Modern Mainframes, Al, and Quantum: Rethinking Enterprise Infrastructure

Every disruption cycle brings the same prediction: the mainframe is finally on its way out. Hybrid cloud was supposed to be the one that stuck. AI is up next. But the data says otherwise.

Daniel Newman, CEO and Chief Analyst at The Futurum Group, and Greg Lotko, SVP and GM of the Mainframe Software Division at Broadcom, bring Tom McPherson, General Manager of IBM Z and LinuxONE, onto The Main Scoop to explain why AI is accelerating mainframe investment rather than replacing it, and what "modern" actually means for a platform running the majority of the world's transaction volume.

McPherson lays out the technical case first: a 5.5 GHz microprocessor built on 5 nanometer technology, 35 billion encrypted transactions processed daily, and 450 billion AI inferences per day at sub-millisecond response times. From there, the conversation moves to where the real ROI shows up. McPherson points to IBM’s own cost-of-ownership research showing mission-critical applications run two to four times less expensive on mainframe than alternatives, and to real-time fraud detection for instant payments as a live example of AI inferencing built directly into transaction processing. Lotko and Newman press McPherson on lift-and-shift versus platform-native modernization, the growth of new talent entering the mainframe ecosystem, and why the industry needs a new word for how AI should relate to existing infrastructure.

Key Takeaways:

🔹 AI is functioning as a tailwind for mainframe adoption, not a threat to it. McPherson points to Code Assistant for Z, where clients using the tool are growing at twice the rate of those who aren't, as evidence that disrupting the platform from within is winning clients rather than losing them.
🔹 Platform-native modernization is outperforming lift-and-shift migration. Newman cites Futurum's research showing organizations that innovate on their existing mainframe platform reach ROI faster than those attempting to move workloads off it entirely.
🔹 The economics of the mainframe scale in the platform's favor. McPherson describes a non-linear cost curve where cost per transaction improves as volume increases, meaning the growth driven by hybrid cloud and AI adoption strengthens the platform's commercial position rather than straining it.
🔹 New talent is entering the mainframe ecosystem at a rate unseen a decade ago. McPherson cites IBM's Z community, 100,000 members strong across 160 countries, with 35% of participants new to the platform, while Lotko notes conference audiences showing 15 to 20% new-to-mainframe attendance compared to single digits ten years ago.
🔹 The language around AI and existing infrastructure needs to change. Rather than treating AI as something bolted onto mainframe operations, Greg Lotko frames the shift as one of infusion: AI built into workflows and processes rather than positioned alongside them.

Every technological shift that was supposed to make the mainframe obsolete has instead made it more central to how global commerce actually runs, and AI is shaping up to be the latest proof point.

Watch the full video at sixfivemedia.com, and be sure to subscribe to our YouTube channel so you never miss an episode.

‍Disclaimer: Six Five Media is for information and entertainment purposes only. Over the course of this video, we may discuss companies that are publicly traded, and we may reference their equity share prices. Nothing discussed during this webcast should be considered investment advice or a recommendation to buy or sell any security. We are not investment advisors, and you should not rely on this content as financial advice. Six Five Media collaborates with technology companies and industry leaders to produce research-driven interviews and multimedia programming for enterprise technology audiences.

Tom McPherson:
Sometimes people have the view that, hey, we use the mainframe, but it's not cutting edge or modern technology. And that's another myth that we're always busting. Today's mainframe, it's a five and a half gigahertz microprocessor, 35 billion encrypted transactions per day. The mainframe is a modern platform.

Greg Lotko:

I'm joined here with my co-host of the most, Daniel Newman. How are you doing, Daniel? Greg, it's good to be with you. All is good and all is going. We're going to broaden out with today's discussion, right? I mean, we're going to talk generally and overall about the IT landscape, right? And we look at this and we see continued innovation. We see customers focused with new challenges, new opportunities. So, you know, the opportunities, the ideas about the continued digital transformation and the innovation that customers are on using IT technology to advance their businesses. But we also have a couple of things coming in here that are causing disruption, right? You think about quantum computing and quantum encryption and what that's gonna do in the marketplace, as well as your favorite topic, AI and the disruption coming from AI. So what do you think, Daniel? There's a lot going on in the market, especially relationship to IT nowadays, right?

Daniel Newman: 

Well, listen, I mean, my favorite topic, at some point when we were doing this show and I would find a way to bring AI within the first 30 seconds of every episode and you'd pick on me a little bit about it, but I think the world has caught up. I think At some point we've come to the realization, you know, as we are staring down the barrel of looks like three different trillion dollar IPOs that are going to be basically rooted in the promise of AI. But at the heart of all of this, right, we are still facing a lot of modernization. We're facing a lot of what I call searching for receipts, meaning all this investment. People are trying to figure out how do we make money? How do we operationalize and create efficiencies and drive productivity in our organizations? How do we do it securely? How do we integrate this with all of the private data? You're hearing more and more CEOs coming out and saying, it's great that these models that have scraped the world's data are a lot of fun, and you can do all these things with them. But it's all that private data. It's all that really critical data that sits on your beloved mainframes, data that sits inside of data warehouses and data lakes that doesn't expose itself to these frontier models. And it's combining those things and basically building infrastructure in an organization that can actually utilize all those things to deliver value. So we are in this really exciting exponential time right now. But yeah, so I'm just going to say I look at it and I say,

Greg Lotko: 

Whether it's relative to modernization, whether it's relative to preparing for a post-quantum world, or the impact to AI. AI is certainly everywhere. And if it's not being used within your institution, if you're not using it technologically, even if it's not within your products, but it's helping you develop your products, it's certainly a topic of conversation everywhere. Whether it be across developers, whether it be across support people, whether it be across engineers, or even in boardrooms or from executives in corporations around the world. Not just those that are developing technology as a product, but those that are developing technology as part of their business to bring their product or service to market. So let me introduce our guests. In this episode, we're featuring Tom McPherson. So he's the general manager of IBM Z and Linux one. And, you know, we want to pull him into this overall conversation, talk a bit about his thoughts around each of these topics, but also share with us why he believes that the mainframe remains crucial to delivering enduring value to the customers that depend on this platform, not only now, but well into the future. So Tom, welcome to the show.

Tom McPherson: 

Thank you, Greg. Great to be here. Good to see you, Daniel.

Greg Lotko: 

All right, so tell us a little bit about yourself and the mainframe space in the area that you're leading today.

Tom McPherson: 

Sure. So Tom McPherson, based out of Poughkeepsie, New York, the home of the mainframe, as we call it. And my role is I'm responsible for the overall mainframe platform at IBM. That's covering strategy, engineering, product, ecosystem. And I've had a lot of great experiences through the years throughout the full Z technology stack.

Daniel Newman: 

So Tom, you heard Greg and I in our preamble there. We have our fun and friendly banter, but in the end, I think we've all come around to this transformation is exponential. It is changing so fast. And you made bets a year ago and gone all in, and you would have to probably change your strategy six to 10 times in just the last six months. We see model innovation happening at great speed. But I think you probably also heard what I talked about with data ecosystem and the provenance. And you're in that middle of that world, right, with with mainframe and the part of the business you run is so much of the critical enterprise and transaction data sits on your systems. Talk a little bit about what the modern mainframe means and kind of what is your business fit into all this? Because every time they try to say the mainframe is, you know, going away, it seems like there's a technological revolution that actually brings its prominence back to the surface. It makes it even more important than what, you know, despite any doubt that's been out there.

Tom McPherson: 

Yeah, it's well said. So from my view, I know on this podcast, we certainly know this, but I will always take the opportunity to point this out. When someone asks who uses mainframes, the answer is everyone, right? If you look across, it's the most innovative platform for mission critical workloads. Every time you tap your credit card, book a flight, take money out of an ATM, invest, get a prescription, you're using a mainframe. So we think of it as like the kind of the heartbeat of global commerce, where many of the world's transactions are running and the economic value of running the transactions on a mainframe is what keeps us winning. I would say the economics of high throughput, highly reliable transaction processing combined with the fact that we're always innovating on the platform is what keeps us really relevant in the game today.

Greg Lotko: 

I like where you started there, Tom. When I'm talking to people, you know, people in my family that have anything to do with IT or just, you know, somebody you meet on the street and they go, you know, mainframe, I don't really have anything to do with the mainframe. I tell them, look, pretty much any day of your life, unless you're sick, you stay in bed and you never turn on the TV or interact with anything technologically, you are hard pressed to go through a day in life without you invoking or having a mainframe somewhere involved. You know, if you make a cell phone call, if you order, if you're ill and you go online and you order medication, if you're dealing with shipping companies, a lot of the systems that you would interact with to stream your television shows or watch things live, there's so many things that we do in our everyday life that we don't even realize a mainframe is behind the scenes.

Tom McPherson: 

Yeah, and you know, the modern sometimes people have the view that, hey, we use the mainframe, but it's not cutting edge or modern technology. And that's another myth that we're always busting, right? If you look at today's mainframe, it's a five and a half gigahertz microprocessor in five nanometer technology, 35 billion encrypted transactions per day. 450 billion AI inferences per day with less than a millisecond response time. I mean, I can go on and on, but it's exciting, innovative, brand new technology. And the mainframe is a modern platform.

Greg Lotko: 

And people think about like X or Facebook or, you know, Snapchat or any of these social platforms. The amount of heavy lifting that the mainframe is doing in the world, the transaction throughput, it makes all of that pale in comparison combined. There's so much more going on on a daily basis.

Daniel Newman: 

It's interesting, Tom. You know, you mentioned about the number of transactions and the volume. But one of the conversations that keeps happening about modernization, and as you, you know, as you said, like this is modern, the technology is modern, it's modern process nodes, it's, it's, it's modern capabilities, it's modern throughput, it's modern, you know, so the architecture is all modern, but at the same time, In the end, I said something earlier about receipts, right? And so you guys have to be spending some effort, spending some time sort of determining that customers right now that are trying to lift and shift versus customers that are modernizing and going all in on platform. The ROI comes quicker, right? If they take advantage of the fact that it is a modern platform, it embraces the best of technology, they don't have to lift it, they stay within the secure realm. I just feel like there's this constant effort to keep trying to push the lift and shift. But I know what we're seeing on our research side is that the companies that are actually innovating on their existing platforms are doing better, they're getting there faster, and their ROI is improving.

Tom McPherson: 

Yeah, it's a good point. We've done some recent studies around the cost of ownership and what you look at in terms of transaction processing. We've done a detailed analysis of workloads. It's showing that mission critical applications are actually two to four X less expensive running on the mainframe. And it's because of the technology, it's because of the architecture, but it's really the balance of that. We have like this non-linear cost per transaction that improves as volume scales up. So all this organic growth that we have, from hybrid cloud, from AI and transactions actually helps the economics of the platform because of our commercial model. And that's helping us win also.

Greg Lotko: 

Right. It's kind of like the model of Costco or the local deli or 7-Eleven, right? 

If you're going to Costco and you've got a lot of things to buy, you get way more of a scale of efficiency and you fill up the back of that SUV. But if you just want to get a single stick of butter, That's not when you go to Costco. That's not when it makes sense. So it is the more you get, the higher you scale, the more you interact with it, the more value, the more efficiency it returns to you.

Tom McPherson: 

Yeah. Well said. Another point. So from my view of, of the mainframe, right. I had this window. I just returned to Z last August. And I was before that I was three and a half years in actually GM at the power platform. So people usually ask me, okay, you're coming back into Z, you haven't been here in almost four years, what do you see? And it goes to the industry points you were making before. Like the first thing I noticed was like the hybrid cloud adoption of the mainframe and kind of the concept of cloud only has gone away and is much diminished. And it's more about leveraging the mainframe in a hybrid cloud environment. So you see many clients that have gone through that journey. Then we're seeing the AI adoption, that's another thing that comes right away and it's because of the data and transactional gravity that we have on the platform and bringing AI to the data, it just makes the most economic sense, especially in transaction processing. workloads. And then we've seen a lot of growth in the community. So good indicators coming back. And it's good to come in with a fresh perspective and see this. And you can see it right away when I talk to clients and analysts.

Greg Lotko: 

Well, the other thing we've seen, and I know we've talked about it, we definitely see More confidence in the platform and using it with these other technologies instead of using other technologies instead of the other thing i know you and i have talked about it we've been out at some of the industry events around the world is. the amount of new to mainframe people that are in this space compared to not only 10 years ago, five years ago, but just within the last few years, I have this habit of when I'm on stage at these big industry events to say, hey, anybody in the audience who's less than five years in the mainframe? Please put your hand up. And that doesn't mean fresh out of university. It could be somebody who is doing something different but is now coming to the platform. And 10 years ago, that would be 1%, 2%, 3% if you're lucky. The last several events I've been in over the last couple of years, it's 15% and 20% of the audience, which is dramatic to see that infusion of people and new blood.

Tom McPherson: 

Yeah, that's great to see Greg. And I know we've worked on that together in the broader mainframe ecosystem, like we're all working on this problem together. One of the data points we like to call out is we have a Z community and there's 100,000 people engaged, of which 35% are new to Z and it's across 160 countries. So there's like, next level energy passion in the community. And I definitely noticed that when I came back in to Z last August.

Greg Lotko: 

And I don't want people to think, I mean obviously I just threw out the 15, 20%, you're talking about 35%. Sometimes people hear those numbers and they think it's all about the service providers and these big bodies that are groups of people that are doing engineering development. It is also actually in the shops that are running the machines around the world that are doing the development. the people that are actually going out to these conferences, whether they be virtual or in person, that it's also rising there, which is very healthy for the ecosystem.

Daniel Newman: 

Yeah, yeah. We talked a lot at the beginning about AI, and it wouldn't be me if I didn't bring this back to that. But IBM as a whole has been very much front and center and very focused on AI. I talked to Arvind very early on when it was basically all hybrid cloud AI, very committed to that. You invested big. into quantum now, that's a big focus. And of course the mainframe, I've watched the earnings. I mean, what a great performance that business has had in terms of contribution to margin. It's incredible. But AI is the top story. Every enterprise is trying to figure this out. You've got to have frameworks in place. You've got to have strategies in place, like just sitting where you sit, maybe through your lens, whether it's how you're thinking about it for your org, because you're kind of seeing it through two lines. You've got your org and you're kind of saying, where does AI fit in how my team works? And then you've got your customers that are building AI on top of the mainframe platform. As Greg had alluded to, it's an and not an or. Like, what are you sort of recommending? What's the advice you have for everyone out there about organizations leveraging AI?

Tom McPherson: 

Yeah, and I need to call out Daniel that I appreciate your insights on AI. I follow you and read most of the content you put out in your team, and it's really good insight on what's going on in the industry. The point that we're emphasizing in IBM is we're in the window now where winners and losers are being decided in the industry. And the main thing we recommend is start now, right? We need to move into a mode where you need to transform your business model, whether that's inside IBM or at our clients, to reimagine your business operations and workflows, and get really into like an AI plus mindset, as Rob Thomas says, instead of a plus AI mindset. And then you apply that to AI for business, and the power of bringing AI to the data and transactional gravity on the platform is really what we're seeing benefits from in terms of our business performance and clients are getting a lot of value in it. So one example we talk about is over the past few years with the transition to instant payments for banking, there's been an increase in fraud rates and new stricter regulations on processing these payments. So you've got the combination of increasing fraud rates, and you have to process the payments faster. That fits perfectly into our strength spot. So we have AI inferencing on platform. Financial institutions can implement real-time fraud detection with sub-millisecond response times designed at the scale of these high-volume transactions and kind of make progress through these new regulations without disrupting their business. And that's a use case we're seeing get adopted more and more so that that's a good example of how the clients are benefiting from the business value. Then of course we have the automation and AI ops and transforming the mainframe experience where we're investing a lot in And I would say culture kind of underpins this off all of it. And I know we all see that, right? It's not just about getting a new tool called AI. You really have to rethink the whole process.

Greg Lotko: 

Yeah, I don't see AI as a thing or one thing. I believe it's about AI infusion. So you think about even as a vendor in the space, how we're using and thinking about AI, whether it be AI frontier models, whether it be using it to expose inefficiencies or security exposures in the code, that's one way to use it infused, not as development plus, but involved in everything you're doing. code generation and allowing it to inspect, infusing it across our product portfolio versus introducing, oh, here's our AI tool. It's about integrating it across the processes, and then that's how customers see benefit. It means for our teams and for our customers' teams, not that they can do more with less, It's about making the developers, the engineers, the DBAs, the sysprogs, way more efficient, effective, so they can speed to time to resolution, but they can also get way more done in their day. So just as it's kind of mainframe and cloud technologies or hybrid IT environments. It's really our people, our technologies and the infusion of AI into that environment to drive value for us, the entire ecosystem and our customers.

Daniel Newman: 

What I'd like to do, because we have that audience out there and there are people, you know, we are alluding to the value of the mainframe, but you two both In particular, it's your day-to-day waking up and thinking about why it's critical that this platform continues to grow, why building the community is so important. Both companies have made significant investments to lead in these particular markets. I guess, just in your opinion, right? And you can take any of these different sort of market shifting narratives that are going on. You could take like, I'll use an adjacent, like the SaaSpocalypse, right? People are like, hey, AI is gonna replace all the business apps. But in the numbers, nothing like that's actually showing up. And there's some people probably that are saying the same thing about the mainframe. Like, yeah, it's not showing up in the numbers yet, but like with AI, all the lifted, it's gonna get so easy. And we're just gonna talk to Claude and it'll stand up all the infrastructure. It seems absurd, but there's a cohort of people that absolutely believe this stuff. So I guess it's just you two. I'd love to get your take on kind of why now is actually the time to almost go the opposite of those kinds of counter of those narratives and really it's invest more because this is going to be a really relevant and continuously relevant. It's going to prove itself to be what it's been in every other technological shift more, not less important.

Tom McPherson: 

Yeah, so the way I would respond to that is we absolutely see clients investing more in the platform. And just like hybrid cloud, when that disruption came through the market, it was viewed as, you know, this could be a problem for the mainframe. Is the mainframe going to play? And it actually lifted the business. And we see the same thing in AI, that it's a tailwind for mainframe if if you are in the right mindset as a mainframe team in a mainframe ecosystem which i certainly are to. Have the conviction to disrupt yourself and when you see these disruptions happening with a i. Don't be in a defensive posture, but play offense. So another statistic I like to call out is we started Code Assistant for Z three and a half years ago. I think it was the first G8. And that was an example of disrupting ourself, of putting out code conversion capability, leveraging AI in the market. When you look at today, the clients that are purchased and are using Code Assistant for Z are growing at twice the rate of clients that are not using it, right? So that's the mindset and that's like infused in the culture of the mainframe, whether you're in IBM or not in IBM, I think it's just the enduring platform culture of always innovating, always co-creating with clients. And not letting our clients down, where this is really the heartbeat of global commerce. So it's that combination of having the culture to disrupt ourselves, co-creating with clients, and always innovating on the cutting edge that puts us in a good spot.

Greg Lotko:

I hear two polar camps talking about AI, and then there's the middle, right? So I hear those that say AI is going to replace all my programmers and all my operations folks. Don't anybody go to school to be a programmer anymore. And then I hear those that go, eh, AI is just the latest tool. It ain't going to change anything. And then the middle is kind of saying, hey, if you're a programmer worth your salt, if you're an operations person worth your salt, you should be learning about AI to see how it can advantage you. And I was having a conversation with one of my folks. And as we're going back and forth, I felt they hit it spot on. You know we've seen this play out before and it's not just the idea of an ant, remember, oh god it was more than thirty five forty years ago. We had all the code generators like bachman etc where you just needed to define to it . It was gonna write your whole system for and it actually accomplished some really good things but what you found. Was there were people who didn't understand anything about programming and just understood business logic, that they were not the best to use these code generation tools. You wanted somebody who understood the technology. And if you look at where AI is going with the code generation, if you even look at using frontier AI models and how you use it to explore your code base, it's all about how well you write those prompts. how well you define the problem set that you're trying to solve, the universe, the data that you can work with, whether or not you can escape your private environment or go out to the internet. So whether or not somebody ends up talking about this as well, we're moving from programming to prompt engineers, It really is about how to use the tool and use it most effectively. And it does allow us to go to the next higher order of magnitude and accomplish more with the tools that we're built on. So while this is going to change the industry and it'll change what programming looks like, It doesn't mean, does not mean, that we're not going to need people who intimately understand technology and can think about how effectively and efficiently to use it, and then can guide an AI to deploy that and accomplish more for us, and then be able to inspect it afterwards to ensure the AI has done what we want it to do. That's where we're moving with the future with AI and those businesses that recognize that they're gonna be the most effective and efficient and the most successful.

Daniel Newman: 

And if i had to boil it down to one thing i think what we're really saying is it's and it continues to be. What we have learned and what we have done, and we are adding it is an end with AI and I like the, you know, what I've seen in building and implementation is it's AI plus, in a way and I know, Rob Thomas says it a little differently but the, the. Utilization of, say, codex or cloud code with the development of mainframes and the historic way that, you know, infrastructure has been stood up or software has been developed is where we're getting the most benefit. And those that understand how to do that, they actually understand full stack code and development are the ones that are able to take most advantage, you know, like. People like me, not computer science, not programmers, have been able to do some cool toys and POCs. But when you really need something to work at an enterprise level with security and guardrails and governance and compliance and data security, those that understood how to build apps are now going way faster. It's kind of that way across all the different knowledge areas where AI can be applied. Tom? Great to hear from you. Congratulations on the role and the move back over. Appreciate you joining us here on The Main Scoop. I'd love to have you back because I guarantee you in six months out, a year out, there's going to be more stories to tell. Thanks for joining us.

Greg Lotko: 

Great to be here. And it's always good to be with you, Tom, and with you, Dan. I'll give you one more thought, Daniel. One word I want to change. We have been talking about technologies like hybrid IT, cloud, and mainframe, mainframe and this. And I think we missed it a little bit. We need one more letter. We need a four-letter word, not a bad one, instead of the three-letter word of AMP. I believe, especially with AI, it's more about with It's about the infusion. The and makes you think bolted next to and just exchange back and forth or a sidecar. AI and a lot of these technologies and the way we look at processes, it really should be when, with. It should be infused. It should be incorporated. And that's where the future is.

Daniel Newman: 

Thank you all so much for joining us. We'll see you all very soon here on the next Main Scoop.

‍