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

Good afternoon, everyone. I'm Toni Sacconaghi, Bernstein's IT, hardware and electric vehicles analyst. And I'm super pleased to have Arvind Krishna, Chairman and CEO of IBM, with us today. Arvind has been at IBM for 30 years. He has architected the Red Hat deal and was appointed CEO in April 2020. And since he became CEO, he's proven that he's a change agent. He's really focused the company on revenue growth and free cash flow and really reoriented IBM to be a hybrid cloud and AI company. So we're super pleased to have you. Thank you for your participation in the SDC.

Toni, it's always a pleasure to be here with you.

So I'm going to lead the discussion. We do have question cards that we will be passing along. [ Temenos ], who just raised his hand, has these cards. He'll walk around the room, if you'd like one, and he'll retrieve cards for any questions that you may have. We try doing this electronically. It's been a massive failure. So we're going to go back to the old way and do questions on cue cards. But I'll plunge right in. And maybe we can just start very generally, like what is IBM today? What is kind of the value proposition or tag line to an enterprise?

Okay. So this started about 3 or 4 years ago. We began to see that there was going to be a massive adoption of technology. Lots of words have been used, digital transformation, technology to scale a business, et cetera. And the 2 transformative technologies were hybrid cloud and artificial intelligence. So what we've been doing over the last 3 years is really almost maniacally focused the company on those. So that led to spinning out Kyndryl because it wasn't opposed, but it was just not necessarily completely reinforcing that proposition, doubling down on both consulting and software and maintaining our infrastructure business. So as you look at the focus on hybrid cloud and AI, that is what we believe is the big opportunity. It's a $1 trillion addressable market the way that we are looking at this. And just by the way, I mean, IT, I think it's $3.6 trillion. When you sit in front of Toni, that's a little hesitant to say that because people like him will do those numbers. In that, about $2 trillion, I would say, is what is left once you take out consumer devices, networking, telecom. And in that 2, we play in that 1. So that's kind of where we are playing. So to shorthand it, you began by calling us hybrid cloud and AI. I think that's the correct shorthand and we are focusing on both consulting and software that helps our clients leverage those technologies to scale their businesses.

Okay. Now you mentioned AI, but you've recently taken a step to divest IBM's Watson Health business, which many associate with AI. The Watson tagline was really what underpinned AI. So help put that divestment step in the context of your strategy. And where does AI really fit in with hybrid cloud?

Yes. So 2 parts, I guess, to that question. So one, Watson Health divestment has got nothing to do with our commitment to AI and to the Watson brand. We remain fully committed to AI. And within that, we remain fully committed to the Watson brand. The Watson brand will be our carrier for AI. Two quick examples. We recently announced a deal with McDonald's. It's called Watson order taking, what will go into McDonald's drive-thrus to replace part of the tech in the drive-thrus. A second great example is Watson AIOps, which is our carrier for how you're going to apply AI to really improve IT operations. So just 2 quick examples of why we remain committed to that, so then why divest Watson Health? It's a question of verticals versus horizontals. So we believe that we are best positioned to take these technologies. We will always have an industry lens but through our consulting team. But we want to work on technologies that are horizontal across all industries. Verticals should belong to people who really have all of the domain expertise that they have credibility in that vertical. And so I think that health care companies, people in medical devices, they will have the credibility to carry out how AI is applied to health in depth. So that's why Watson Health. So we divested it. And should be -- should happen this month. And the new owners should be able to then leverage it because I think it will be best to grow stand-alone.

Arvind, was -- because the original strategy for Watson was horizontal, and then it became vertically focused. You brought Promontory in the financial space. You bought a couple of companies in the health care space, and the go-to-market was vertical. Was the decision to reverse that largely because it was difficult to gain traction? Or did you -- and so it was really a consequence of you tried and it didn't work? Or conceptually, do you feel like even it's just -- horizontal is kind of the way that we need to go?

So 2 different parts. One, if I look at it, I actually am completely convinced that AI as applied to health care, to financial services, to compliance, in that case, regulatory compliance, is going to be a massive market. Now that said, is going to be, I think it's a decade, maybe 2 decade-long journey to get there. And the people who get there, I think the world has shown, are those with massive depth in the domain area. So that would say that it wasn't so much an error of not going to get there, it's a question of right strategy, but it's going to take a lot more time and you need a different set of domain expertise to get there. To succeed in either of them, you're going to need doctors and nurse practitioners to speak to the buyers of Watson Health. That's not the IBM go-to-market field force. So there's a misalignment. Ditto in Promontory, you're going to need X regulators and accountants to go talk to people worrying about financial compliance. So that's a little bit different than us. On the other hand, there's 2 different answers in there. So Watson Health makes sense by itself. If I look at Promontory, it makes a lot of sense where we do work with banks on AML and on KYC because those are more general data problems. So sort of 2 different answers to those. So I think it's a question of the timing. So I think it will succeed. I think AI in health care, I believe, is a given, not even like a 99%. There's 100% going to happen. And part of that, Toni, actually does reflect on why I'm such a strong believer in AI applied to the enterprise and in all domains. I think you'd all agree that the world is going through a demographic shift, meaning there are fewer people with the skills than more people with the skills. And two, you have inflation and labor cost. You put those 2 together, and that will then argue for AI and automation in general being applied to more and more domains. I don't believe this trend is going to reverse in the next few decades. And so that is what is going to get all of these fully drawn out.

You talked about hybrid cloud, and I'm wondering -- we have a number of tech people in the group because I know them and I see them. We have some that are non-tech people. So maybe you can just sort of almost define hybrid versus multi-cloud. And ultimately, is IBM making a bet on hybrid that's differentiated or multi-cloud that's differentiated.

Yes. So I'm actually going to separate singular public cloud as opposed to multi-cloud from hybrid, and we can talk about multi bit on that. So if I look at the world, we believe that most of our clients are going to use multiple public clouds. And at least 70%, maybe 80% of them will also use something on private. Private, on-premise, same difference, I'm not going to debate that semantic. So why is that? If you look outside the United States, sovereignty on where data resides is a big driver of the reason for both private or sovereign cloud versus the big public clouds. In most countries, regulators are, for many industries, in financial services, in insurance, in telecom, in health care, in government, are going to insist upon multiple public clouds, not just a singular. So if we say that for about 80% of our clients, they need multiple public and on-premise, that's hybrid. Now to me, the difference if somebody says, well, maybe it's just 3 public clouds and there is 0 on-premise, it's the same difference. So I don't actually call that out as much as the answer is not a singular public cloud for all of our clients' names.

And in terms of IBM's acquisition of Red Hat, was that ultimately, in your view, an ability to help clients with multi-cloud or with hybrid cloud? Because really what OpenShift does is enable multi-cloud in a way that may be unique. So maybe you can speak to that.

Actually -- so the answer is yes, does it aim at hybrid or multi-cloud. So if you want to go -- the cloud providers are very good at doing what I will call cloud native, but that means you are locked into the APIs of one cloud and taking that application, which may compose of one, is not going to go over to the other. It's a complete rewrite. So the ability for applications that do need to reside on 2 or more clouds, then Red Hat gives us a great portfolio to allow people to do that. I always make fun, and Toni and I have talked about this before, it seems to come back every 20 years. 20 years ago, it was Java with right ones run anywhere. Now it's sort of categorization with the right ones deploy anywhere, I think, is the word as opposed to run anywhere. And so it gives us that. But I actually believe it goes further into the on-premise world as well because you don't then have to make the decision today or where you're going to run eventually. 90% of what you do, I'll acknowledge it's maybe not 100%, 90% of what you do, you can decide, I want to do my development maybe on a public cloud, I want to deploy it on private or it's been 2 years, I might get -- I changed my mind and say I want to deploy it on a different public cloud or for different country reasons or data locality, I might deploy it on one cloud in one country but another cloud in another country. So the Red Hat portfolio gives us all of that. And that is the reason it was so attractive for IBM. It gives our consulting team a great platform that they could take to clients with this value proposition. It led all of our software portfolio to get rewritten to run on OpenShift. So we can run in all of these cloud environments, both public and private.

I want to zoom out a little bit and talk more about your portfolio and what you're seeing. I think on the last -- I saw you about a month ago, you were at the World Economic Forum as well. And I think you did say I see something mild, I think, was the quote about careful evaluation of decision-making in Europe. And so I'd like your perspective on the health of enterprise spending globally and specifically what you're seeing in Europe.

Okay. So if I look at -- so let me first caution by saying we are much well aware of B2B spend as opposed to B2C. I would have no expertise on B2C, except what I read. On B2B, we are not really seeing a slowdown in demand yet. So if I look at the Americas, it's as strong as it was. I think the reason for this is back to demographics, labor inflation, technology offers a way through those problems as opposed to it is part of those problems. If I look in Europe, I don't see it yet. But we can't be naive. When energy prices are going up as much as we can see them going up in some countries, when there are possible food shortages, when they do have to deal with the refugees, then it's prudent to assume that they are going to spend time, energy and money on those topics in certain industries. If that happens, what we are seeing, and I mentioned not mild, I said I'm seeing that decision-making could become slower because there is caution about what is the forward-looking environment going to be. I'm probably one of the more optimistic people. I actually think that we have a lot of tools. We have full employment generally in a lot of places. So when that happens, I don't really see a downturn in GDP in the very near term. Now as interest rates keep ratcheting, if inflation keeps building, it doesn't slow down, then history tells us that we could see something. But I expect that in the bulk of the areas we operate in, it's going to be pretty shallow and pretty quick because it's just a response to financial tools. And there's a lot more financial tools that have been used by governments over the last now 15 years than were historically used. So history is a guy, but not a perfect guy to what could happen. I also believe, Toni, that technology will fare better than most other industries because technology has turned into what is a possible answer to many of these problems as opposed to being part of those problems.

And just to follow up, Arvind, are you -- you mentioned the U.S. and you mentioned Europe. You didn't address Asia Pacific and China, so maybe you can comment on demand there. And then just to be clear on Europe, you said you haven't seen it yet. Are you seeing any anecdotal evidence of more scrutiny or more careful consideration among deals? Or are you just extrapolating and giving the macro backdrop?

So let me answer your Asia question, you're right. I didn't -- we're seeing Japan's good, South Asia is good, ANZ is good. In China, I just have to be cautious. Given the emergence from the lockdown only this week, I think we have to begin to see what will happen there over the next few months. But that's the only place where there could be a bit of caution. That said, we haven't seen it yet, just to be candid. Yes, in consulting projects, even though that's a knit for us, in China, you see some because people were not allowed to leave the house. So it was hard for them to do some kind of consulting work. Some they could do. But that, I view, is not -- that's not secular. That's just temporal and it will go away. Now to your Europe question, no, I'm not really seeing it yet. I am extrapolating for the bulk of our business, which is Western Europe. Yes, we actually gave the size of our impact from Russia because we got out of Russia in -- right after the war broke. And that was $300 million worth of business for the year. So we were quantified about that. Eastern Europe, we're not yet seeing it, albeit the business there is not that big. And in Western, maybe near the end of the year, so that's pure extrapolation, not -- no hard data in the -- that we can see it.

Okay. So we've talked a little bit about this before, but I think it is beneficial to provide your perspective. So IBM was kind of growing, we think, about minus 2% per year pre-pandemic, prior to your becoming CEO. And we've spent a lot of time trying to make sure we took out acquisitions and divestitures and currency. And it was kind of about minus 2, a little better during mainframe cycles, maybe a little worse if the comps were tough. You're now talking about mid-single-digit growth. So that's a really significant change for a company that's $60-plus billion in revenue, was almost $90 billion in revenue. So maybe you can do the waterfall of what you think has kind of structurally changed for what appears to be kind of a 7-point delta in growth rate?

So I'm going to do it 2 ways. I'm going to do it a little bit by a, I'll call it, a financial buildup of the vectors that have changed. It's part of the way I think, but it's also part of the way at least a subset of the room will think. The second way I'm going to do it is by portfolio, and that's more the way that it gets acted on, albeit I'll accept that it shows up in the first set of vectors. So we divested Kyndryl, and Kyndryl was at the IBM level probably 1.5 points of headwind. So not having that, you would say that the growth rate improves by 1.5 points, part one. Part two, the minus 2 that Toni referenced, and I'm not going to debate that math because it's not meaningful, is that Red Hat is going to add about 1.5 points. So if you think of Red Hat today as being between $5 billion and $6 billion, even if you take 15% as a growth rate, that's $900 million. On $60 billion, that's 1.5 points. Three, we have really unleashed acquisitions, and this is acquisitions that fit in our cash flow. So that's why we talk about $3 billion to $4 billion a year. About 1/3 of them being in consulting, about 2/3 of them being in software. So then if you go do the math there, you can say, depending on a given year, 1 point to 1.5 points comes into revenue growth from acquisitions. I won't actually belabor the point that if you're buying stuff that is higher growing over time, actually, that's a virtuous cycle, so it actually is more, but let's just count 1.5 points. And fourth, we have done a number of big partnerships. So consulting for us has become a bit more focused on IBM than it should be. I'm not going to say exclusively because there always are partnerships with SAP and Oracle that were there before that are still there. But unleashing consulting to be great friends with AWS and Azure and Adobe and Salesforce, and we can see the growth there. Our bookings there are increasing 50% year-on-year, and those turn into revenue down the road. So that's about 1.5 points to IBM. It's much more than that. It's probably 4 to 5 points to consulting right there. So if you look at all of that, you could say that there's about 6 points in there, if you just do the straight math of the 7 that you're asking about, Toni. And then there is another point that is going to come from how we have simplified our go-to market, much more focus on the incentives for what behavior. I actually believe there's a lot more there in the medium term, not necessarily the short term. But for the short term, I'll sort of put 1 point there. The other way of looking at it is by portfolio. By doing all the actions of Kyndryl and Red Hat and the tuck-in acquisitions, 70% of IBM's portfolio is now in high-growth areas. So that's a little over 40% in software and about 30% in consulting. Infrastructure is going to be roughly flat. So if you look at those 2, they are both playing in markets that are close to 8% growth. So that's why you say if 70% is an 8% growth, then if you perform like the market, that's about 5% to 6% of growth. And do we still have work to do in different pieces? So some are doing better. Consulting is not doing 8%. It's 17% in the last quarter. We pointed at low double digits for this year, high single digits for the medium-term model. So that's a bit above where the market is there. Software, we pointed to mid-single digit in the short term. And as we improve, it could do better, but that's kind of where we are. So I did it by the vectors, so that's an easier way to look at it compared to history. And then we did it by portfolio, which kind of positions that, okay, we are in the right places if we grow at market rates. Does that help, Toni, to decompose that?

That's great. That's the way analysts like it, so appreciate that. You talked about the big partnerships. You mentioned 4, AWS, Azure, Adobe, CRM.

SAP, Oracle.

Okay. I was going to say, you talked about kind of the big 6. And if we think about those businesses, and I think you've mentioned some of this publicly before, so that's why I'm asking you. Maybe you can kind of dimension where they were at and where you think they can get to. And then what do we think about the margin implications of that? I would suspect that the margin profile might be lower on those kinds of deals, but over to you on that.

Yes. First, I don't think -- actually, if anything, I think the margin profile on that will be good for consulting, not lower. So now let's go through them. So the 2 that predated and have for a long time are SAP and Oracle. That said, they were probably flat-ish for the last few years in terms of growth. There was a small business on Salesforce, on CRM. I'd say it was good, but not growing very, very fast. And I'll put that on us, not on them. The other 3 are almost nonexistent 2 years ago. If I go back to the beginning of 2020, we had almost no business on AWS, Azure or Adobe. So if you look at our consulting acquisitions, which work like a catalyst. If you hire a few hundred people with AWS skills, you want thousands. We have over 10,000 certified. But those few hundred allow us to then retrain the others very quickly because they come with that knowledge. So in consulting acquisitions, once you pick your area, it allows you to do 4 years of acceleration into one. That's really what a consulting acquisition achieves. Very different from a technology acquisition where you put into your go-to-market channel and take it global. So having done those then, that is allowing us to go scale those. My aspiration is pretty clear. We want those to all become -- to be measured in billions. So whether -- is that $0.5 billion or is that $1.5 billion or is that $2 billion? That in the next few months and years will tell us. Now that is good for our partners because if we are working in them, their technology gets deployed a lot faster. So there is a win for the client. It gets done. It's a win for our partner because there's a market opportunity, and it's a win for us because we get the consulting work associated. No, the margins there are actually as good as in all the rest of the consulting work, Toni. We don't see because these are technologies in high demand. And back to my point on labor and inflation shortage, so we have the skills to get it done. You can command appropriate premium. No, you cannot extort by charging prices that are unreasonable, but you can charge appropriately for the market and for that business. So hopefully, that dimensions it. And that gives a sense of what we can do with those. I actually believe that we can grow multibillion-dollar businesses on many of those. I mean we have in the past with SAP, and I fully expect we'll do it with a large subset of these 6.

And Arvind, you talked about it there exclusively through a consulting lens. Ultimately, is there some pull-in in the software portfolio? Or conversely, you could say, look, if you're actually migrating more people to alternative platforms, could that accelerate the migration away from IBM's existing software platform? So if we look beyond the lens of consulting, how do we think about the implications of that?

I actually think that it's a flywheel effect. So if I take one of the software ISVs, not even for the moment, AWS or Azure, I'll come back to those. Invariably, there is work to be done around integration. That's a great place where technologies like MQ and our automation portfolio play a role. There is a great play for data integration from existing repositories into one of those. Those are half a decade or a decade-long technology deployment. So having our consulting footprint there then gives us a benefit. It's never exclusive. So I don't what ever want to make it sound like that.But it gives us a much higher win rate for the technology side of the portfolios that go in there compared to if you were not there at all. So it's a plus, not a minus. If I look at most of these, it's not who's going to deploy SAP. I mean it would have -- if it was going off some custom development in the past that might have been running on our software or hardware, that's going to happen regardless. So that's kind of baked into the numbers. Now we are there, we can instead say with RISE that maybe runs on our cloud. We can say with RISE, maybe it runs on Red Hat as opposed to alternate technologies. So you get a benefit compared to what might have happened. So I definitely see it going down that route more and more. And it really leads to a lot of areas because while the SaaS applications are extremely good at what they do, they don't do everything for the client. And so the ability to come in then with enterprise workflows based on AI for our consulting team becomes an additive opportunity. Now going to AWS and Azure, actually, it works both. We do a lot of work on their native APIs and services. We also do a lot of work on Red Hat and OpenShift on those clouds. But being there allows us to mix and match those as appropriate for a client, and that's a wonderful opportunity. We want consulting to play where the clients want to go. And if they want to go there, then we can help by saying, what is fit for purpose by pointing to the different options that are there for our client.

Right. Delving a little bit in the software, aggregate portfolio and you were talking about your businesses and their growth profile. So if we look in the software, about a $26 billion business. Transaction processing, about $8 billion. You've talked about kind of mid-to-high single-digit declines over time. That's your legacy mainframe software business. Red Hat, about $5 billion. You've talked about teens or high teens kind of growth rate. So those are pretty explicit. Most people know about those. But you then have roughly $12 billion, $13 billion. You're going to divest some of that with Watson Health, call it, $12 billion, which is non-Red Hat hybrid platform and solutions. What is in that? And what do you feel confident will grow in that? Because that's had an uneven set of growth over the last 5 years. So what's in it, where are the growth opportunities? And how can investors get confident that, that will grow.

So one from this past January, we are explicit that we will give you the growth rates for our software portfolio in 5 vectors. You mentioned 2 of them, Toni, Red Hat and mainframe a.k.a. transaction processing. I'd say it's maybe smaller than you mentioned, but not dramatically, okay? The 3 others, so that it's clear to all of you on where we -- and how we are performing as well as the opportunity are automation, and I'll come back and decompose it, data and AI and security. Okay. So all 3 of those, we believe, are massive growth opportunities. So if we look at automation, we might have made the most progress over the last 2 years on that one. Let me acknowledge since we were -- so let's call it flattish to slightly negative on these parts of the portfolio. What is the differences? And I would say that all 3 were probably in the same zone. And so automation has gone already into the mid to upper single digits over the last 2 quarters. I think it actually hit double-digit in one, but probably just above mid in the other. So let me call it mid to upper. What made the difference? There is a lot of core technologies in there around business process management, workflows, integration that people need. But if you're not innovating a lot, if you're not helping people take labor costs out, then they tend to look at you and not really buy a lot more. So what we did there with IT ops, what we did there with Watson Orchestrate, what we've done there with a couple of acquisitions like Turbonomic and Instana has what has powered that up. I would look at you and tell you, I think there is more opportunity there. But let's get it before I tell you to up the model from above mid in that one, okay? Security. I think that's a little bit of execution on us. Some quarters, we do double digits, like 10%. Some quarters, we fall flat, 0. Some quarters, we do the mid to upper single digits. The model there is going to be mid to upper single-digit growth. And the demand is there in that market. That's on us to go convert all of that demand into real outcome. The one way we need to improve, I'll be upfront, is on data and AI. And the reason for that is on AI, I mean, and Toni alluded this Watson Health, we were too focused for too long on trying to go after what I'll call black box AI or certain applications. We really need to help people, I'll call it, do much more enterprise AI, apply it to your HR processes, apply it to how you're getting work done inside the company, apply it to quote to cash. That's much less aspirational than health or wealth management or financials, et cetera. And as we kind of make that pivot and as we do some tuck-in acquisitions there, I expect to see that then get from the flat where it has been into the mid-single digits. So as you think about that mid-single digits, if we just do your $13 billion up, I mean, if you say it's 5, 6, you need to begin to grow at about $300 million a year. And we believe that's fully possible based on kind of what we can see in our capabilities. But we got to be focused as applying -- as opposed to trying to do everything, it really focused in on the enterprise AI use cases that I talk about.

And Arvind, we've talked a little bit about this. I don't want to go down too much of a rabbit hole here. But if I look at kind of reported cloud and cognitive PTI like pre-pandemic, it was like $8 billion, $9 billion. These are actually reported, and they include Kyndryl results. And this year, your software business, as you reported, which is essentially your cloud and cognitive business, well, PTI, that is close to $7 billion. So -- and if I go further back in time, it looks like PTI from the software group was even higher. So what has happened -- and the business has been roughly the same size. So is that a fair characterization? What has happened and why does that change going forward?

Look, unlike Wall Street, Toni, I don't debate the numbers. Because even if the exact number is slightly off, directionally, you're asking, look, it was higher before. It became lower. You have some improvement now. You didn't say that, but I'll say that. There is some improvement in '22, which I think even Toni will grant compared to '21. And what gives? Why did it come down those recovering? And why should it keep improving? Look, I think it came down mostly because you have a certain fixed cost for owning the company, and a lot of it will get now put on the back of software as opposed to in other places. And so that's the reason for it coming down. Also, by the way, I mean, like as we acknowledge, it was minus 2 revenue, you were losing revenue over the past decade. And so if you lose revenue in a software business, you tend to lose profit disproportionately.

But wouldn't -- isn't the margin profile of the transaction processing mainframe business generally higher than it is, particularly for an acquired new automation business?

Sure.

And so having you -- if TP is going down 7% a year, haven't you mix shifted into lower-margin businesses? And why does that dynamic not persist if TP is continuing to go down 7% and that's 40-plus percent operating margins.

Because software businesses at scale can easily operate in the low 30s or very high 20s of PTI margin. So yes, you're right. If you have a really mature software business on which you don't have to invest a lot in go-to-market and marketing, its margins are going to be much better. But as you point out, those tend to be declining. So you are actually declining absolute profit. So -- but if you're growing, if you're acquiring a business or innovating and growing a business, that once it gets to scale is going to be in the high 20s or low 30s, that is highly accretive for IBM as a whole. So it's not that I'm looking at that trade-off of one versus the other. I'm looking at, are we competitive in the overall software space? And in software to be competitive, I believe that the ones at scale need to be in the high 20s or low 30s. And so that's the goal to go get there. And at the math you do, as you can acknowledge and I'll acknowledge, we are a couple of points below that. So that is opportunity to be garnered over the next couple of years...

Red Hat, which was a profitable high-growth software company was still well below that high 20s when you acquired it.

Low 20s.

And again, if you think about transaction processing, if BMC and CA are indications kind of where it's at, that's 40%-plus margins, those are going down. Red Hat is growing at 20s kinds of margins. There clearly is some mix effect, right?

Oh, yes. But I believe that we'll get Red Hat also into the mid to upper 20s over time because you're right, it was at 20%, but that is related to a $3.5 billion business. As you get to $5.5 billion -- and as you know, we've been pretty hands off and kept Red Hat very neutral. But I don't believe that we will actually underinvest in technology, in direct sales and support. But over time, we will merge more of the back office. We don't need 2 cash desks. We don't need 2 treasuries. We don't need 2 payroll systems, just to give some examples. And as businesses scale, we know this, right? From looking around the industry, the aggregate in all those things put together is about 10 points. So that's -- you won't get all of that, but we will get some percentage of it. And that's how you'll see us raise it over time.

And just shifting gears to the consulting business. You talked about some of the drivers and the partnerships and how that's helped growth. You've now done 4 quarters of double-digit growth, but the comps get a lot tougher. So how do you see kind of the rest of the year? And more importantly, 2023 for consulting given it's obviously easiest to grow practices probably initially from low basis plus your comps are much tougher.

Yes. So 2022 first, we believe we'll have low double-digit growth in the year, all in. So given that the first quarter was in the mid-teens, you can see that we've got to have growth pretty close to that. You can't have just one quarter in the mid-teens and the rest kind of flat. It doesn't work. So we already stated that -- maybe I forget, I think maybe on the call, maybe -- there. So that gives you a sense of 2022. So that does say that maybe by the end of the year, it could get down into the high single digits or very low double digits. But maybe not high double digits. For 2023, what you said so far is that our midterm model for consulting is high single digit, right? Now I can tell you that when we look at the pipeline, when we look at the signings, our book-to-bill that is still running at 1.1for the full last trailing 12 months, so that is a signal of why we have confidence in the numbers that we are talking about, Toni.

And can you talk -- because this was surprising when we spoke about a month ago about -- I always think about consulting being a transactional business, kind of 6- or 9-month contract life, not a big backlog-driven business. Can you tell us what is the average contract? How big is your backlog in consulting, let's say, January 1 as a percentage of annual revenues?

So we don't talk about the backlog in consulting that much, but I will tell you, our average contract length is about 22 months. So you can think of that as 2 years. So it's not a transaction, meaning it's not an in-quarter or even in-year business. The moment you go to 24 months, Toni, basic math tells you that about half of your business at the beginning of the year and the other half we got to earn during the year, and that gives you then the base for the next year of the half. And so that to me is a good model because -- and that kind of tells you also why you're not on inflation within into us. It takes you about a year to kind of get new pricing through the system. So when inflation comes in, your labor cost goes up, but you're not getting it back in terms of price. But you do within a year, not multiple years, but about a year herein.

For 2023, after a year where you said you're going to be at the high end of your model, 6% growth, you have a mainframe cycle. The Red Hat accounting will be largely entirely behind you. Why should investors be confident that you can grow mid-single digits again in 2023?

So look, if you look first at consulting, and I say high single digit, let me round that up to let's call it, whatever, 8%, 9%. But then consulting is well over $20 billion. So that gives you a fair amount at the IBM level almost 3 points for IBM in growth. Infrastructure, yes, you're right. The mainframe cycle starts now this month. However, you would have only had a bit more than 2 quarters. So it's not going to be a full minus 1. So you can call that maybe flat going from '22 to '23, maybe a little bit worse than flat. But it's not a full headwind of the point because if you're a plus 1 and long term I call infrastructure flat, plus 1 means you should get a minus 1. But I don't think it will be the full minus 1 because I think that will be spread over '23 and '24. So it's not a full headwind. Then if you look at Red Hat, even if you slow Red Hat down into the lower teens, but at a $6 billion base, that's giving you another $700 million or $800 million, almost 1.5 points at the IBM level. So if we add up those, that's about 4.5. So we need to go get another 0.5 point from everything else. So that's -- by the way, just the new acquisitions will give you that. So that's why I have confidence in our mid-single-digit model.

Speaking of acquisitions, we met a month ago, you certainly sounded much more open about M&A, in part given software valuations have come down. So maybe you can talk a little bit about your M&A strategy? And are you -- particularly in light of valuations, are you amenable to a much larger deal?

First, our general philosophy is that you should try to live in your cash flow. That's a general philosophy. So if you have $11 billion of cash flow, you put 6 into the dividend, you put 1 into internal CapEx, that leaves you 4-ish that you can go do things. And we've been open about that model, so that's what you should expect from us on a going-forward basis. But cash flow does increase from year-to-year, so that aperture opens up in future years. You can also think about always borrowing a little bit from the next year or 2. And if you look at a 3-year picture, that gives you round numbers maybe $20 billion of flexibility. But I also hesitate because if you do a big thing and consume that and then you do nothing for 2 years, that's also not so good. But in the market where valuations have come down, though I believe it may be a few more months before the owners of assets maybe fully recognize that, then that opens up a possibility. Would we do something larger? It has to fit my criteria. The criteria are, it's not just a bolt-on for revenue's sake. It has to be somewhat synergistic with IBM. Does it drive more Red Hat? Does it drive more consulting? Does it drive more software? Can we help it grow faster than it did before, not just what was its growth rate? So the synergy has to be there. It has to fit in our strategy lanes, which we have been explaining for the last 45 minutes. It has to fit in there. It has to get cash flow-accretive relatively soon. It doesn't mean in the first month or the first 6 months, but let's call it 12 months or maybe just beyond it has to get cash flow accretive. So if it fits all those criteria and if the right strategic asset is there with the right financial profile and the right synergy profile for us, we would take a look at it. And there are many ways to then think about it. Is it just based on pure debt? Is it based on quick paydown of debt? Is it based on some mix of some equity vehicle and debt? It all depends on the details, Toni. But it has to be one that then meets all those criteria, not just because we want to go to cash flow accretive, that puts a good rigor on it. So not just revenue for revenue's sake.

Right? That's helpful. You did mention $20 billion. I think your debt capacity could probably give you $15 billion or $20 billion incremental plus you can borrow against future free cash flow. Is conceptually like $20 billion sort of the cutoff number in your head? Or you were just sort of saying that illustrative point?

Using that as an illustrative point. No, it's not a cut-off number. And the reason I opened up to other vehicles, you opened up one right now is I don't think there is a fixed number, but one should also be reasonably cautious, conservative and careful, I just sort of say that as opposed to opening it up to, okay, any kind of number as possible. Since there are some deals out in the market right now. So...

And accretion is -- so if you did a lot of it with low-cost debt, you're covering the cost of debt with cash flow, just to be perfectly clear on how you define accretion on free cash flow.

Yes.

Okay. Speaking of cash flow, you're guiding for $10 billion to $10.5 billion this year, $35 billion over the next 3 years. So that implies pretty significant free cash flow growth over the next 2 years to get to $35 billion. Why should investors be confident about that?

Okay. It's dependent upon the revenue. So we're seeing the revenue comes from software and from consulting. We're not seeing revenue comes from infrastructure. So if you look at software and if you look at its margins there, then even if you're in the mid-20s, that gives you about $0.5 billion of free cash flow growth from there. Consulting growth will give you about 1/4. So it's a mixture of 2/3, 1/3 in terms of cash flow growth. So it is very much on the mid-single-digit revenue growth. There is about $0.5 billion, I would turn on and tell you of, I'll call it, final cleanup from the Kyndryl spin that we're spending this year that we don't have to spend next year in '23 or '24. So about one of the total comes from that cleanup, the rest comes from the revenue growth.

And speaking of Kyndryl, I think you said when we last met that you don't have a purchase obligation in terms of minimum volume. And we know this year that, contractually, Kyndryl will purchase a couple -- $2.5 billion worth of IBM stuff, principally infrastructure software and associated services. Why should investors not be worried that ultimately, over time, that number could fall dramatically.

So $2 billion to $2.5 billion I'll just say that as opposed to $2.5 billion, it will be somewhere in there. As you said, it isn't a minimum or a maximum purchase commitment. So that's why there's a range. We said about 3.5 points for IBM for the full year in 2022. Look, we don't have minimum purchase commitments because in the end, it's based on their business. But a very large portion of this total is around mainframe, software and hardware. We don't really see that going away inside the Kyndryl base. And if it does, it will go somewhere else. So I think we'll keep that revenue. That's the reason why I'm so confident about the aggregate. Will there be some puts and takes? For sure. They might make partnerships with other people, but it's in the small. I don't think a mainframe, there's a choice in terms of which hardware and software they're going to use.

So you would expect sort of growth profile or declines in the case of mainframe software commensurate with the broader transaction processing business?

Right. In this case, I think it will be not so much of the broader transaction. It will be more flat even than the broader transaction processing because the broader transaction processing includes whether you have a loss of clients or capacity or substitutions. Those are less common in the outsource business base.

So Arvind, what -- as you stand back and you say, look, I've articulated a pretty ambitious change agenda. I have financial targets that certainly on paper, mathematically compute in terms of resources. What are the biggest risks? Because certainly, in hindsight, there have been growth strategies at IBM. And I think collectively, where they have fallen short has typically been -- the growth has not been fully incremental. It's either cannibalized the core or the core has fallen off at a more significant rate. And that's kind of undermined the aspiration for growth. So why is it different this time? And what do you see as kind of the biggest risk of core here 3 years from now, and you said, yes, we had a great '22, '23 was 2% instead of 4%, 5% and '24 is 1% or 2%? It's a really long question. I'll just turn it over to you.

I think you're almost answering the question, Toni. So the biggest risk is execution risk. And then you say, why will the execution fail as opposed to strategy or market? So by execution risk, I mean, we tend to invest in the wrong areas inside as opposed to where there is a market in software; or in consulting, if you run after contracts that don't have the appropriate margin, right? In consulting, we want to get to a low-teens margin as opposed to a single-digit margin. Because if you get into contracts that are lower margin, then it's going to be very tough to deliver high quality, and that will cause dissatisfaction of the client. In doing software, you've got to be very focused on where is there a return for our clients. The areas for me are clear. There's automation, there's AI, there's security. You've got to remain focused there because there's so much opportunity. So sort of going too wide is the biggest distraction risk we have in execution. So that is one that I am maniacally focused on to make sure that we don't spend our expense and our capital on areas beyond those. I would call that risk number one. Look, in our business, as always, technology is fast moving. So while I don't see it, is there going to be a better way to do automation than AI, okay? That's a risk. I don't know how to quantify that except by saying you got to keep an eye on the market and an eye out there. Always, you got to make sure you can maintain your talent. We're in a world where there is going to be a churn, a higher churn rate than before. What we have right now, we can live with. Will it get a lot worse? It's a risk, but not one that is in my top 2.

Right. Arvind, thank you for the time, and thank you for being at our conference. Appreciate it.

You're welcome, Toni. Thank you.