International Business Machines Corporation (IBM) Earnings Call Transcript & Summary

Welcome back. Dinesh is with us from IBM. He's SVP of Products. Thank you, Caitlin, too, who's in the front row and the IR group for being here as well, and she's been at the company for many, many years and has tremendous knowledge of the story. So thank you so much for sharing your story. Thank you for making my Masters experience enjoyable while I was on a family trip out of the country. I couldn't figure out how to get access, but IBM helped power the Masters experience. So my entire family thanked me for being -- I was excited, and I wasn't running off my head cut off during our family vacation. So thank you for powering that experience.

Thank you, Brent. Good to be here. This is my favorite conference partly because of the location.

Yes, yes. So when I first sat down with Arvind, he had said software, software, software. I don't know if people truly believed him. But as you said, software has gone from 25% of your revenue to 45% in the last 5 years. And all the other things matter in IBM, but it's clearly a software-first move. And I think that's obviously paid off in a great stock performance. It's paid off in good financials. Your multiple has gone higher. You're ahead of that. So what's gone well? What are the things you still want to get done in the next, call it, 3 to 5 years, your true north for the software business over that period?

Yes. Thanks, Brent. So you're right. If you look at the last 5 years, right, I mean, we have transformed IBM into a software company. Today, from 25% to 45%, close to 45% is what our software revenue is. 5 years ago, we said hybrid cloud and AI. And if you look at the last 5 years, we have really stood by saying hybrid cloud and AI. When all the growth and focus was on hyperscalers, everybody was moving to cloud, and we said hybrid cloud. And we said not every customer is going to move every single data and application to a single hyperscalers, especially given the geopolitical situation we are in today, more and more customers are saying, how do I keep my data where I have the control behind the firewall? Or how do I keep it in a particular MZR or region? So I think that decision/strategy that we put forward 5 years ago on hybrid cloud and AI has paid off. Now within that, right, we said there's 4 core platforms. One is automation because we said as more and more generative AI and AI become prevalent, what is the outcome they're driving? They're driving automation. And we really went on organic and inorganic growth phase for automation. The second pillar that we said is data. The growth in unstructured data because of generative AI is a place where we want to make big bets. I think we announced what's next in our data, and I'll kind of come and give a little more color on it. The third one pillar is Red Hat. I mean, it is what hybrid cloud is all about. When you look at Red Hat, right, how do we do Red Hat virtualization? How do we do application modernization using OpenShift? How do we continue the growth in Red Hat? And the fourth pillar is TP. I mean, as more and more code being written, as more and more applications coming online, the transactional is also increasing. How can we really double down on TP on the mainframe side of things. So those 4 core pillars is what we are really focused on from a software transformation perspective. Overarching with hybrid cloud and AI has really paid off really well.

Yes, it has. Many ask, how does Red Hat keep this type of growth? The growth rate has been really strong. What is keeping those such elevated levels of growth?

Yes. So on the Red Hat side, I mean, I would say, one, obviously, the application modernization. So if I have an application, I want to take that and run it on hybrid cloud. I want to use cloud-native containerized software to really bring it. I mean, Red Hat OpenShift can enable you to do that. The second one is automation, right? Ansible really enables you to automate your application deployment, your application management, all those things. The third one is virtualization. I mean, especially with the challenges that's going on with VMware and pricing and all those things, how can we really make sure that clients have an alternative choice to do virtualization, right?

They don't like 1,000 points of price increase from Broadcom?

Right? So I mean, clients are looking at options. I mean they might not completely move off, but they're looking for options. And Red Hat virtualization gives you that option, right? So that part of the portfolio. So when you ask what will drive the growth in Red Hat, those are the 3 core pillars. And all 3 core pillars are very critical for customers because, one, I want to modernize my application. Or if I'm writing a cloud-native new application, I obviously want to run it on any cloud, right? That is not -- I cannot just go use EKS, which is just for one hyperscaler. So there's a huge opportunity on that space. And that's why our automation business is also doing so well because there's a lot of synergy between Hashi and what we have in Red Hat.

Your CFO called out virtualization on the last earnings call, and I kind of like did a double check like why are we talking about virtualization? It seems -- and then I think everyone kind of understands why. Is there a big migration that's happening you're seeing from other platforms to IBM?

When it comes to virtualization, you mean? Yes, we are seeing interest, right? I mean customers -- I mean, obviously, these are platforms that has been around for decades. So any time you have a platform that's so core to your enterprise, it takes time, it takes planning. It takes a lot of -- bringing the different teams together. But we are seeing tremendous amount of interest from clients on -- and that shows up in the pipeline. We covered that in the first quarter earnings call, partly because of what customers are thinking through into the next generation of applications.

Yes. We -- this elephant keeps showing up in the front of the room every session. So we want to push the elephant out of the room with the macro that like just push that out. It's on everyone's mind, the tariffs, what's going on. We really haven't seen it maybe show up in a big way in the software industry yet. But what are your thoughts on what -- how customers are behaving right now? Is there any change how you're seeing them behave? Maybe it actually helps you because you have so many products, you can consolidate so many other solutions. In a time of uncertainty, they can come to you and have that certainty of consolidation. How do you see it?

So I think any time there is a disruptive element happening in the market, what do customers look for, right? They look for 2 or 3 things. One is the vendor or the software diversified, meaning is the diversification that can allow you to go through that turbulent period, right? So we are well diversified. Two, 80% of our software revenue is recurring. So that means we have less to minimal impact from these tariffs. Three, the tools. In these times is when customers look for optimization. The tools that can automate and optimize, bring more productivity and all that. We have it. I mean when I look at our automation tool, Brent, we probably have the best automation tools, both from IT and business automation that can run on any cloud or behind the firewall. So when you look at all these things, I would say, I mean, the impact to IBM as a company is less to minimal. But then obviously, I want to kind of peel back into the tools aspect because when customers are looking to go through these turbulent phases, they say, how can I really do something to take the human out of the loop? How can I use automation, right? So you look at HashiCorp. Now I want to go deploy it in 5 -- 3 different clouds and on firewall. Terraform can really enable you to go do that. Or I want to do secrets management for my applications on any hybrid cloud or behind the firewall. I can do that with Vault, right? So clients, I think these are the times they will even more look for tools that can enable them and to optimize through this journey because cost becomes a big picture item for them. How do I optimize my cost? How can I make my people more productive, right? I mean COVID is another example. That was a turbulent period that we went through. But look, I mean, we did pretty good, actually really good because we have the set of tools. And last but not least, I mean, we have the trust of the customer, which is a huge element in all of this because who do they trust to come into their enterprise and enable and help them go through this space.

The CEO of NVIDIA said inference is exploding and the amount of compute resources it's going to require is insane, whatever the number, you said 100x or whatever the exact number was. But when you think about what's happening with AI and inference and what we're about to go through, what does that mean for IBM?

Huge opportunity, huge. I think this is probably the biggest opportunity for our lifetime, especially for IBM. Two reasons; one, the data. I mean, when you look at the data, right, where is the growth in data happening on the unstructured side of things? 90% of data growth in enterprises is all around unstructured data. Now you cannot just take that unstructured data. You want to make sure that, that unstructured data is combined with the structured data to give value to the customers. Now if you look at -- I want to stay on the data for a minute, Brent, because -- the reason I say that's a huge opportunity for us is because if you look at the last 12 months, it has gone through a phase. We started with assistance, then we went to RAG, then we went to vector database. Now we are talking about agents. So this field is changing so much, so fast. But the reason I'm so excited about data and the opportunity that we have is that RAG enabled you to take a document, chunk it, put it in a vector database, do a search, similarity search and come up with answers. But it's not giving you the accuracy that you -- a customer expect. So when I call a company to ask about my bill, accuracy matters. I mean don't tell me that $5 went up because whatever reason, I need accurate answers. We can do it because why? We can combine RAG with SQL to give you better answers. How many vendors can do that? So the opportunity for us as IBM in the unstructured space is huge. Two, agents. As agents takes off, what are agents? Agents are basically a tool -- a set of tools with the brain. Those tool could be API calls. Those tools could be workflow. Those tool could be database calls. Who can do those tools, and we have tools in every single space, who can do that in enterprise better than IBM? So our watsonx orchestrate that we announced 4 years ago, even before agents became so popular, it's all about orchestration of tools that we can do using generative AI. So I think the opportunity for us, I believe, in generative AI is not just about inferencing and models and all those things. Obviously, we are playing there with Granite and all that. But the biggest opportunity for us is in the agent space because we have those tools. The second opportunity is on the data side, how do we combine unstructured data with their enterprise data to give real value. I'll give you one client example. So we are working with a major network operator where they're saying, look, they really want to automate that customer who calls to understand about their billing every month. 80% of their customer calls is all about bills. They want to automate it. They cannot get the level of accuracy right now using RAG, they're getting about 75%. They want to get to 95%. We are working with them with watsonx.data to combine their enterprise data with the documents, with the RAG to give them to a 95%. All of a sudden, instead of a customer calling and spending 5 minutes on the call, they can use a chatbot and answer in 30 seconds. 4 minutes and 30 seconds of their life, they got back to do something else. So that is the opportunity that we have out there from an IBM perspective because we understand the enterprise side of things that not many vendors can claim to combine it with what generative AI can do.

Your book of AI business is multiple billions. I think it was $6 billion was the last number, but the majority of it is services. And so everyone asked us, when we flip from a book of business of services to book of software?

Right. So -- we are -- I think we are in the early stages of adoption from the client's perspective. Obviously, they're doing the bookings from a services perspective. Two areas that I see really resonating well; one is the developer side of things, which is watsonx Code Assistant, right? So if you look at WCA for Z, we are really doing well. Why? Customers want to modernize their applications. So if you're running COBOL application, how do I take that COBOL application and convert that to Java applications. Why? Because now I have the skills who can go do those application modernization, all those things, right? So that space, we are seeing a lot of, lot of opportunity. So that -- I think we are at the early stages. We'll see the deployments. We will see more happening there because a developer space, application space is one space generative AI is going to disrupt. The second place that I see a lot of opportunity is the line of business. because I have a workflow that is 8-way join, very complicated. How do I split it into multiple pieces. Using generative AI, I can end-to-end automate it, right? So that line of business where watsonx orchestrate is playing with agents and assistants, is another place where we are seeing tremendous amount of interest because every customer is saying, how do I like really make my line of business much more productive? Because my developers are productive using generative AI. They are writing 30% more code. But my line of business with my applications calls a workflow, it still takes time. How do I make that much more productive, right? So we are seeing a lot of opportunity there. And I think the last piece that I just talked about is the data, the governance of the data and the unstructured data and the unstructured data security is the third place where I think we will see a lot of opportunity going over the next few years because right now, all the focus is on the models, how do we do inferencing all those things. But the next phase is that how can I make these inferencing calls to my back-end systems? How can I get the data trained for it? So that's where I think the next few years where IBM will really benefit from it.

Where would you say -- and I mean this makes a ton of sense that it seems like a lot of these enterprises are confused. So they come to you and say, you roll all the architects and they're like, well, you're not going to build a house without a business -- without an architectural plan. So let your SIs get this done. And the question we get is like, when do you think AI reaches an inflection in software? Is it later this year? Is it 2026? Is it '27? A lot of companies here have said, hey, we got no AI revenue today. It's -- I mean it's really early for the software industry, but everyone is waiting for this moment. When does it go from NVIDIA infrastructure services and flips to the software world? When do you think that we see that in a bigger...

I think that's a really good question. I think we will see it in multiple ways. I say that is because -- so I'll take -- I'll go back to the developer persona, right? So today, there's a huge tremendous amount of focus on improving the productivity of developers. Why? Because every enterprise are going to live and die with 2 things. One is applications and two is data. You're a data-driven enterprise. You want your applications to be up and running all the time. You want to have more applications come online, right? So those are the 2 core things. On the application side, the disruption is happening using generative AI. This is where developers are writing more code and all those things. The reason I said there is other peripheral places where we will see the benefits of generative AI, which hasn't come yet. And that's why I said we are at the early phases of it. I'll give you an example, which is resilience, right? As application developers write more code, what happens? You create more applications. As more applications come online, am I ready to absorb that many applications? Do I have the infrastructure? Do I have the observability? Do I have the recoverability of it? I'll give you a simple example, right? Look at CrowdStrike as an example. When CrowdStrike incident happened, if they would have done canary deployment, which is a phased deployment, they would have caught that problem in Australia or Singapore. They did a global deployment, which brought down the whole global economy. So the same thing is going to go through the face of generative AI, which is customers are going to say, am I resilient? How resilient am I, right? That will become a topic or -- that is a place where IBM has such a -- we have a tool called Concert that can come in and give you a resilience score. So that is all coming as more inference comes, as more productivity of developers comes, as more applications come, that is an area. Nobody is talking about it yet, but that is in the next 2 to 3 years come. What is the risk of me bringing so many applications? How many vulnerabilities am I fixing as part of it? How many am I not fixing, right? That is one. Second, the unstructured data security. Nobody is talking about it. How do I discover my unstructured data? How do I make sure it's encrypted, at rest and in transit. So that's another phase that will come. So there's so many peripheral elements of generative AI that will drive more and more growth in software that we haven't seen it. So when is the inflection point? It's just -- we're just getting started, I think. But there's huge opportunity on many of these peripheral phases, not just on inference and models and all those things, but the money will come from some of the other areas like unstructured data security, the agents, the resilience of the application, how do I measure it? How do I drive it, all those things.

We keep hearing from Databricks and everyone else like 85%, 90% of these AI projects aren't live. They're still in the lab. And it seems like it always comes back to what you're talking about is getting your data right, getting your governance right, all that. So this -- it seems like you have a great position to get ready for the AI moment but even Microsoft called out, it wasn't AI that drove the last quarter. It was the core workloads getting your data ready to go.

Exactly.

We're still on the bulldozer, get the land set up before we can put the AI buildings up. I don't know what the right analogy is.

But I think you named it right, which is how do I make sure I have my data ready, which means how do I discover my data? That's the first step, right? Then how do I catalog my data? How do I govern my data? And then how do I make that data self-service available for my data scientists and others. That is the phase that we have to go through, and there's tremendous amount of data that's being produced if you look at the unstructured. It's not just document or text. There is chat, there's e-mail, there's images, there's videos, there's audios. So that data growth is proliferating. How do I manage that data? And then how do I combine it with my structured data where I can bring it all together, right? I mean if I take Dinesh as an example, all the things that I tweet or I put it on LinkedIn, those are all unstructured set of things. But then I have a structured phase. If I were to combine those 2, can I predict what Dinesh will do in the next 3 months? Or what is his behaviors? Can I use -- understand his behaviors to monetize it, right? So if I go to Starbucks every morning and I went to Boston, can somebody predict me and say, hey, Dinesh, I know you go to Starbucks every morning. There's a Starbucks 5 minutes away from you. Why don't you go? So all of a sudden, I monetize my behavior into something that is a business-oriented; discovery and solution. I think that's where the money is, that's where the world will go, but data is going to play such a big role.

Well, we're seeing that with Salesforce making the acquisition of Informatica, which investors were not super excited about. But when you think about M&A yourself, it's been a big part of the strategy. I guess some have asked, obviously, we're not going to talk about what you're going to do next, but like the pace of what you're willing to do and how you balance that with organic growth inside the company.

Yes. So I kind of talked about the organic side of things, right? I mean there's a few places where we are doubling down on the organic side of things. One is obviously the resiliency to say, how can we really do make sure as more application comes on, we give enterprise resiliency score and actions you can take to improve resiliency. So we did concert there from an organic perspective. The other place where I talked about disruption is the line of business. We did organically did watsonx orchestrate, which is all about agents, agents ops, domain-based agents, assistants, how do we automate the line of business. That's all organic, right? So data security is another place where we are organically driving to say, how can we really make inroads into unstructured data security. On the inorganic side, I mean, we have done a really good job on automation side of things, which is HashiCorp. We have done cost from an Apptio perspective, Turbonomic, Instana. So we have done a portfolio on the automation side. I would say probably the best automation portfolio in the industry that can run on any cloud. Where we will double down is on the unstructured data side of things. And you saw we just closed DataStax acquisition, which is all about NoSQL database, JSON format, all those things. So we have doubled down on the data side of things. So from an inorganic, we are always on the lookout that fits our strategy of these 4 platforms, whether it's automation, data, Red Hat or TP to say, how do we really bring a core focus on these 4 things from an inorganic perspective. So all the acquisitions we have done is really focused on these 4 pillars to say, what holes do we have? And can we fill it in organically? If not, can we do it inorganically?

There's been a view that Oracle feels like they're obsessed with building infrastructure and maybe you're more obsessed with building the software for AI. Is that the right way to think of it or not?

Yes, exactly. We are focused on building software that enables you to do generative AI much more seamlessly, whether it's using unstructured data, whether you're using governance of the data, whether you're using unstructured data security and the peripheral things like automation and also resilience.

I know we're still trying to chew through AI, but everyone goes, well, they got this little side cart thrust that could just be a turbo boost with Quantum at some point. Where do you think -- maybe 40,000-foot view I know we're at a time. But like where is Quantum today? When does it start to take off?

That's another area that we talk about. I mean we talk about Qiskit and what we are doing in that space. I do think there's a huge opportunity on the Quantum side of things. And you see some of the companies out there in the market itself that are making inroads. And for us, I think it's a major bet, the fifth pillar, I would say, in the next few years, I think we would see tremendous amount of opportunity on the Quantum side of things.

Stay tuned. The next chapter. Dinesh, thank you so much.

Thank you.

Good to see you.