Good morning. Please welcome to the stage Rob Thomas and Bruce Ross. Well, good morning, everyone. It's nice to be here as you wrap up your breakfast and enjoy the morning. We're going to get started here right now with Rob Thomas. I'm delighted to have Rob Thomas, Senior Vice President, Chief Commercial Officer at IBM. I was going to say at RBC. No. I still have problems, Rob, because I worked at IBM at one time, as you know, and sometimes that slips in, but that was only a decade ago. Right. Exactly. It sticks with you forever, so. We'll always welcome you back, Bruce, but you're doing great work at RBC, so. Well, thanks. Having an interview in front of 200 of your closest friends, yeah, always a great thing. Anyway, thanks so much for joining us this morning. Just as we get started, Rob, you've had a wonderful career so far at IBM, just maybe tell people a little bit about yourself and the journey that you've been on in your role today. Yeah. I started in consulting at IBM, so did that for a number of years, and then moved to microelectronics, because IBM used to be in semiconductor manufacturing, ASIC design. We still do a lot of chips for our own systems, but we spun off the microelectronics business a number of years ago. Lived in Japan for a couple of years, and then when I came back from Japan, I started in software. I think that was 2006, 2007. I've spent really all of my career in software since then. Ran engineering for a while, spent a good bit of time on M&A, go-to-market. Really been focused on software for most of my career at this point. Yeah. Well, a few topics that I know everybody will be interested in getting into this morning, so thanks for that. Those of you who know that IBM, wow, that's a meteoric career path and one that a lot of people would like to duplicate, so congratulations on that. As we think about IBM, back 2019, pivotal year, Red Hat acquisition. I think it was 2021, the spin-off of Kyndryl as well, changes at the organization and strategically repositioned the company. I just wondered if you could give people some insight into what your strategic direction is going forward and how this all plays into it for you. I think Arvind laid it out really well, to his credit, when he took over in 2020. At that time, he said, "We're going to be focused on hybrid cloud and AI." I think at that moment, not everybody understood what that meant. Because while hybrid cloud, that seems different than public cloud, we're moving everything to the public cloud. I think the world has kind of come around to that view in the last few years, that public cloud is critical, but most companies are going to do something on private cloud as well, on the edge. We've put a lot of focus into hybrid cloud and what it means to be a great technology provider and consulting services provider on hybrid cloud. He also talked about AI at that point, and that was before the Netscape moment with ChatGPT. Interesting that we actually started our investments in generative AI in 2020. It was probably one of the least popular investments at the time, where we put a couple hundred million USD into building infrastructure to start doing model training, and this was before this was really on anybody's radar. I think probably the key reason for the progress we've made, and there's still a lot more to do, is it's truly about focus. We said we're going to be hybrid cloud, AI. That will be the focus of the company. When you choose those types of segments, working with partners becomes more important because you can't do hybrid on your own. We've put a lot of investments into ecosystem. We've put our software onto AWS and Azure. We've built consulting practices around AWS and Azure. There's never one thing, but I'd say just thoughtful focus in terms of strategic priorities. That drives our product investment. It's also driven the investments that we've made in consulting, and I think we're getting some positive results, but there's always more to do. Yeah, I think you're also as a big consumer of your capability, I think you've also been the engine room for high volume transaction businesses and one that we can really rely on. It's those investments you talk about, how they build on top of companies that process over 1 billion transactions a day has been pretty impactful. I think to your credit, Bruce, and your team, you guys push us constantly. I think you're probably a perfect example of why we thought hybrid cloud would be a significant market, which is you guys were aggressive in moving applications to public cloud. At the same time, you leveraged the mainframe. You're like, "Well, we need these things to talk to each other. We need a level of connectivity." I think you were one of our most significant early customers on OpenShift, Red Hat OpenShift, for modernizing your applications. I think it's been a really good partnership. Your team always pushes us. Even some of the stuff we've started to do on AI I think are good examples. I think it's a good example of technology driving business value, I would say. If you think about some of the things we're doing together around risk, payments, other areas, I think it's places that are important to RBC. Yeah, exactly. Especially in today's world, which stepping back for a second, just talking about the macro environment that we're in. We've got a lot of factors going the wrong way on us in terms of climate change, interest rates going up, economies in trouble. Obviously, global issues that are around the world in different places and different markets. You're in a unique position, I think, amongst firms, regardless of technology, but just firms in general, being in about 170 odd markets. What's the global economy look like going forward for you, and how does that impact your investment thesis? Which parts of the world do you see are going to be better or worse, and how do the investments you're making going to be? How do you juxtapose the two: tougher world and yet need to invest? I do think it's very different based on where you go. About, I guess, two years ago, we decided to get a little more aggressive in our investment in different markets around the world. We were basically looking at what are trillion-dollar GDP countries that are growing north of 5%, so ideally 8%-10% range, that are putting technology at the forefront of what the country's trying to do. You look at that, it's not a huge list, but you see countries in the Middle East, UAE, Saudi Arabia would fall into that. India's in that category. Indonesia's in that category. The challenges and opportunities there are very different because in some of those, there's actually not a lot of inflation, as an example. What they all have in common is an aggressive growth agenda that's driven tops-down in the country, and technology's at the forefront. In many cases, they don't have all the skills they need, so that's a pretty good match for us. I was in the Middle East middle of October. The progress we've made there in a year has been remarkable to see. We'll just keep going because they're going to keep going. The other interesting surprise to me, probably one of the biggest surprises that doesn't get talked about much, is what's happening in Japan. People don't think of that as a high-growth market, but the government in Japan is probably more aggressive now than they've ever been on investment. We've announced the work that we're doing on semiconductor technology, which is really driven by the government sponsorship, working with a company called Rapidus. This has gone from an idea to ground has been broken. You can actually see the fab being built, and this is like in a one-year period. I don't think I've ever seen Japan move this fast. I do think we're at a unique time where governments are realizing the importance of technology in terms of what they need to do as a country, and we want to be a part of that. If you just think about GDP growth, I think classic equation is something like GDP growth is about population growth plus productivity growth plus debt, and not every country has the benefit of population growth. It's going to be about productivity. That's really only going to come from technology, and then you know the credit part probably better than I do, so I'll leave that one to RBC. Well, it's interesting. One of the things we think about quite a bit right now is the whole relationship with chips and you talked about it and the U.S. position versus the Asian markets, the instability over there, the access to chips. Your supply chain's global in nature. How do you think about that? The U.S. government's put a lot of money into chips, but that capacity, as you said, like building a fab plant takes years to bring online. Will the U.S. be self-sufficient in chips, or is it actually, no, they're going to be better, but not able to do it on their own? Self-sufficiencies would be hard to imagine at this point. It just takes a long time, right? If you look at what comes out of Taiwan today, primarily. We have a big partnership with Samsung, so we get a lot of our chips out of Korea as well. Self-sufficient in the U.S. feels like a ways off, but I think the government's doing the right things to start thinking about investing. You talked before about sustainability, supply chains. Obviously, we spend a lot of time thinking about how do we de-risk what we're doing from a supplier perspective, which I think we've done a good job on. The sustainability topic is one that comes up everywhere. I think it was about two years ago, we actually announced a software product line focused on sustainability, I think people are starting to realize now that sustainability is not just good for the climate, but it's actually good for business. A lot of our work is around asset management. How do you optimize assets that are deployed? How do you extend useful life, predictive maintenance, and analytics? We're even doing ESG dashboards now. We bought a company called Envizi that gives you a view of how are you doing against your goals to get to zero carbon emissions. I think sustainability has shifted from it's just the right thing to do to it's also good business, and that's been good to see. Yeah. It's a big thing for us. Just as a big Canadian company, we've become the poster child on climate, it's a big deal, not just for us as within the pressures that you see day to day, but it's actually what our clients are asking for. The ESG strategy, specifically on climate, is a huge deal to us, thanks for that. Back to technology innovation. When you juxtapose a bit of an uncertain marketplace, I know that spending is down certainly in the Western markets, you say, "Okay, but gee, there's all these great technologies that you see, like artificial intelligence, quantum computing, GenAI." I was out in Silicon Valley the other day about three months ago, I'm starting to hear this term, "Well, there's classic AI, Bruce, and then there's GenAI." I'm sitting there going, "Really? Classic AI that's only been around for two years." We started investing in this in 2015 when we created Borealis AI, we have about 100 PhDs that only focus on that today. You're a great partner of ours in that. It's a big space, right? Breadth of space from everything from the compute power that you need, to the data you organize, to the LLMs you build, the capability you bring to market. Just talk to us for a second about, you've made some great investments and announcements, whether it was Granite or more. Maybe just talk about across the breadth of that space, where's IBM really putting its energy, and where does it see it playing in this ecosystem? I kind of laugh about your comment on old AI versus new AI. I forget who said this, but I think it was a computer scientist who made the comment that once it works, it's no longer called AI. It's like a forever moving goalpost, because first it was machine learning, data science, that's AI. We're like, "No, that works, so now it's no longer AI." It will be a forever moving goalpost, I think, for everybody. We're really excited about what we're doing with watsonx. We launched that in May. watsonx is a platform for GenAI. There's a studio, which is called watsonx.ai, for building foundation models. To your point, we've populated with some IBM models. We've also partnered with Meta to bring in Llama 2. We partnered and invested in Hugging Face to bring in open source models. It's basically a way to build foundation models. We've got watsonx.data, which is a modern data architecture for AI. Just this week, we announced watsonx.governance, which probably has a bigger response than I expected, which is really how do you monitor AI that's running in your company to understand how decisions are being made, to be able to create a report that you can put in front of a regulator, which is very important in regulated industries. Really happy with watsonx. We've also built a layer on top of watsonx that we call Assistance. These are different, more functional or line of business type use cases. We've got watsonx Orchestrate, which is for automating back office tasks, largely. I think one example in IBM was we've automated 90% of the work done in some of our HR service centers using Orchestrate, just a way to automate repetitive tasks. Probably one of the things I'm most excited about, which is related to your point on Granite, is something called watsonx Code Assistant, which is about making developers more productive. Granite is a foundation model we built for code. We support about 115 different code languages. We started with COBOL because there's about 300 billion lines of COBOL in the world. Every customer is trying to figure out, how do I run this more efficiently? How do I modernize applications? Because some I may not want to run on the mainframe, some I may want to keep on the mainframe. Since we made that generally available a couple of weeks ago, I'd say the client interest has been off the charts. Interestingly, the biggest use case right now is not even necessarily modernization or migration, it's just code generation. If you don't have enough developers or you can't hire enough developers, being able to use a code assistant just to generate more code can be hugely valuable. With Code Assistant, we're seeing 85% acceptance rates, meaning watsonx will recommend code, developers accept it 85% of the time. That's a huge productivity boost. Yeah, I think it's a couple of those areas that we're really taking advantage of inside the bank, and I think actually others of our size would be absolutely doing, is this whole area on, if you think of DevOps as everything from It starts in Agile with a team, and it ends up in cloud. The code generation piece of it is the piece that hasn't been that highly automated. GenAI is a ability to drive that. We see at minimum a 30% productivity improvement in that code development space. You start to translate that across thousands of developers. Now it'll take time, but we see this as the next frontier of our efficiency and capability to get things to market. I think that your COBOL piece, like a lot of us, while we talk about the new stuff, we talk about digital and everything else, we still sit on these massive legacy systems I talked about earlier. The ability to be able to, as our generation gets older and people start to retire, how do you ensure that you understand that and have that intellectual property on that code to be able to support it, manage it, transform it in the future? That's not easy. I think this type of capability on the COBOL side, we had been making phone calls before they got started on that to help with that space. The one thing that's interesting on this is that with Granite, the thing that's interesting about generative AI and foundation models is the model is generalized for code It supports 115 or so languages. We fed it data for COBOL and for Red Hat Ansible. Those became the first two products because if you basically feed it a lot more code for those two, then you can actually turn it into a product. I think the ability to productize this is almost unlimited. We're thinking about doing Code Assistant for Java. We're working with a lot of banks. What I've learned is that a lot of banks have custom programming languages that they built somewhere in the '60s, '70s, or '80s on many trading floors. Now they're trying to figure out, can we take Code Assistant and generalize to use for their languages? We're finding actually very good results for that. I think this area of code is just getting started. Back to the LLMs. Obviously compute has been an issue. The amount of compute that you need has actually forced people into specific models. You see where OpenAI has been going. You've actually built a number of what I call purpose-driven models by functional capability. Sort of two questions. One, do you see us continue to go down to that purpose-built LLMs? Two, are they going to get smarter so we can run them with less compute underneath them? I believe the only sustainable competitive advantage for generative AI is going to be proprietary data. Maybe not the models themselves. The models will be important, but if you can't feed it proprietary data that solves a business problem for your organization, then I don't think there will be a long-term ROI here. Why do I say that? I think the models itself are very important. In building a base foundation model, the investment is not for the faint of the heart. Round numbers, feels like it's $500 million or so to build a base model. I don't think you're going to see thousands of companies do that, but that's why we build base models. If we're working with somebody like RBC, you can bring your proprietary data to tune the model based on your data. Effectively, that becomes your model. I think that will become the default approach for many companies. We also announced about a month ago that we're now indemnifying our models, which to us was a key point on if a client wants to work with us and tune based on their data and they're using our model, we think we have to stand behind it in terms of providing indemnification, because it gives clients comfort to be aggressive in how they pursue this. I don't think you're going to see every company building foundation models. I think there'll be a handful of those. There'll be a lot in open source, we're doing a lot of work in PyTorch because we think that will be the driver of a lot of innovation around foundation models. I think proprietary data tuning models is going to be where a lot of the competitive advantage comes from. Then being able to, for us, it's how do I create that proposition for the client where I can have a debate with you on your financial lifestyle on your mobile phone or whatever piece of glass you want, wherever you are, whenever you want to have that "debate." I think that this is a key component of it. I think that now at the client side for financial institutions, the difficulty is how do you do it in a safe way that meets your own standards and the regulatory standards? I think that that indemnity that you talked about, to us, that was really important. It's part of what we call an overall control plane of ensuring that we've got bias management, we have explainability, we have a level of control that we know where the data is, who's accessing it, who's got a right to access it, and that we can explain all that easily to a regulator. We also get asked as a financial institution that anything that we take to market, we've got to be able to hold the third party responsible as well. That indemnity that you put forward is really part of what we would call a control plane. Great. Thank you for that. I think it's a key part. I think the other piece, your team's been aggressive on this, to their credit, is around governance. What we've announced on watsonx governance is really the analogy I would use is think about when the FDA came out with the nutrition label. I think this was like the '60s, it was people should know what they put in their bodies. They should know impact of that type of thing. This is basically a nutrition label for AI, which is, do you understand what models are running? Do you understand how they're performing? How does this correlate to the enterprise risk that you're willing to take? I think that this is really going to take off. You probably don't need governance if you're running one model, because you can put enough people around it to develop the reports, watch it. The minute you get to 10s, 100s, 1,000s, 100,000s, which is where most larger companies will get, you have to do this in an automated fashion. You have to use AI to monitor the AI, basically. Right I think we're just getting started there. Let's talk about the data problem. You raised that, we would absolutely agree with you. Inside any institution, certainly a large one. We have these. While we talk about external data and being able to use external data, I don't think we leverage our internal data well enough. That's what everybody's looking for. You talked about organizing the data to be able to leverage the AI. What's your perspective and your investment thesis around data, preparing the data, so that you can take advantage of the models? I'd like to say there's no AI without IA, meaning information architecture. I think most companies still have a lot of catch-up to do on information architecture, data strategy, to be ready to do AI at scale. I think we're at an interesting inflection point in data. If you look at data warehousing market, it goes back, I guess, 30, 40 years, I think we're at the fourth, I would call it an epoch of data warehousing. The first one was OLAP. People had databases. They started doing online analytical processing, that led to the creation of Oracle, Db2. There was a bunch of players in that initial epoch. The next one was really about appliances. We bought Netezza. It was a key part of IBM. Oracle did Exadata. Teradata started to get bigger. Appliances was kind of the second piece. I'd say the third wave is what we're at the tail end of, which is the architecture around separating compute and storage. It was a huge unlock from appliances, but it's also very expensive, and you end up with your data kind of trapped in different silos. You've kind of recreated the problem somewhere else. The reason I say we're at the start of a new epoch is there's a lot happening in open source right now, which is kind of what we've been contributing to and starting to commercialize. It's things like Iceberg, which is an open table format. Your data is no longer locked up. You can store it in an open format. Anything can access it. We're doing work with Presto, which was a query engine that came out of Facebook. The next version is called Velox, which is basically a high-performance query engine. We made the product available in July. The results we're getting and proof of concepts are remarkable, like 2X performance over a separated compute and storage architecture is at half the price. I think there's going to be a major change in the data landscape. This, we've seen this before. I said it's four epochs, and it's like 40 years. It plays out over 10 years. I think we're at the end of this last chapter. We're at the start of this next one, and I think AI will be the catalyst for this because the minute you get excited about AI, you just want more data, and you don't really want to have to deal with, "I need to do a 90-day data quality project or a data movement project before I can get started." I think open architecture is just going to win for data. Yeah. That whole point on, do you move the data to the compute or the compute to the data? How that sits for us, it's federated. It is a hybrid cloud environment because we just can't move that kind of quantity of data. You think of like a billion events a day. You're not going to route them through your core systems and then route them into a public cloud. You'd get killed on ingress and egress fees, right? Yeah. Oh, yeah. How do you be able to operate across the two? Switching topics to you now, on you, if you're okay with it, cyber, right? Cyber, ransomware activity is up over 70% this year. We certainly have to look out not only for ourselves, but every single third party that we deal with. We deal with over 1,000 of them. The world's getting more and more difficult in this space. You guys have made a lot of investments there. What's the future of cyber for IBM in terms of investments, and how do you position this? It's one of the areas where I go to Silicon Valley. I think there are over 700 startups in the cyber space, and everybody's got a great point-to-point solution. As an organization, for us, it's massively difficult to integrate these. What's your point of view? I think you just hit on one of the key challenges. The security market has always been incredibly fragmented. You end up with the problem that you just described you all have, which is you've got thousands of tools, and you become the integrator which is not really ideal from a client perspective. As we think about our security software, one of our big focuses is integrating this as a platform. We have a great install base with QRadar. We're kind of at the heart of a lot of companies in terms of how they administer, monitor, and think about their cybersecurity strategy. We have more to do to bring that together as a platform. You want to integrate identity, you want to integrate threat management, more work to do there. The area I'm probably most excited about is data security. I think if you look at cyber as a software market, that's probably been the place that's been ignored for the last decade. There was a lot of focus on data security around 2010. We acquired a company by the name of Guardium, which has done incredibly well. There's a couple other players in the market, but there hasn't really been a next wave of data security. That's what we've started on now. It's an organic build that we've been doing. We also acquired a company called Polar Security, which has a great team and some great assets to augment what we're doing. I think this notion of multi-cloud or hybrid cloud data security is going to be at the forefront for the reasons on ransomware that you described. That's probably one area I'm very interested in that we're investing in. Yeah. Back to the Not that I want to talk about legal terms all day, that indemnity that you put into place, the work that you're doing on your own control plane, how that all comes together in front of our own risk committee of our board, our regulator, ourselves as we manage it's all massively important. In fact, talking to some smaller regional banks that we've talked to in the past, some of them, because they haven't been able to make that kind of investment in the security side or the regulatory side, has actually held them back from being able to move things into the cloud environment. You will actually slow yourself down if you can't make those investments. Cyber is our highest growth area of spend, on a percentage basis, not on a pure dollar basis. Well, if you think about it, say, in the last 20 years, how many technology topics have really become board-level topics? Certainly, cybersecurity has. I think generative AI has. That's all happened in the last nine to 12 months. It's not often that what you and I do every day becomes something that becomes heavily discussed in every boardroom, but I feel like those are two topics where it's certainly happening. Do you agree with that? Yeah. Board meetings are nice to avoid. Yep. They take a lot of preparation. You're right, if you ask Dave McKay, what are the top three risks that he deals with, I can absolutely tell you that cyber would be number 1 or number 2 that he worries about. I said, "Well, boss, that's my job to worry for you." Even with me worrying for him, he still worries about that in one of the top three. You're right. I think that the regulatory standard keeps going up, and it needs to in many respects. The United States as a market leads that globally. You'll see that. You hit the regulatory standard here, you tend to be able to achieve it around the world. We only have a few minutes left, I just wanted to come back to M&A. You've talked a lot about M&A activity. I know everybody here is interested in, "Well, hey, what's IBM's next move?" You made a huge acquisition yourself with Apptio in the FinOps space. You've got a lot of things that are on your mind. How do you look at the market space for IBM's direction with organic growth, with $6 billion plus of R&D spend, but obviously a market there where you make acquisitions. How do you think about this, Rob, and just generally your areas of interest? It's balancing this. Focus for me every day when I come into the office is really on organic innovation. How are we driving R&D? That's why we're really pleased with things like watsonx, because I think we're getting our rhythm back in IBM in terms of being a great innovator, building new products. M&A is a great way to accelerate things, because on one hand, we could probably build anything, but that's probably not the fastest way to get to market, because sometimes if somebody else has a couple-year head start and we can capitalize on that, we can get it to a lot more clients way faster and drive growth through M&A. I think Apptio is a good example. We probably could have done that. It probably would have taken us 10 years because they've done an incredible job. They've collected data on $450 billion worth of IT spend. You don't get that overnight. Acquiring that put us in a unique position where we can really be an advisor to clients on your technology spend, how can you better optimize it. The part of Apptio that most people haven't focused on, a little bit lesser known, is something called Targetprocess, and that is about the labor that goes against the technology spend. Everybody focuses on cloud spend, technology spend. Makes sense. It's kind of easy. With Apptio, you can figure it out in a couple of weeks how you're going to save money. What the labor piece brings is what is my total cost going against any given project? I'll tell you, when we sit in front of a CIO and say, "Here's the technology spend, the cloud spend, and then by the way, the labor against that," there's like an aha moment. I think we're just getting started with Apptio. Feel really good about that. We're always looking for what else we should be doing in M&A. I think we've done 35 or so acquisitions since Arvind took over, about 70%-80% in software, the rest in consulting. It will remain a key part of our capital allocation strategy and growth strategy. Yeah. We were an Apptio client before IBM purchased them. For us, we publicly said we continue to focus on our efficiency, and this is actually one of the core components of us being able to really not only understand that spend but manage the spend by business unit. From my seat, I can have a very fact-based discussion with any of our business leaders on exactly what those investments are. I think banks are really good at managing their costs and less good at actually managing the benefit of those investments and realizing those benefits over time. This is the type of thing that really helps with that. I'd like to just end on leadership. Rob, you may or may not know this, but Rob's got a blog that is publicly accessible. I've spent some time on that. One of the things you talked about most recently was excellence. I just thought, when I think about what you're doing with IBM, your focus on excellence, Rob, and how the IBM company is turning itself around. I know you're mid-flight on that, but you've made really substantial progress in the last number of years. Maybe just talk a little bit about you're managing such a huge, globally dispersed team, great communication method of it, but just your thoughts on excellence and what that means to you and your team. I think when you look at businesses that are struggling, they normally have, I'd say, two problems. They're probably unclear on priorities, and cost is probably in the wrong places. I think the primary role of a leader is to figure out those two things. How do we define priorities and get everybody aligned to those priorities? How do you allocate capital so that cost is in the right places as opposed to the wrong places? To your point on large organization, distributed teams, you can't just say, "These are the priorities," and assume it will get done. The question is, how are you going to motivate people? How are you going to excite them? I do spend a lot of time thinking about that. I actually write more stuff behind the firewall internally than I do externally because I've found it's a good way for me to get my own thoughts together. It's like if you have to sit down and say, "What do you want to get done in the next year?" The first five versions are normally really bad. You start to refine it. You get it to a level of simplicity you think people can understand it, then it becomes a really good way to communicate what we're doing. I share a lot of writing internally because, one, it's a way to convey, this is what we want to do, this is what's important, but you also get a lot of feedback on, well, this is why that's a bad idea. The thing that's great about IBMers is they'll always speak their mind, you can quickly learn what's going to work and what's not going to work. I think writing is an underrated leadership tool for communicating. It's like there's something about people being able to sit down and really see what you're thinking. You can do videos and stuff like that, we do that, but I think there's something unique about writing, that's why I like doing them. Yeah. I think really good insight there. Just as big organizations, I think the need for us to be able to compress those organizational structures. In the days when we started work, you knew there was a CEO or a senior executive team, they're way up in the stratosphere. You never heard from them other than maybe once a year you got a note, now there's that ongoing dialogue, some of which is great, some of which you want to hide under your desk. I think it was great to have you here. Thank you very much for your openness on what IBM's up to and the things you're doing, and thank you as a client of yours. I just wanted to say a big thank you for all that you and your colleagues do for us. Please join me in thanking Rob for being here this morning. Thank you, Bruce.