very easy to manufacture them on semiconductor lines.

And at the end of the day, manufacturability and cost to manufacturers becomes critical to get deployment and to get ease of use going forward. So but there is there are, things to watch in this space from others. I’ll always be respectful because there’s intellect and there is money out there.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Moving on to capital allocation. How how does IBM think about allocating excess cash between continue to make acquisitions, which you’ve been doing, versus potentially resuming your stock buyback versus growing dividend or paying down debt? And I have a follow-up on m and a.

Arvind Krishna, Chairman and CEO, IBM: Look, the capital allocation strategy is pretty straightforward. Number one, we’re pretty committed to our dividend. But as we get a growing cash flow and we were over 12 last year, we committed 13 and a half for this year, the number one use after the dividend is always going to be M and A as long as it leads to growth and as long as it has a few important metrics, but you said M and A next, so we’ll do that. Now as that number keeps growing, I’ll use the word, it gives us optionality. Do we do more on M and A?

Do we do more on stock buyback? Do we do more on paying down debt? But those are great options to have without having to pick them right now.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Okay. Fair enough. So within within M and A, you’ve been very active and, particularly in software. So when you think about the current state of IBM software portfolio versus hybrid cloud and AI strategy, what are some of the outstanding strategic areas of interest that, you could perhaps pursue with, inorganic growth m and a?

Arvind Krishna, Chairman and CEO, IBM: Look. It’s got to fit three criteria. Number one, it’s got to be in our strategic lanes. We don’t wanna open up new ones at least at this time. The strategic lanes are very simple.

Hybrid cloud, so things that help the Red Hat portfolio. Neuralink was probably the NeuralMagic was the last one around VLLMs. If I look at automation, we’ve done a lot there, and we’ll probably keep doing it there. If I look at data and AI, data stacks was the last one that closed yesterday. So that’s an example of one that’ll help us on unstructured data and how do you make that more real for enterprises.

Keep expecting that as you look at these. But the lens is strategic. Next lens is we have got to be able to get synergy. Otherwise, we can make it grow faster than it would by itself. And we bring both our global distribution as well as the conjunction of all the other properties we have in IBM to help make that happen, and we’ve done that.

Whether we did that to Red Hat originally or we did it to Hap2, we did it to Turbonomic, we made them grow faster than they were growing on their own. And the last little piece is, if it’s of any size, not for little tuck ins, if it’s of any size, we wanna make sure that it’s free cash flow accretive within two years. So those are sort of our criteria, and I gave you sort of the lanes that we focused on.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Kathy, we’ve got a bit of time for two or three audience questions. So, first of all, what part of the IBM transformation story do you feel is most misunderstood by the market?

Arvind Krishna, Chairman and CEO, IBM: Look. I think that our flywheel of growth is probably the most misunderstood part of it. And I begin with, we’re very, very proud of our now hundred and fourteen year history in the technology space. As Mark began by saying, that is somewhat unique in the tech space of people who go back. I mean, the world tends to evolve here really, really fast.

Now that is not something that you want to say and defend. That’s not why I’m proud of it. I’m proud of it because that means that we have trust established. We have a thousand odd clients around the globe who do appreciate and who do trust us. That makes it a lot easier for us to bring in all of the innovation.

So the doubling down on r and d and m and a and adding to the software capability is because it allows us to drive lot more productive and efficient growth because of the incumbency. Then we take that deeper into the market with all of the partnerships and all of the ecosystem growth that we’re doing. And as we lean to new areas for consulting, it builds again on that innovation, but it allows consulting to be able to grow. So I kinda call that the flywheel, and I think that that is not very well appreciated in the market that that does give us a couple of points more efficiency and a couple of points more than efficiency turns into GP and profit and cash flow. And so it gives us a model that, will actually have increasing returns over time as opposed to linear.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Another question here from the audience. Can you discuss the AI pipeline growth and, timing to convert the pipeline to revenue? The sorry. The the question moved quickly. So and can you give an example of pipeline opportunity of implementation, and the revenue?

Arvind Krishna, Chairman and CEO, IBM: Sure. Look, I’ll take a couple of, different examples here. So first of all, when we look at that pipeline, when we say the billion, 20% out of the five end of last year, So that’s software. Software is real revenue. That’s not a pipeline.

When we say but the other 80% is in consulting, that is signings. So signings is more than pipeline. Signings is that the client is committed to go spend it, but consulting signings typically play out over two to three years. So that gives you a sense of where it is. So let’s take two very real examples.

I’ll take one of our health insurance clients. They have signed up for a huge amount of AI. By the way, it’s both software and it’s consulting. How is that playing out? They are using it to deploy.

They’re improving their call center. They’re improving how they process claims. They are improving how they give answers to people when they ask complicated questions on insurance. They’re looking at not deploying it like our client, Xero, inside their own enterprise processes that has got little to do with how to service clients, but how you run the enterprise itself. So that plays out with consulting going in there and doing the work for them.

And as they deliver the work, then the signing converts into revenue. So pretty straightforward, much like consulting has played out on Internet, much like consulting has played out on cyber, much like consulting has played out on doing application deployments. I’ll take a second example. We do a lot of work with our partners, and I’ll pick on Palo Alto Networks. We’re doing a lot of work with them on much of the same.

Can we begin to use generative AI to improve how you do threat detection? Can you begin to use GenAI on how you begin to make customer service more effective? And so it becomes a place that we work with our partners and we work with our clients to go deploy AI to help make them a lot more productive.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Another question here from the audience. How will IBM commercialize its quantum technology in the in the future?

Arvind Krishna, Chairman and CEO, IBM: And and and will people wanna buy a quantum computer? I don’t know. If somebody has a quantum computer that can solve some problems that nobody else can solve, will some people want to buy that? Maybe they’ll want to buy it as a service as opposed to as a physical computer because it does take some unique conditions. They’re not they’re not too hard.

I think any sophisticated client can run them in their own data center, but maybe some will just say, I’ll rent capacity on a quantum computer. Then they’re gonna turn around and say, can you help us take maybe maybe a random or maybe not so random an example? If we can create a new pricing algorithm that somebody in the capital markets can use to price a good in a few seconds as opposed to running a Monte Carlo overnight, Maybe there’s some money in that to be had. Not that financial customers ever care about milliseconds of advantage, do they? I mean, maybe some do.

And if I look at people in materials, maybe if you can take 40% of the oil out from a reservoir in the ground as opposed to 30%, maybe there’s some money in that also. So I look at these as all opportunities. The reason I hesitate from giving a very, very precise answer is we should always have a bit of respect for the market and see, do people wanna pay for the hardware? I actually believe the cycle goes like this. First two, three, four, five years of any new technology is gonna be very much on the infrastructure.

It then very quickly moves to help me deploy it, and I wanna make a lot more money there. And then the third wave, which lasts longer, is all of the applications on top, which is what creates the value. But that’s the way it’s gonna play out.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: That’s fair. It’s very early, obviously, to call exactly. Another question from the audience. What is the impact of DOGE, Department of Government Efficiency, on the growth rate of the consulting activity, average revenue per contract, slower top line growth due to tighter margins and longer cycles, etcetera?

Arvind Krishna, Chairman and CEO, IBM: Look. I mean, step one, we are probably as subject to those as anybody else. But if I look at it, you’ve not heard us talk about it or complain because if there’s something discretionary, like we did lose there were maybe dozen contracts where there’s a question asked. There were a couple that were canceled, but they were around areas where the government has decided to invest a lot less, like USAID. The bulk of the work we do, I call it critical.

I don’t think the VA is gonna stop processing claims. I don’t think that they’re gonna stop paying the employees who are there, so that’s payroll. I don’t think they’re gonna stop looking at Social Security payments. So I look at the bulk of our work, it’s in the critical areas. It’s not in what I would call discretionary.

Was it a couple of projects? Fine. They might get canceled. My reaction is, well, there’s a lot of critical work to be done. So let’s go focus on that and find growth overall.

So, yep, there’ll be a couple that’ll come down, but I think that the growth will outweigh those. So I don’t look upon Doge as a headwind or tailwind for us.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Arvind, we’re just about out of time. I just wanna give you one moment for any final concluding comments.

Arvind Krishna, Chairman and CEO, IBM: Look. I think I concluded where I started. We’re a very different company. Half software as opposed to only 20% software, mid single digit growth for the last many years and accelerating, cash flow growth that is much higher than our revenue growth. I think that if I look at our investors and really good custodians of capital in terms of the returns we’ve been giving them on that, That’s kind of where I would close.

Mark Neuen, Bernstein’s US IT Hardware Analyst, Bernstein: Arvind, CEO of CEO and chairman of IBM, thank you very much for joining us. Thank you, everyone.

Arvind Krishna, Chairman and CEO, IBM: Thank you.

This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.