“Now Ben, to your question, how do we get confidence? Because our guide says we stay at that high 120-plus percent throughout 2026 overall.”

In context

“James Kavanaugh : Yes. The questions kind of go together, right, because we run this as a mainframe stack platform economic multiplier as we talk about. Every dollar we land on mainframe, we get $3 plus of software. Mainframe overall, yes, down 42%. It's off a 70% growth last year in the launch quarter overall. You look at the historical seasonality of mainframe, it's always down in the fifth quarter as we wrap around the launch. Now with that said, I think we were very open, transparent. It was below our expectations. But with that said, let me put this in perspective. One, we're still at nearly 130%, a record across the history of IBM of a program to program. Put that in dollar figures. From our most successful program, [ z16 to now z17 ] 5 quarters in, we're at nearly 130%. That is $1 billion more of mainframe revenue. That is $3-plus billion more of software stack on top of that versus where we were at a similar point in time in the z16 cycle. And I would tell you that's the value that we bring around innovation in mainframe unmatched in terms of reliability, security, resiliency, scalability, AI on the chip. That's why the mainframe handles over 70% of the world's transaction volumes in terms of value, 140 million MIPS in the marketplace that we run core mission-critical systems across every industry. 45 of the top 50 banks, 4 of the 5 top airlines, 7 of the top 10 retailers, over 70% of the Fortune 100 run on mainframe. Now Ben, to your question, how do we get confidence? Because our guide says we stay at that high 120-plus percent throughout 2026 overall. We have looked at key indicators, installed base, upgrade cycles, propensity models, but what drives demand and purchasing requirement in mainframe? I would boil it down to 3 areas: one, capacity workload. It's the most important determinant. 85% of the installed MIPS capacity out there in the marketplace today running all those core mission-critical workloads are either stable or growing. Clients are adding capacity and workload to mainframe, the viability. And by the way, that's coming in new AI workloads, analytics workloads, Linux-based workloads. And those MIPS are growing program to date over 15% to 20% installed capacity. So that's number one. Two, economic factors. We don't talk a lot about this, but I think it's important for our investors to understand things like TCO, total cost of ownership, consolidation, lease propensity. Depending on the size and complexity of the workload, we have anywhere from a 2 to a 15x TCO advantage running on the mainframe. That's why Arvind said in the prepared remarks, we do not see any evidence of clients migrating off mainframe. And lease propensity, which is a great indicator, we're actually in a very nice sweet spot as we enter second half from an upgrade. And then finally, AI-driven value. When you look at it, applications, data, security, all on the platform. We do $450 billion inferences per day at 1 millisecond with 8 9s availability. We've got clients that have already purchased over 50% of our Spyre inferencing. And those clients that have purchased that are growing MIPS capacity, the way we monetize value by over 3x faster than others. So then you get to TP. The reason I went through that on mainframe, how do you monetize the value of that 3x multiplier in TP. The more capacity that's in the marketplace, the more monetization opportunity, the more price opportunity, the more value creation opportunity. That performance, we have all the confidence in the world in Z in the second half. But when we look at TP and the models that we have done, given how much MIPS capacity is out there, the first thing that we should all remember, that is a great future indicator of monetization and revenue growth opportunity for us to go capitalize on that. The key question then becomes timing. And when you look at it, unlike the high-value innovation we bring in many of our data, automation, et cetera, the client has capability around their purchases of MLC, they could run on an OpEx model without doing a big purchase for 3 years for quite some time. But eventually, that technology upgrade cycle catches them and it moves in IBM's favor. And the economics, as you can imagine, the price differential moves in IBM's favor. So when you look at that, we have all the confidence in the world around a growth vector of TP. But when we look at the guidance in the second half, I think prudently, when we're looking at 6% to 8%, 6% being the anchor, 8% approaching double digits in the second half, we've got TP down low single digits to mid-single digits overall because we're going to look at that as a 2027 growth vector opportunity for us.”