This conversation at IBM Think '26 examines artificial intelligence, IBM Z evolution and ARM integration, with a focus on hybrid cloud strategies, sovereignty and quantum. Topics include AI-optimized Z systems, native ARM support, storage analytics, security and enterprise deployment considerations.
Ric Lewis of IBM Systems provides insights on Z systems optimized for AI workloads, native Advanced RISC Machine architecture ARM integration, storage innovations and the role of quantum. Lewis highlights built-in AI for inline security, storage analytics and a client-driven "client zero" co-design approach; they emphasize sustained Linux workload growth and broader workload compatibility.
John Furrier and Dave Vellante of theCUBE Research probe deployment timelines, performance and security for mainframes, supply chain impacts and how IBM positions Z, Power and cloud services for agentic AI workloads. The hosts note accelerating infrastructure demand and rising sovereignty pressures that shape enterprise strategies.
Key takeaways: hybrid architectures are inevitable because data is distributed. Z systems demonstrate continued Linux workload growth and will natively support ARM to broaden workload compatibility. Built-in AI enables inline security and advanced storage analytics. A client-driven co-design model guides product development while sovereignty and infrastructure demand shape enterprise strategy.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
IBM Think 2026. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for IBM Think 2026
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for IBM Think 2026.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
IBM Think 2026. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to IBM Think 2026
Please sign in with LinkedIn to continue to IBM Think 2026. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Ric Lewis, IBM
This conversation at IBM Think '26 examines artificial intelligence, IBM Z evolution and ARM integration, with a focus on hybrid cloud strategies, sovereignty and quantum. Topics include AI-optimized Z systems, native ARM support, storage analytics, security and enterprise deployment considerations.
Ric Lewis of IBM Systems provides insights on Z systems optimized for AI workloads, native Advanced RISC Machine architecture ARM integration, storage innovations and the role of quantum. Lewis highlights built-in AI for inline security, storage analytics and a client-driven "client zero" co-design approach; they emphasize sustained Linux workload growth and broader workload compatibility.
John Furrier and Dave Vellante of theCUBE Research probe deployment timelines, performance and security for mainframes, supply chain impacts and how IBM positions Z, Power and cloud services for agentic AI workloads. The hosts note accelerating infrastructure demand and rising sovereignty pressures that shape enterprise strategies.
Key takeaways: hybrid architectures are inevitable because data is distributed. Z systems demonstrate continued Linux workload growth and will natively support ARM to broaden workload compatibility. Built-in AI enables inline security and advanced storage analytics. A client-driven co-design model guides product development while sovereignty and infrastructure demand shape enterprise strategy.
>> Hello, I'm John Furrier, host of theCUBE here in Boston on location, IBM Think with Dave Vellante, co-host. We're the co-founders of theCUBE. Covering, Dave 17 years, our 17th year covering the industry with theCUBE and of course IBM. A lot of changes happened. We got Ric Lewis, the senior vice president here of the infrastructure division at IBM. Ric, great to see you again. Thanks for coming on theCUBE and the IBM hosted studio. Appreciate it.
Ric Lewis
>> Happy to do it. Great to see you guys again.
John Furrier
>> We were chatting out in the hallway, almost like a historical view of the industry talking about the different waves. Now the waves are coming together probably bigger than any wave in the past combined. The infrastructure is the hottest area right now. You look at all the AI build out. It's accelerating the value for agentic. Hybrid has now become the standard. You start to see visibility into the global scale. What's it like for you these days at IBM? Because you're leading the charge, you got Z, which has evolved and continuing to grow by the way. So that's been phenomenal. What's your take right now on the market as AI infrastructure and what's IBM's position there?
Ric Lewis
>> It couldn't be more fun right now. The industry is as exciting as I've ever seen it in my entire career. Every boardroom is having a conversation around what are you doing on AI? What's the plan? What's the infrastructure that underlies that? So the conversations are rapid and you talked about waves. They're big waves and they're coming in big sets quickly. And the conversation has changed radically in just the last 18 months with agentic and how quickly that's changing everything. In infrastructure, we're seeing tremendous momentum and have been for quite some time and that makes it a lot of fun. A lot of decisions to make, but a lot of fun.
John Furrier
>> You got a lot of infrastructure products under your purview. One that we've been fascinated with has been, not the rise, but the continuation and the growth of Z, the Linux support, a lot more workloads coming on Z. You got a lot of real estate on the board and you have a dual architecture with Arm you guys announced on April 2nd. What was the motivation? What changed? Why did that come about?
Ric Lewis
>> You hit it. It's the momentum. So Z, I don't think most people would guess this, but we've been growing 20 to 30% program to program on Z for almost a decade now. And that means people are bringing on new workloads. The AI technology that we've put into Z is really resonating with clients. We have AI built on the chip. We have AI cards that we plug in the systems that we started shipping last year. All of that is resonating. So the conversation we're constantly having with clients is we'd like to bring more workloads to the platform, but we don't want to have them just be the ones that run on Z OS, et cetera. We have LinuxONE systems, but we have clients that are like, couldn't we just bring some of our ARM ecosystem apps that are surround our apps and put those on there? It would have the same security profile. And as long as it's performant, we'd love to just put those on there. So we're making that a reality and designing it.
John Furrier
>> What was the challenge there? Because actually I mentioned a lot of real estate when I saw Z, it's got a lot of, it's not as dense as say it's like super computers. What was some of the challenges with ARM to make that work, if any? And where are you on the curve from marquee customer to full production? Get us through some of that adoption.
Ric Lewis
>> Yeah. So we haven't shipped that capability yet. We have shipped our AI capability. This will be in the next generation of Z systems. Key challenges really always is about making sure that you do it in a performant way. So we haven't announced exactly how we're doing that, but I'll give you some hints. We're going to talk about that a little bit later this year, but it's not going to be emulation. It's not just simple software kind of stuff. We're doing hardware to make sure that we execute in a performant way and have that performance right up to the standard that a client would expect from a Z system.
Dave Vellante
>> So okay, you can't tell us exactly what it is, but maybe what it's not. It's not replacing Z and Power.
Ric Lewis
>> Absolutely not.
Dave Vellante
>> I don't think you're co-designing the silicon necessarily?
Ric Lewis
>> No, no, no. We're integrating ARM technology directly natively into our chips is what's essentially going. So it'll still run all the same Z apps the same way it does. It'll just be able to execute ARM as well. So we'll-
Dave Vellante
>> So expand and give more details on how we do that. A lot of technologies that we developed over the years to be able to-
Ric Lewis
>> So it expands the market. It opens up the ecosystem.
Dave Vellante
>> It's not Graviton.
Ric Lewis
>> Correct.
Dave Vellante
>> It's not a new chip, but it's making ARM a first party citizen inside of it.
Ric Lewis
>> It'll be part of our next generation Z system. So just like as we come out with Z15, 16, 17, we're really creative on those names, you can tell.
John Furrier
>> Z18.
Ric Lewis
>> It'll have this capability.
Dave Vellante
>> What does that do for the economics?
Ric Lewis
>> Just helps. We've already taken a huge step on economics. You would be surprised to know. So we have our Z normal transaction running our database, CICS, IMS, Db2, all of the traditional apps that run on there. But we sell mainframes by the MIPS, by how many instructions are run on those platforms. We've had 3X the MIPS growth in Linux transactions versus those core Z transactions. So people have already been doing app stacking and moving Linux apps onto this. That already was a huge leap in terms of ecosystem and cost, making it simple for them. This will be yet another leap because you don't have to make it to run ZLinux. You actually can use RMAPs. So excited.
John Furrier
>> I just put out a post last night on call Hyperscaler 3.0. It's set in the frame for what's coming around the corner. I want to get your thoughts and your vision on this because Hyperscale 1.0 was AWS, Azure out there and got Google. 2.0 includes the neocloud and you see a lot of action on neocloud's huge demand for GPUs and the infrastructure. 3.0 would be many suppliers across a portfolio of services fully distributed, but the neocloud is getting a lot of attention. How does IBM infrastructure fit in this wave of hyperscale 2.0, which is emerging today and then 3.0, which is fully distributed computing hybrid?
Ric Lewis
>> Yeah. I mean, you said the keyword right at the end of that question, hybrid. I mean, we went from hybrid during phase one of the clouds that you talked about there was almost like, no, no, there'll still be something left. That's not the case anymore. The whole, especially AI has made hybrid not an option but an absolute inevitability. And there's a reason for that. That's because of the data. AI is all about data. If you have bad data, you have bad AI. That data is not in one spot. I think there are players who've said, "Oh, it'll be in one spot." But you talked about it. It's going to be in neocloud, it's going to be big clouds over here. It's going to be on prem, it's going to be everywhere. And if you want to get value from that AI, you have to get at that data. And so that's how we view the evolution in this next wave is it's hybrid. It's hybrid everywhere. And so you got to make sure that your systems make it easy to pipe that data. You see a bunch of announcements from us on that. I have them in storage, making it super easy to get at that data. But you see also bringing AI capability to the data instead of requiring people to take data somewhere to get AI capability on. That's why we built AI into both Z and Power. That's why we built fusion and storage to make sure that data path is simple and allows you to run AI.
John Furrier
>> Heterogeneous has always been the market. And we were talking before we came on camera about how those one chip, x86, now you got GPU, all kinds of processing capability. But today the demand is so high for infrastructure for all kinds of compute at many levels. How does that impact your view of the enterprise? Because in this 2.0 phase, before we get to three, the enterprises are now adopting code, automating code building aid that's going to pave the way for agentic, which is data driven.
Ric Lewis
>> Absolutely.
John Furrier
>> So the infrastructure has to fill the void and that piece of the market is the enterprise. Neoclouds will flower up. We think that's going to be, I think we would agree on that, but that's going to impact the potential consumption for the enterprises. How do you view that?
Ric Lewis
>> I view it as it's going to be a heterogeneous environment. I mean, we've had this conversation of at one time in the industry, everybody said there's going to be one processor and that's all there's going to be. And all the graphics and everything's going to move on to that. Well, we've seen how that's happened. There are specialized processors, RM processors, X86 processors, there's Z processors, there's Power process. It's all about specialization and doing it right. I see that the same in AI infrastructure. Inferencing is not a one size fits all problem. Training, lots of investment going into big training right now. But over time, more of that investment's going to go into tuning, fine-tuning and really inferencing. And that inferencing needs to be optimized for the specific data problem for the data where it is. And so that's how I see it. And we're working really hard to make sure that we're ready for that eventuality.
Dave Vellante
>> So thinking about that heterogeneous environment, you're in the many nines business.
Ric Lewis
>> Nine-ish. Exactly. So 99.9999, whatever.
Dave Vellante
>> I think about my RACF days. I mean, that's the most secure environment, but as the environment gets more heterogeneous and you bring in things like ARM, how do you maintain that level of security, many nines, uptime, et cetera?
Ric Lewis
>> It's core to our value proposition so you know we spend a lot of time thinking about that. And it's one of the most attractive things about a Z platform. We have people doing app stacking of apps that were on x86 processor just because it's one footprint that they can update from a security point of view and they know everything's rock solid. So keeping that security that I just talked about, but keeping that reliability and keeping everything up all the time is more appealing in this era because now as you go to agentic, you can't have things going down. It's doing work that people were doing and you got to make sure that it stays up and keeps working like that. So we spend a lot of time, like when we add new capabilities and new features, we make sure they're in the same paradigm, the same testing, the same everything that we've been doing to deliver those nine nines of uptime. And we do that now with the AI processor that's integrated in that. We make sure that it's tolerant to errors and things that happen and it goes all the way up through our stack, not just in the hardware, but through firmware to OS to the apps that sit on top of that in the platform.
Dave Vellante
>> Arvin in his keynote today talked about the gap between those that are really driving hard in AI and those that are somewhat lagging or directly significantly lagging. Why that lag? How can you help them compress that time to value?
Ric Lewis
>> So usually what we're seeing is clients are dealing with a couple of tough choices and how to get after it. First, there's the thing we talked about, are they hybrid? Can they really get out all of their data and how is that data? Second, what is that data? We talk about it and we say that data can either be a gold mine or a landfill. And if you just take AI to your landfill, you're just going to dig a lot faster in a landfill. It's still not a gold mine. So you have to be really careful about that data architecture and make sure that you are ready, you have the right data. For example, we have this feature in our storage called Content Aware and it allows you to figure out where is the gold in this whole gold mine that's in there and how do I structure that and make sure it's ready for AI? So those two, so it's hybrid then to the data platform. And then the third is you got to scale this and to scale it, you need to use AI in all of these areas to help you scale because otherwise what we find with clients sometimes is their infrastructure is growing faster than they can figure out how to keep it up and reliable. So we have AI enabled control that we embed in the systems to make sure that they can achieve the reliability that they want to achieve and they can still get the security and all of those capabilities. So it's like the old days to build a new CPU chip, you needed the old chip to run the processing. Well, to build this new era of AI, we have to use a lot of AI technology built in so that we can build what we need for the next era of AI.
John Furrier
>> Yeah. And that's client zero that you guys have pioneered. I love that reference to the mainframe having that growth and it ties to the agent world. You can have fit the AI for the workload and that fits into the infrastructure. Dave and I spent a lot of time with our team trying to squint through where the pressure forces are and the infrastructure, mainly on demand. Two that jump out I want to get your reaction to is what Jim Kavanaugh and I talked about at the NYSE was this AI builder persona. It sits at the center of the C-suite, hyperscalers, and then all the data people and the organizations, but it's not just developers. It could be a C-suite person, a line of business. So the builder is going to put a lot of pressure on tokens, more compute. And the other one is sovereignty. With the global climate, people are digging in the telcos in these countries to keep the AI revenue in. So there's a lot of sovereignty conversations that's a little bit more business, but it does put pressure on the infrastructure to be responsive, performant, trusted. How do you view this rise of this demand for this AI builder role, which is going to power the agentic and then the sovereignty piece?
Ric Lewis
>> Yeah. So I'm going to go in reverse order. Sovereignty is our home court. I mean, if you think about it, we do 70% of the world's financial transaction. The most important data that travels around the world is in our Z systems. Those are sovereign systems. They're on prem, they're in environments, they're in banks, they're in financial institutions, they're in insurance. Sovereign is our home court and we know this space extremely well. This AI builder, what we have to do is provide a landscape, a place for those people to come in the sovereign environment and make sure that they have AI optimized cores to bring to that data kind of problem, that they have AI-ready data pipes that I talked about before and the control, the AI-enabled control so that they can keep that environment safe and secure for the application that the people are running. Builder is the new thing. I think there's a whole bunch of roles here. There's builder, there's enterprise architect. They're all going to have to collaborate to say, "What is it I'm trying to achieve? Is it the sovereignty? Is it the global distributed nature?" There's going to be AI in all of the spaces. There's plenty of room for players to operate and we're trying to make sure we can deliver across or as a service across that entire landscape.
John Furrier
>> Yeah. And we go to a lot of events and we see the slides up on stage on the big screen architecture, role of storage is becoming more prevalent in the architecture. Memory, we see what's going on in the supply chain. And you mentioned you guys do a lot of AI and certainly that's the client zero success story for IBM so props to the team. How have you changed your product approach with AI specifically on the infrastructure side to meet those demands?
Ric Lewis
>> Sure, sure. From development, meaning we're now all using, Bob, you heard Bob announced here for our development, both in terms of chip. A giant chip project is really a giant software project that then becomes hardware at the end. So we're using those kind of capabilities, but then a giant chip project has a ton of firmware and OS on top. So we're using software to develop those elements of the stack. We're using AI to help us with our software development in those areas of the stack, but it's not just there. It's also how we're doing our customer support. I own the TLS organization, which is our worldwide network of 13,00 people that support all these platforms. All of our call work is now automated with AI. You can do self debug of your systems if there's an issue of any kind using AI. So it's changed how we do that. It's changed how we manage our global supply chain. I own the global supply chain that's positioned around the world to build all these systems.
John Furrier
>> Get your NAND orders in early.
Ric Lewis
>> Exactly. Exactly. And it has helped us already to make sure that we're well positioned for these supply shocks and things that are going on. So it already has changed my group and my role as much as a lot of the other transformation that we've been doing for the last five years.
Dave Vellante
>> We've mentioned storage a few times. You mentioned NAND. We saw Sam right before the keynote, Sam Warner, he was at a spring at his step because like you say, storage is booming. It's amazing. I mean, six, nine months ago, everybody wanted all flash, all flash correct me I'm wrong, but you still have a hybrid architecture.
Ric Lewis
>> Absolutely. We do flash, we do spinning disc and we do tape, and our tape business is booming.
John Furrier
>> Well, I'm not surprised.
Ric Lewis
>> As you can imagine. As large format spinning-
John Furrier
>> Tape is dead, long lived tape.
Ric Lewis
>> Exactly. As large format spinning disks become astronomically expensive, people are saying, maybe I can put some of that on tape. So we have all three. The cool thing, again, back to your how is AI changing what we're doing, our new flash system has AI management built into it, first in the industry of its kind. There's been a revolution in our infrastructure group of just innovation and some risk taking and being ahead of the curve. We introduced AI in these platforms before there was a ChatGPT moment as we all think of it. So being on the cutting edge and innovating like crazy has turned infrastructure into not just a big profitable business for IBM, but a growth business again and it's something we're super proud of.
John Furrier
>> Arvin obviously can't do a keynote without mentioning Quantum. How is the Quantum Heron line impacting or what does that connect into Z power and ARM?
Ric Lewis
>> So over time we view the world, it's not like Quantum comes and replaces classical computing. It's yet another specialization. I think of it as almost like a co-process or it's a big honk and co-processor that does problems that are beyond imagination. We talked about-
John Furrier
>> It solves problems that compute can't get to.
Ric Lewis
>> That compute can't get to. So it's a different class of thing. Just like GPUs went off and did pixel math and that kind of thing better than CPUs could do. Quantum does a whole new category of things in chemistry and biology and those things that CPUs really can't do very well. So I see it as a whole new branch of computing, except you could still do pixel math with digital kind of the Bitmath that we were all doing. It just wasn't quite as good at it. You can't even do these problems that are in quantum. So it's a big emerging area for-
John Furrier
>> We're seeing an emergence of pre-quantum engineering being facilitated and accelerated by supercomputing. You have a lot of that.
Ric Lewis
>> Yes.
John Furrier
>> How do you see that extending? We're seeing GPUs being used. We're seeing large scale. How does your portfolio fit into that kind of precursor, the preamble or the pretext of that?
Ric Lewis
>> Regular computing, supercomputing, quantum will all exist in parallel. We talk about fit for purpose computing all the time in IBM and they'll be fit for purpose. There will be problems that are best solved and only solved in quantum. There'll be other problems that are best solved in your super computing hardware and there's other problems that'll be best solved in regular CPUs. There'll be better other problems that are transaction problems that are best solved in a Z mainframe system. It's all about fit for purpose and making sure you have the right tool for the job and we're making sure that we're playing in those spaces.
Dave Vellante
>> And you see that when you talk to the labs and you visit, I was just recently in Knoxville at Oak Ridge and you can see they've got the quantum that's coming and they've got the classical computing, they've got super computing and they've got all of the above because it's right fit for the right workload.
Ric Lewis
>> Yeah.
John Furrier
>> It's all true. Question on the enterprise. We're seeing, again, we hinted out earlier about the adoptions coming with the agents obviously and the global sovereignty piece. Where's the excitement in the enterprise? Obviously on prem is back, never left, but it's more relevant now with the data on prem. What's your vision on how the enterprises will interact with this hyperscale 3.0 environment managed services with the neoclouds and the hyperscalers, on prem services, manage their own token factories? How do you see the enterprise over the next few years consuming and deploying the tech?
Ric Lewis
>> My team gets tired of me hearing talking about how segmentation is the answer to a lot of your questions, a lot of your... And I see that the same way. This whole distributed environment is about bringing the right capability to the right data for the right workload. And I just see a proliferation of that specialization, if you will, that I just talked about, the fit for purpose. But you could hear that and think what you mean is a whole bunch of new islands and silos and things like that. I actually don't see that. I just see specialization, but in the big picture, it's still hybrid computing. We're still bringing containers with OpenShift to that entire environment. We're still bringing sovereign core to be able to manage those environments. You can see us gluing that hybrid cloud together. We talk about our strategy of AI, hybrid cloud, and quantum. Our job is to bring those three really strong technologies and make them fit seamlessly together and deliver that to clients around the globe on their most important problems, whether it's healthcare or their most important data, which is all the transactions of the world, all that stuff, that's where we play.
John Furrier
>> How would you say this market is right now? Because we're seeing a lot of co-design, roadmap sharing. I mean, hardware and software businesses would do that in the past. What's changed the most in this wave, in your opinion, in terms of how you go about your business? Is it joint development with customers? We're seeing a lot more co-design moving away from supplier relationship and also as a leader, how has this changed your job and your ability to execute?
Ric Lewis
>> Two things, two simple answers. One, speed. It's a lot faster now. You got to be just moving like crazy. Everything is accelerated. So that's a difference and that there's a whole bunch, there's a half hour talk we could do on that. The other is client intimacy if the world is evolving into specialization and solving specific problems is absolutely paramount. Some people have said, "Well, were you guys geniuses?" There for a while when we first introduced AI processing inside of Z before there was a ChatGPT and all of this wave of AI, people were like, "That seems like that might be a little crazy." Then we looked like geniuses.
John Furrier
>> It's a good bet.
Ric Lewis
>> And it was a good bet, but where did it come from? It actually came from clients. It didn't come from us, "Hey, I think we could put together these widgets and make this happen." We had our financial clients coming to us saying, "Fraud is a billion-dollar industry for it's a major problem. We need a way to be able to do that." Fraud a few years ago is you get a note from your company seven days later that says, "Hey, we think there might have been a fraudulent transaction." Well, great. It's already happened. It was seven days ago. They needed that in line in the system.
John Furrier
>> We had the customer loops. Dave, we were telling us on our last podcast that the trend now is faster loops with the customer, more deployed engineers, whatever you want to call it.
Ric Lewis
>> Correct. So what do we have to do to solve that? We had to embed AI technology in the processor to be able to do inline fraud detection and that forced us to get aggressive about AI before the whole world was talking about it. Well, it worked out great for us and I see that as we go. They don't invent for you the clients, but if you listen really closely, they'll tell you kind of what they really need and it forces you to get to that next place.
Dave Vellante
>> And they just laid out the value prop and the economics. I mean, you can now compress this at the point of transaction.
Ric Lewis
>> Correct. Correct. You stop the transaction, so there's no-
Dave Vellante
>> So that's money in their pocket.
Ric Lewis
>> Correct.
Dave Vellante
>> The insurers who are covering it.
Ric Lewis
>> Correct. Hundreds of millions of dollars back to our base.
John Furrier
>> Well, there's billions of dollars being spent, but there's trillions in, whether it's GDP or productivity, a lot of headroom. Ric, thanks so much. I guess my final question is quickly lay out the ARM relationship. What does that mean? Do an explainer video real quick on what is the ARM IBM Z relationship all about? What does it mean?
Ric Lewis
>> Once again, back to our clients first, they want to be able to put more workloads inside of a Z platform due to its fantastic security, reliability, the AI that's already built in and there's a whole ARM ecosystem of apps and capabilities that you don't have to port, you can bring those to the platform. And so making that as seamless as we can is just another win for the clients.
John Furrier
>> Demand for workloads. All right, thanks so much for-
Ric Lewis
>> Thank you....
John Furrier
>> having us on the camera.
Ric Lewis
>> Appreciate it.
John Furrier
>> I'm John Furrier, Dave Vellante, Cube on location in Boston for IBM Think '26, where the gaps are being closed. AI wave is continuing to accelerate the infrastructure demand is at an all time high. And of course that's going to feed into the AI era and of course sovereignty and behind that's quantum and physical AI. So much tech action happening. Of course, we're doing our job to bring that to you. Thanks for watching.
>> Hello, I'm John Furrier, host of theCUBE here in Boston on location, IBM Think with Dave Vellante, co-host. We're the co-founders of theCUBE. Covering, Dave 17 years, our 17th year covering the industry with theCUBE and of course IBM. A lot of changes happened. We got Ric Lewis, the senior vice president here of the infrastructure division at IBM. Ric, great to see you again. Thanks for coming on theCUBE and the IBM hosted studio. Appreciate it.
Ric Lewis
>> Happy to do it. Great to see you guys again.
John Furrier
>> We were chatting out in the hallway, almost like a historical view of the industry talking about the different waves. Now the waves are coming together probably bigger than any wave in the past combined. The infrastructure is the hottest area right now. You look at all the AI build out. It's accelerating the value for agentic. Hybrid has now become the standard. You start to see visibility into the global scale. What's it like for you these days at IBM? Because you're leading the charge, you got Z, which has evolved and continuing to grow by the way. So that's been phenomenal. What's your take right now on the market as AI infrastructure and what's IBM's position there?
Ric Lewis
>> It couldn't be more fun right now. The industry is as exciting as I've ever seen it in my entire career. Every boardroom is having a conversation around what are you doing on AI? What's the plan? What's the infrastructure that underlies that? So the conversations are rapid and you talked about waves. They're big waves and they're coming in big sets quickly. And the conversation has changed radically in just the last 18 months with agentic and how quickly that's changing everything. In infrastructure, we're seeing tremendous momentum and have been for quite some time and that makes it a lot of fun. A lot of decisions to make, but a lot of fun.
John Furrier
>> You got a lot of infrastructure products under your purview. One that we've been fascinated with has been, not the rise, but the continuation and the growth of Z, the Linux support, a lot more workloads coming on Z. You got a lot of real estate on the board and you have a dual architecture with Arm you guys announced on April 2nd. What was the motivation? What changed? Why did that come about?
Ric Lewis
>> You hit it. It's the momentum. So Z, I don't think most people would guess this, but we've been growing 20 to 30% program to program on Z for almost a decade now. And that means people are bringing on new workloads. The AI technology that we've put into Z is really resonating with clients. We have AI built on the chip. We have AI cards that we plug in the systems that we started shipping last year. All of that is resonating. So the conversation we're constantly having with clients is we'd like to bring more workloads to the platform, but we don't want to have them just be the ones that run on Z OS, et cetera. We have LinuxONE systems, but we have clients that are like, couldn't we just bring some of our ARM ecosystem apps that are surround our apps and put those on there? It would have the same security profile. And as long as it's performant, we'd love to just put those on there. So we're making that a reality and designing it.
John Furrier
>> What was the challenge there? Because actually I mentioned a lot of real estate when I saw Z, it's got a lot of, it's not as dense as say it's like super computers. What was some of the challenges with ARM to make that work, if any? And where are you on the curve from marquee customer to full production? Get us through some of that adoption.
Ric Lewis
>> Yeah. So we haven't shipped that capability yet. We have shipped our AI capability. This will be in the next generation of Z systems. Key challenges really always is about making sure that you do it in a performant way. So we haven't announced exactly how we're doing that, but I'll give you some hints. We're going to talk about that a little bit later this year, but it's not going to be emulation. It's not just simple software kind of stuff. We're doing hardware to make sure that we execute in a performant way and have that performance right up to the standard that a client would expect from a Z system.
Dave Vellante
>> So okay, you can't tell us exactly what it is, but maybe what it's not. It's not replacing Z and Power.
Ric Lewis
>> Absolutely not.
Dave Vellante
>> I don't think you're co-designing the silicon necessarily?
Ric Lewis
>> No, no, no. We're integrating ARM technology directly natively into our chips is what's essentially going. So it'll still run all the same Z apps the same way it does. It'll just be able to execute ARM as well. So we'll-
Dave Vellante
>> So expand and give more details on how we do that. A lot of technologies that we developed over the years to be able to-
Ric Lewis
>> So it expands the market. It opens up the ecosystem.
Dave Vellante
>> It's not Graviton.
Ric Lewis
>> Correct.
Dave Vellante
>> It's not a new chip, but it's making ARM a first party citizen inside of it.
Ric Lewis
>> It'll be part of our next generation Z system. So just like as we come out with Z15, 16, 17, we're really creative on those names, you can tell.
John Furrier
>> Z18.
Ric Lewis
>> It'll have this capability.
Dave Vellante
>> What does that do for the economics?
Ric Lewis
>> Just helps. We've already taken a huge step on economics. You would be surprised to know. So we have our Z normal transaction running our database, CICS, IMS, Db2, all of the traditional apps that run on there. But we sell mainframes by the MIPS, by how many instructions are run on those platforms. We've had 3X the MIPS growth in Linux transactions versus those core Z transactions. So people have already been doing app stacking and moving Linux apps onto this. That already was a huge leap in terms of ecosystem and cost, making it simple for them. This will be yet another leap because you don't have to make it to run ZLinux. You actually can use RMAPs. So excited.
John Furrier
>> I just put out a post last night on call Hyperscaler 3.0. It's set in the frame for what's coming around the corner. I want to get your thoughts and your vision on this because Hyperscale 1.0 was AWS, Azure out there and got Google. 2.0 includes the neocloud and you see a lot of action on neocloud's huge demand for GPUs and the infrastructure. 3.0 would be many suppliers across a portfolio of services fully distributed, but the neocloud is getting a lot of attention. How does IBM infrastructure fit in this wave of hyperscale 2.0, which is emerging today and then 3.0, which is fully distributed computing hybrid?
Ric Lewis
>> Yeah. I mean, you said the keyword right at the end of that question, hybrid. I mean, we went from hybrid during phase one of the clouds that you talked about there was almost like, no, no, there'll still be something left. That's not the case anymore. The whole, especially AI has made hybrid not an option but an absolute inevitability. And there's a reason for that. That's because of the data. AI is all about data. If you have bad data, you have bad AI. That data is not in one spot. I think there are players who've said, "Oh, it'll be in one spot." But you talked about it. It's going to be in neocloud, it's going to be big clouds over here. It's going to be on prem, it's going to be everywhere. And if you want to get value from that AI, you have to get at that data. And so that's how we view the evolution in this next wave is it's hybrid. It's hybrid everywhere. And so you got to make sure that your systems make it easy to pipe that data. You see a bunch of announcements from us on that. I have them in storage, making it super easy to get at that data. But you see also bringing AI capability to the data instead of requiring people to take data somewhere to get AI capability on. That's why we built AI into both Z and Power. That's why we built fusion and storage to make sure that data path is simple and allows you to run AI.
John Furrier
>> Heterogeneous has always been the market. And we were talking before we came on camera about how those one chip, x86, now you got GPU, all kinds of processing capability. But today the demand is so high for infrastructure for all kinds of compute at many levels. How does that impact your view of the enterprise? Because in this 2.0 phase, before we get to three, the enterprises are now adopting code, automating code building aid that's going to pave the way for agentic, which is data driven.
Ric Lewis
>> Absolutely.
John Furrier
>> So the infrastructure has to fill the void and that piece of the market is the enterprise. Neoclouds will flower up. We think that's going to be, I think we would agree on that, but that's going to impact the potential consumption for the enterprises. How do you view that?
Ric Lewis
>> I view it as it's going to be a heterogeneous environment. I mean, we've had this conversation of at one time in the industry, everybody said there's going to be one processor and that's all there's going to be. And all the graphics and everything's going to move on to that. Well, we've seen how that's happened. There are specialized processors, RM processors, X86 processors, there's Z processors, there's Power process. It's all about specialization and doing it right. I see that the same in AI infrastructure. Inferencing is not a one size fits all problem. Training, lots of investment going into big training right now. But over time, more of that investment's going to go into tuning, fine-tuning and really inferencing. And that inferencing needs to be optimized for the specific data problem for the data where it is. And so that's how I see it. And we're working really hard to make sure that we're ready for that eventuality.
Dave Vellante
>> So thinking about that heterogeneous environment, you're in the many nines business.
Ric Lewis
>> Nine-ish. Exactly. So 99.9999, whatever.
Dave Vellante
>> I think about my RACF days. I mean, that's the most secure environment, but as the environment gets more heterogeneous and you bring in things like ARM, how do you maintain that level of security, many nines, uptime, et cetera?
Ric Lewis
>> It's core to our value proposition so you know we spend a lot of time thinking about that. And it's one of the most attractive things about a Z platform. We have people doing app stacking of apps that were on x86 processor just because it's one footprint that they can update from a security point of view and they know everything's rock solid. So keeping that security that I just talked about, but keeping that reliability and keeping everything up all the time is more appealing in this era because now as you go to agentic, you can't have things going down. It's doing work that people were doing and you got to make sure that it stays up and keeps working like that. So we spend a lot of time, like when we add new capabilities and new features, we make sure they're in the same paradigm, the same testing, the same everything that we've been doing to deliver those nine nines of uptime. And we do that now with the AI processor that's integrated in that. We make sure that it's tolerant to errors and things that happen and it goes all the way up through our stack, not just in the hardware, but through firmware to OS to the apps that sit on top of that in the platform.
Dave Vellante
>> Arvin in his keynote today talked about the gap between those that are really driving hard in AI and those that are somewhat lagging or directly significantly lagging. Why that lag? How can you help them compress that time to value?
Ric Lewis
>> So usually what we're seeing is clients are dealing with a couple of tough choices and how to get after it. First, there's the thing we talked about, are they hybrid? Can they really get out all of their data and how is that data? Second, what is that data? We talk about it and we say that data can either be a gold mine or a landfill. And if you just take AI to your landfill, you're just going to dig a lot faster in a landfill. It's still not a gold mine. So you have to be really careful about that data architecture and make sure that you are ready, you have the right data. For example, we have this feature in our storage called Content Aware and it allows you to figure out where is the gold in this whole gold mine that's in there and how do I structure that and make sure it's ready for AI? So those two, so it's hybrid then to the data platform. And then the third is you got to scale this and to scale it, you need to use AI in all of these areas to help you scale because otherwise what we find with clients sometimes is their infrastructure is growing faster than they can figure out how to keep it up and reliable. So we have AI enabled control that we embed in the systems to make sure that they can achieve the reliability that they want to achieve and they can still get the security and all of those capabilities. So it's like the old days to build a new CPU chip, you needed the old chip to run the processing. Well, to build this new era of AI, we have to use a lot of AI technology built in so that we can build what we need for the next era of AI.
John Furrier
>> Yeah. And that's client zero that you guys have pioneered. I love that reference to the mainframe having that growth and it ties to the agent world. You can have fit the AI for the workload and that fits into the infrastructure. Dave and I spent a lot of time with our team trying to squint through where the pressure forces are and the infrastructure, mainly on demand. Two that jump out I want to get your reaction to is what Jim Kavanaugh and I talked about at the NYSE was this AI builder persona. It sits at the center of the C-suite, hyperscalers, and then all the data people and the organizations, but it's not just developers. It could be a C-suite person, a line of business. So the builder is going to put a lot of pressure on tokens, more compute. And the other one is sovereignty. With the global climate, people are digging in the telcos in these countries to keep the AI revenue in. So there's a lot of sovereignty conversations that's a little bit more business, but it does put pressure on the infrastructure to be responsive, performant, trusted. How do you view this rise of this demand for this AI builder role, which is going to power the agentic and then the sovereignty piece?
Ric Lewis
>> Yeah. So I'm going to go in reverse order. Sovereignty is our home court. I mean, if you think about it, we do 70% of the world's financial transaction. The most important data that travels around the world is in our Z systems. Those are sovereign systems. They're on prem, they're in environments, they're in banks, they're in financial institutions, they're in insurance. Sovereign is our home court and we know this space extremely well. This AI builder, what we have to do is provide a landscape, a place for those people to come in the sovereign environment and make sure that they have AI optimized cores to bring to that data kind of problem, that they have AI-ready data pipes that I talked about before and the control, the AI-enabled control so that they can keep that environment safe and secure for the application that the people are running. Builder is the new thing. I think there's a whole bunch of roles here. There's builder, there's enterprise architect. They're all going to have to collaborate to say, "What is it I'm trying to achieve? Is it the sovereignty? Is it the global distributed nature?" There's going to be AI in all of the spaces. There's plenty of room for players to operate and we're trying to make sure we can deliver across or as a service across that entire landscape.
John Furrier
>> Yeah. And we go to a lot of events and we see the slides up on stage on the big screen architecture, role of storage is becoming more prevalent in the architecture. Memory, we see what's going on in the supply chain. And you mentioned you guys do a lot of AI and certainly that's the client zero success story for IBM so props to the team. How have you changed your product approach with AI specifically on the infrastructure side to meet those demands?
Ric Lewis
>> Sure, sure. From development, meaning we're now all using, Bob, you heard Bob announced here for our development, both in terms of chip. A giant chip project is really a giant software project that then becomes hardware at the end. So we're using those kind of capabilities, but then a giant chip project has a ton of firmware and OS on top. So we're using software to develop those elements of the stack. We're using AI to help us with our software development in those areas of the stack, but it's not just there. It's also how we're doing our customer support. I own the TLS organization, which is our worldwide network of 13,00 people that support all these platforms. All of our call work is now automated with AI. You can do self debug of your systems if there's an issue of any kind using AI. So it's changed how we do that. It's changed how we manage our global supply chain. I own the global supply chain that's positioned around the world to build all these systems.
John Furrier
>> Get your NAND orders in early.
Ric Lewis
>> Exactly. Exactly. And it has helped us already to make sure that we're well positioned for these supply shocks and things that are going on. So it already has changed my group and my role as much as a lot of the other transformation that we've been doing for the last five years.
Dave Vellante
>> We've mentioned storage a few times. You mentioned NAND. We saw Sam right before the keynote, Sam Warner, he was at a spring at his step because like you say, storage is booming. It's amazing. I mean, six, nine months ago, everybody wanted all flash, all flash correct me I'm wrong, but you still have a hybrid architecture.
Ric Lewis
>> Absolutely. We do flash, we do spinning disc and we do tape, and our tape business is booming.
John Furrier
>> Well, I'm not surprised.
Ric Lewis
>> As you can imagine. As large format spinning-
John Furrier
>> Tape is dead, long lived tape.
Ric Lewis
>> Exactly. As large format spinning disks become astronomically expensive, people are saying, maybe I can put some of that on tape. So we have all three. The cool thing, again, back to your how is AI changing what we're doing, our new flash system has AI management built into it, first in the industry of its kind. There's been a revolution in our infrastructure group of just innovation and some risk taking and being ahead of the curve. We introduced AI in these platforms before there was a ChatGPT moment as we all think of it. So being on the cutting edge and innovating like crazy has turned infrastructure into not just a big profitable business for IBM, but a growth business again and it's something we're super proud of.
John Furrier
>> Arvin obviously can't do a keynote without mentioning Quantum. How is the Quantum Heron line impacting or what does that connect into Z power and ARM?
Ric Lewis
>> So over time we view the world, it's not like Quantum comes and replaces classical computing. It's yet another specialization. I think of it as almost like a co-process or it's a big honk and co-processor that does problems that are beyond imagination. We talked about-
John Furrier
>> It solves problems that compute can't get to.
Ric Lewis
>> That compute can't get to. So it's a different class of thing. Just like GPUs went off and did pixel math and that kind of thing better than CPUs could do. Quantum does a whole new category of things in chemistry and biology and those things that CPUs really can't do very well. So I see it as a whole new branch of computing, except you could still do pixel math with digital kind of the Bitmath that we were all doing. It just wasn't quite as good at it. You can't even do these problems that are in quantum. So it's a big emerging area for-
John Furrier
>> We're seeing an emergence of pre-quantum engineering being facilitated and accelerated by supercomputing. You have a lot of that.
Ric Lewis
>> Yes.
John Furrier
>> How do you see that extending? We're seeing GPUs being used. We're seeing large scale. How does your portfolio fit into that kind of precursor, the preamble or the pretext of that?
Ric Lewis
>> Regular computing, supercomputing, quantum will all exist in parallel. We talk about fit for purpose computing all the time in IBM and they'll be fit for purpose. There will be problems that are best solved and only solved in quantum. There'll be other problems that are best solved in your super computing hardware and there's other problems that'll be best solved in regular CPUs. There'll be better other problems that are transaction problems that are best solved in a Z mainframe system. It's all about fit for purpose and making sure you have the right tool for the job and we're making sure that we're playing in those spaces.
Dave Vellante
>> And you see that when you talk to the labs and you visit, I was just recently in Knoxville at Oak Ridge and you can see they've got the quantum that's coming and they've got the classical computing, they've got super computing and they've got all of the above because it's right fit for the right workload.
Ric Lewis
>> Yeah.
John Furrier
>> It's all true. Question on the enterprise. We're seeing, again, we hinted out earlier about the adoptions coming with the agents obviously and the global sovereignty piece. Where's the excitement in the enterprise? Obviously on prem is back, never left, but it's more relevant now with the data on prem. What's your vision on how the enterprises will interact with this hyperscale 3.0 environment managed services with the neoclouds and the hyperscalers, on prem services, manage their own token factories? How do you see the enterprise over the next few years consuming and deploying the tech?
Ric Lewis
>> My team gets tired of me hearing talking about how segmentation is the answer to a lot of your questions, a lot of your... And I see that the same way. This whole distributed environment is about bringing the right capability to the right data for the right workload. And I just see a proliferation of that specialization, if you will, that I just talked about, the fit for purpose. But you could hear that and think what you mean is a whole bunch of new islands and silos and things like that. I actually don't see that. I just see specialization, but in the big picture, it's still hybrid computing. We're still bringing containers with OpenShift to that entire environment. We're still bringing sovereign core to be able to manage those environments. You can see us gluing that hybrid cloud together. We talk about our strategy of AI, hybrid cloud, and quantum. Our job is to bring those three really strong technologies and make them fit seamlessly together and deliver that to clients around the globe on their most important problems, whether it's healthcare or their most important data, which is all the transactions of the world, all that stuff, that's where we play.
John Furrier
>> How would you say this market is right now? Because we're seeing a lot of co-design, roadmap sharing. I mean, hardware and software businesses would do that in the past. What's changed the most in this wave, in your opinion, in terms of how you go about your business? Is it joint development with customers? We're seeing a lot more co-design moving away from supplier relationship and also as a leader, how has this changed your job and your ability to execute?
Ric Lewis
>> Two things, two simple answers. One, speed. It's a lot faster now. You got to be just moving like crazy. Everything is accelerated. So that's a difference and that there's a whole bunch, there's a half hour talk we could do on that. The other is client intimacy if the world is evolving into specialization and solving specific problems is absolutely paramount. Some people have said, "Well, were you guys geniuses?" There for a while when we first introduced AI processing inside of Z before there was a ChatGPT and all of this wave of AI, people were like, "That seems like that might be a little crazy." Then we looked like geniuses.
John Furrier
>> It's a good bet.
Ric Lewis
>> And it was a good bet, but where did it come from? It actually came from clients. It didn't come from us, "Hey, I think we could put together these widgets and make this happen." We had our financial clients coming to us saying, "Fraud is a billion-dollar industry for it's a major problem. We need a way to be able to do that." Fraud a few years ago is you get a note from your company seven days later that says, "Hey, we think there might have been a fraudulent transaction." Well, great. It's already happened. It was seven days ago. They needed that in line in the system.
John Furrier
>> We had the customer loops. Dave, we were telling us on our last podcast that the trend now is faster loops with the customer, more deployed engineers, whatever you want to call it.
Ric Lewis
>> Correct. So what do we have to do to solve that? We had to embed AI technology in the processor to be able to do inline fraud detection and that forced us to get aggressive about AI before the whole world was talking about it. Well, it worked out great for us and I see that as we go. They don't invent for you the clients, but if you listen really closely, they'll tell you kind of what they really need and it forces you to get to that next place.
Dave Vellante
>> And they just laid out the value prop and the economics. I mean, you can now compress this at the point of transaction.
Ric Lewis
>> Correct. Correct. You stop the transaction, so there's no-
Dave Vellante
>> So that's money in their pocket.
Ric Lewis
>> Correct.
Dave Vellante
>> The insurers who are covering it.
Ric Lewis
>> Correct. Hundreds of millions of dollars back to our base.
John Furrier
>> Well, there's billions of dollars being spent, but there's trillions in, whether it's GDP or productivity, a lot of headroom. Ric, thanks so much. I guess my final question is quickly lay out the ARM relationship. What does that mean? Do an explainer video real quick on what is the ARM IBM Z relationship all about? What does it mean?
Ric Lewis
>> Once again, back to our clients first, they want to be able to put more workloads inside of a Z platform due to its fantastic security, reliability, the AI that's already built in and there's a whole ARM ecosystem of apps and capabilities that you don't have to port, you can bring those to the platform. And so making that as seamless as we can is just another win for the clients.
John Furrier
>> Demand for workloads. All right, thanks so much for-
Ric Lewis
>> Thank you....
John Furrier
>> having us on the camera.
Ric Lewis
>> Appreciate it.
John Furrier
>> I'm John Furrier, Dave Vellante, Cube on location in Boston for IBM Think '26, where the gaps are being closed. AI wave is continuing to accelerate the infrastructure demand is at an all time high. And of course that's going to feed into the AI era and of course sovereignty and behind that's quantum and physical AI. So much tech action happening. Of course, we're doing our job to bring that to you. Thanks for watching.
John Furrier
Dave Vellante
