Well, welcome, everyone, again. Now it's on, I can hear. My name is Michal Katz, and I head the Investment of Corporate Banking division at Mizuho. I am joined by Rich Gallivan, who is our Chairman of the technology group at Mizuho. We are thrilled to welcome Arvind Krishna, who is the Chairman and CEO of IBM. As a business leader and technologist, Arvind has architected the building and expansion of new markets for IBM in artificial intelligence, cloud, quantum computing, and blockchain. Over his 30-year career at IBM, Arvind led a series of bold transformations, we'll talk about some of them, including, most recently, the acquisition of HashiCorp announced in April. Arvind is also a member of the board of directors of the Federal Reserve Bank of New York, and he serves on the board of directors of Northrop Grumman. Given his vast experience and expertise, this will undoubtedly be an insightful conversation, and we very much thank you for joining us today. With that, why don't we get started? I think it'd be an understatement to say that we live in a data-driven market that's constantly focused on every macroeconomic progress, trends, and obviously its implication on monetary policy, which coupled with the geopolitical tensions that we're having, and not to mention an upcoming election. Those have certainly created an uncertain backdrop broadly for investors, but I would say for many at large. Why don't we start maybe from your vantage point at IBM and perhaps given the front seat that you have to a lot of dialogue, both with clients as well as in the political circles, where do you think we are in the economic cycle? What do you think the implications or potential implications of either a new administration or a continuing administration on the economy? Yeah. Michal, thanks. It's a great question, and I spent the last 24 hours in D.C., some of those conversations were indeed happening. I'm glad you put the question as an economics question because I think that the economics is very different sometimes than a market reaction. Markets tend to want to forecast down the road, not just what's happening. I'm going to be sounding a little bit contrarian and a little bit much more optimistic than perhaps all of you are used to. Let's sort of maybe go around the globe, I'll reflect a little bit on our own experience, but actually, given the footprint of Mizuho, it'll be interesting. I'm really, really bullish on Japan. I think that what is happening in Japan is likely going to be, in many segments of the economy, maybe very high single digit, low double-digit growth. Why? Japan is deciding to move way forward. They're going to invest. They're going to invest in re-industrialization. They're going to invest in onshoring. They're investing in the defense complex. They want to digitize government. You would say, "Wait, haven't they done all those things?" The answer is not really, and those are all happening as we speak. I think many of those same flavors go on in Southeast Asia as people are worrying about diversification of supply chains and where do they put a footprint and where do they want to do business. India is probably the single biggest beneficiary of that and given that the politics looks very stable right now, I'll use that word, it looks like that 7, 8% is within reach. 6 is already there. That's a baseline. If 7, 8 is within reach, that's very different if you then look over a five-year picture. The Middle East is on fire. I think the people who are skeptical of what is happening in Saudi and the UAE will actually look back with chagrin at themselves. I think the economic changes to really build a non-oil and gas economy, not to say oil and gas is not important, through the Middle East is there. By the way, that's half the world's population that we've just gone through. I think Latin America, many people were worried with the politics about a year and a half ago when in late 2022, early 2023, I would say candidly, a lot of left-wing governments looked as if they were taking over, but they were quite clear. They need the economies to grow. They need to have business investing in those countries and we have found that it has remained remarkably robust. Maybe not as robust as the Middle East and Japan, but very robust. I don't discount 4% or 5% growth as bad growth. That's good GDP growth. You get into North America. I actually personally believe that we are going to get decent growth here. Maybe not 6, 7, 8, but 2 to 3 I think is reasonable. By the way, at actuals, you take all this and you say, "Hey, 3% growth with 2%, 3% inflation implies that actual you're getting 5 to 6." I'm an optimist when I look at this. I got to acknowledge the uncertainty you mentioned. By the way, I think the biggest uncertainty is unfortunately geopolitics. Something that the world thought was behind it. That's actually not true. There has always been skirmishes and wars. In the past, they were mostly impacting tiny bits of the GDP. The economies and the markets ignored it. Those are bigger now because those impacts could become far more contagious than they have been, and we've seen the impact of Ukraine on energy prices plus gas for Europe is significant. The impact of a Middle East conflict, if it goes bigger or wider, could also become significant in certain parts of the economy. Of course, what could happen in the South China Sea is a big question mark. I don't think it's imminent, but that's the biggest one. The demographics, I think, is an underplayed concept. The number of skilled people in the world is no longer growing, which is the first time in written history that that's the case. From 1990 to about 2020, we benefited from about 400 or 500 million people from China coming to the global workforce. That is now on the way down, not up. Japan is declining. Korea is declining. India has, I think, reached stability, but because of the age of people, another 10, 20 years, you'll still have more entering the workforce, and then it'll decline. Europe is in decline. North America is in decline. I know it doesn't feel like it, but North America is in decline in terms of skilled labor. This is going to be a question of how do we, in the middle of some inflation, I've got some interest rates. This is not 14% or 10%, but it is 5%, 6%. The demographics, how do you take care of it? Hence I get bullish on technology because technology, in effect, is a digital worker. To go back to Economics 101, capital is always "a worker." Capital and labor are substitutes for each other in a very, very crude economic framework. I would say technology is a bit wider than capital, but it becomes another substitute for actually both of those. Sorry, very long answer, Michal, to your question, but that's kind of how we see the world. You implicitly also got a sense of where we see investment capability for ourselves and where we see good returns going down the road. Love the optimism. I really do like the fact that there's the discussion about the bifurcated economy. It's great kind of doing the tour around the world. Given the fact that 50% of the population is going to the polls, if we just center on the U.S. economy for a second, how do you envision either a Biden or a Trump administration impacting the macro or, say, the economy? Look, the biggest question is going to be what happens with international trade relations and international taxes. I think those are the two fundamental questions. Internally, we'll get to questions on deficits and where are we going to spend money. I'm going to sound really unusual. I don't think the last two are really big questions. If you build up deficits too much, interest rates will go way up. If interest rates go up, within two years, this will self-correct. I think the U.S. system is so wonderful. Every two years, you get an election cycle. You have a balance of power, first, not just at the federal level between the House, the Senate, and the White House. You also have the states. You shoot interest rates up, and you make inflation go up. Whoever's in there causing all that is no longer going to be in a position to do what they do. I really do believe that in a very fundamental way. That means we are left with then what happens to international trade and what happens to taxes. We do know that at the end of 2025, certain tax cuts done in 2017 all go to expire. When I talk to people, by the way, the idea that those are just going to go away and leave people in chaos is completely wrong. People understand that those help make American business more competitive, and people understand that we want the U.S. economy to be robust, which means we need the companies to be in a robust position. I am not enough of a policymaker to predict to you exactly what form it'll take, but it's not going to be chaos and back to 2016. It is going to be something that allows us to go forward. By the way, countries all over the world are trying to figure out how do we tax more. Maybe they should also look at how they can become leaner in the government spending, would be my advice, but I admit that I come from the world of business, not government. We get down to, I think, the one uncertain issue, which is how much nationalism and how much protectionism is going to be there. I think that is probably the one big question mark that is there. By the way, that is not going to impact some bells. I think India, South Asia, U.S., I think, will remain fine in that. I think the bulk of Europe and North America will probably be fine. I actually think the north-south corridor here will probably be fine. Could there be some edges? Could there be some particular segments where there's some issues? Maybe. That only leaves one large block where there could be some issues. I don't think that that'll impact growth by more than about 1%. That's kind of where I land up. You asked where do I get the 1%? I'm not an economist. I'm an engineer and a business guy. Let me admit that. But from reading about international trade, I think David Ricardo is probably the easiest to kind of intuit. Okay, you have comparative advantage, if you don't want to take advantage of comparative advantage, every 10% of global trade is a 1% impact on your national GDP. I kind of go to, okay, if you have that much of an impact, maybe it'll impact by 1%. That's a one or two year, and then you kind of baseline it out, and you go forward. Just to kind of stay on one more topic as it relates to the broader, I would say, more regulation. Technology, as you talked about, particularly given the demographic issue, is going to continue to be core to how we work, live, socialize, engage, and what have you. At the same time, despite tech really being foundational to all that we do, has come at the crosshair of both politicians, regulators, and to some extent, perhaps some of the population that is being disrupted by it. IBM is a company that has been acquisitive. Can you just share with us some of your thoughts around the political dynamics and the regulatory environment as it relates to technology, not technology's ability to continue to partake in M&A deal-making. Yeah. Maybe I'll mention two aspects of regulation. The first, which we underestimate, is does regulation allow innovation and disruption and new economic business models to flourish, or does it put a damper on those? If I look at the U.S. versus Europe or even the U.S. versus the Far East, I would tell you that the U.S. has been far more embracing of, "Well, we do want the economy to grow." I think the single best example is probably the internet. Sure. In the mid-1990s, when the internet was coming up, the U.S. put out, whether it was Section 230, whether it was all around credit cards, let's allow e-commerce to flourish. As a consequence, what did we get? We got all of these companies that grew up here in the U.S., then all of the shareholders benefited, but a lot of the gain did come back into the U.S. GDP. Europe was much more careful. If you get worried about all the ills and you create really high guardrails, well, those effectively also act as a damper on the companies growing up and on innovation. I can take the pharma industry. You can take so many cases where that's the case. The first one I'll point to is, the conversation in D.C. and in the States, even on AI, does go, "Well, we are worried about it." But if my reaction is to over-regulate, I don't really not want to have all of the innovative companies growing up here. I think that that's a good balance, and as long as you walk down that balance, it's probably a good thing. I think that's actually far more important than just the M&A. Look, on M&A, it's clear. The FTC and the DOJ are being much more stringent. They're asking many more questions than they were four or five years ago. I think we go through waves on these topics. I'm not a lawyer, so we should probably ask, and I'm 100% sure this is an investment conference. Putting aside the theory, we're going to do that. What are they after? They don't want massive consolidation. They don't want somebody to accrue way too much power. The way we look at it is, look, in certain parts of the market, if you have 30, 40% share, we would probably stay away from doing anything that is M&A related there because it could cause a lot of concerns and questions. We would stay away from trying to do things which would cause a two, three-year series of questions. By the way, our track record is clean. We haven't had, I think, a long review in at least the last 10 years that I can remember. It is part of a selection process. We kind of go down what things fit our strategy. If they fit our strategy, we have to find areas of synergy, meaning is it the geography growth? Is it that when we combine products, one plus one equals three? It's got to be accretive. Absolutely, we will look at is it likely going to cause a question or a big delay with regulators or not? If it's not, we are much more likely to go forward. I won't put it where there's a question, we'll say an absolute no, but I'd say it's less likely. The other three have to be higher odds than to sort of cause that question to come up. That's how we think of it. By the way, I think there's plenty of opportunity, despite all the questions that they raise, all it means is that you got to stay away from some areas, but the others are still wide open. Well, Arvind, as bankers and investors in the audience, we all want to see more M&A. I can say that with certainty. You've talked about anticipating the shift. We are your friend then, because we do a fair amount. Absolutely. Wonderful client, as are we. It's a symbiotic relationship. You've talked a lot about anticipating the shift and the move from corporate clients experimenting with AI to deploying AI at scale. IBM has a comprehensive strategy around AI. You've got the watsonx platform. How do you think about this evolution in the current environment, and where are your customers in their life cycle in terms of AI deployments when they engage with IBM? Yeah. Rich, maybe I begin with some actual hard data, and then I'll go into more of my opinion and speculation. Great. We tend to poll our clients. I think the last time we polled was like 1,200 different clients. It has moved. 40%, I think if I remember, said that they are making what I would call reasonable progress. Maybe the word beginning to scale is appropriate as opposed to scale. Another 40% were still kind of kicking the tires. I call it experimenting like crazy. That's just poll data. I now get back to the questions that concern people. There is a lot of FOMO on AI out there. I got to get going. I actually advise our clients, and this is my opinion, it's like, please don't do that. Having 100 different projects is actually, I think, harmful, not helpful, because you're not giving them enough scale. You're not giving them enough resource. It does take a level of investment and a level of people, and by the way, process and culture change to take full advantage. The advice I give people is find your five. Five could be three, could be seven. Call it find five, get them to begin to scale. As soon as they begin to scale, now you can begin to add more. That is sort of my advice. Why do I say that? This is no longer going to be a maybe. As I look at all the areas where AI could help, is it on customer service? Is it on reading documents? Is it on summarizing documents? Is it on internal processes. Is it on quote to cash? Is it on accounts receivable? Is it going to be around digital labor by that example for programming? The answer is yes, it's going to be all of those. You've got to embrace it and go forward. By the way, if you don't change your process, so let's suppose you had a certain accounts receivable process and you now use AI to automate away half of that, if you don't change your process, then it's not very useful. There is no ROI. You've got to also begin to say, and when I say process, what is the risk going to be? How do I worry about all the aspects around, are you paying the right person? Do you have all the right controls? Those are going to change as they get based on AI. You've got to think of the process change along with it. Where do I think the world is? I think actually about 20% have understood that and are beginning to scale. I'll give you a great example. One of our clients in Western Europe is doing 800,000 calls a month using AI to go address them. By the way, 40% higher customer NPS at the end of it, not lower. That tells you that this is there. It's a big gulf for teams who have been measuring their people on how many calls per hour do you answer, to then say, "No, if it gets deflected completely, that's the much better answer." That's sort of process change that has to be there. I think where we are in the world, I kind of use a baseball analogy, Rich. AI is kind of in the first innings of the baseball game. You kind of got the players in the field, but you don't quite know how it's going to go, who's got momentum, we've got to begin to scale or kind of get your momentum going. Whereas cloud is more like in the fifth or sixth innings, where there is a lot of momentum. You kind of know what you have to get done. I think that's kind of the analogy I would use in addition to the numerical way of where we are. Following up on that, I know IBM has espoused a very strong message around your hybrid cloud strategy, and as a leader in hybrid cloud with an intentional strategy to remain cloud agnostic. I'm curious how you work with your clients to help them avoid vendor lock-in on the watsonx platform, while at the same time being able to utilize IBM as a successful deployment. Can you give us some real-world examples of some of your clients who are using your platforms and avoiding vendor lock-in and driving the expansion of the business? For sure. It's not so much vendor lock-in as making sure that people have flexibility. I'll say it's kind of fit for purpose. They can take advantage of capabilities wherever they exist, and don't have to feel that they're only on one. If you go and look at banking in Europe, there's a lot of regulators who are asking the question that, "Prove to me that you're not dependent on only one vendor." That's not a vendor lock-in question. That is much more a question of resilience. If you have to be resilient, then how do you operate on two infrastructures if you're picking one cloud? There are many who will say, "Oh, I'm going to do all my office productivity on this suite, but I want to put some of my web video traffic on this one, and I want to do data and AI on another one." Without taking any names, it's pretty clear kind of how that begins to play out. Every SaaS property is sitting on yet sometimes their own infrastructure and sometimes one of the 3, and people don't quite know where it is, and they should. It's as a service, it's up to your T's and C's to make sure that you're getting all of the capabilities that you need. Take a couple of real-life examples. Take somebody like Delta Air Lines, a great customer of ours. They have picked AWS to be their primary cloud provider. They leverage Red Hat and OpenShift as a layer with the idea that, well, if I need to maybe bring something back on premise, I could. I think they would prefer much rather that they don't have to do that, but it remains an option. If they have to take things to another cloud, then I call it's not going to be completely free, but probably 80%-90% of the work would be common as opposed to then bespoke to go and do that. I think that's a great example. They're all in on AWS, and I think as long as they have a great relationship, that'll carry on. I very much doubt it. I think keeping optionality if you're a business is really, really important. Take a large global banking client. They're clear they're going to use 2. They want to use a big private cloud premise because that fits their brand image, it fits their regulators, it fits the sovereign needs of the nations that they operate in. They also want to leverage a public cloud because it gives them speed and nimbleness. How do they leverage then our capabilities, be it Ansible or be it Red Hat or be it down the road, maybe HashiCorp, et cetera, to go achieve the best of both worlds are real examples of how that begins to work and how that begins to play out. In the U.S., you can often think through U.S. lens, but step out into a non-U.S. lens. I think every nation is going to turn around and say, "I need some things to be sovereign." What do they mean by sovereign? It doesn't mean by using that word, not nationalistic and it means that in case something happens, not that anything has ever happened, right? Nobody blew up a pipeline under the ocean. People wonder about if fibers get snipped under the ocean, I can still be able to operate. If geopolitics leads to people saying, "Hey, I don't necessarily want to give you technology," people still want to be able to operate. There is going to be sovereignty is going to become more and more important in this next 10-20 years of the worlds we're living in. People are going to say, "I need the ability to operate for my most critical infrastructure," 1. Two. There is going to be a homogeneity risk. If there is a lot of homogeneity, then there is a risk of cyber attacks, there is a risk of those infrastructures going down. People are going to say, "I need diversity." That, by the way, is behind the resilience question that the banking regulators are asking. Those are going to drive people into the solutions that I'm talking about. Those are really, really big. The last one we should talk about, because we talked about AI before it, is the data gravity in AI. It's expensive to move petabytes of information from one place to another. I don't think those costs are going to go down a whole lot. If it is going to be really expensive to move massive information, then you're going to want to do your AI training and your AI, kind of where that data is. Once you have a model, you could choose to run it in a different place, but there's a lot of competition that goes on in that first phase. That's going to happen where the data is. I think that these things are going to go together and people are going to say it's an and. It's not an or, meaning I need to only be in one thing in a hardened form. People are going to say, "I'm going to keep some things here. It fits my brand. I need these here. Sovereignty drives me here. I need to put this on a big public cloud because I get the variability in terms of cost and bandwidth," and so on. People are going to say as they always have, "I want the best possible answer." Sort of a point here on the deployment goes to the AI. People often say, "Are you going to have a big model or a small model?" By the way, the big models we're talking about get trained on trillions of terabytes of data. I'll say trillions of parameters. They have model sizes in the hundreds of billions. They do take a certain amount of infrastructure to then train and certainly to run. Then you have people with 10, 20 billion parameter models. People say, it's not a question of Those models, the very large ones, are good for some things. Anybody here ever moved homes? Well, most people here at least got from college to apartments to homes. You probably used an 18-wheeler to move homes. I'm guessing most of you don't drive an 18-wheeler every day. No. Just because you can do the thing, but you're not going to use an 18-wheeler to go to the grocery store or take your kids to school or maybe go on the family trip for a week in the summer. It kind of becomes fit for purpose. There are some things for which the really big models are really useful, and there are plenty of things for which the smaller models are going to be much more useful without a compromise on quality. Your smaller car is as fast as the 18-wheeler, but it can only carry four people. It's not going to carry 20 tons. You don't carry 20 tons every day. If you have no idea what you're doing, the really big model is useful because it might have a better sense of what to do. If you know that all I'm doing is summarizing English, a small model is great. If all I know is I want to understand Arabic, so I don't need the one model that can do Arabic and English and programming and haikus and give me a letter in the voice of John. That has a purpose, but that's not what most businesses want to do. I'm very much focused on the B2B world, so I think it's going to be an and. You would use the very big ones for some things, and those come at a cost to run and to operate. You're going to use, for the majority, once you know what you're doing, you're going to use something that is great, equal in quality, but is going to be maybe 1% of the cost to run. I think that's how this world is going to play out over the next 12 to 24 months. Arvind, that's super interesting. I just want to double-click here. We've talked a little bit about AI. We touched on hybrid cloud. How do you think about the interconnectedness of those two technology platforms effectively? Yeah. How does that inure to the benefit of IBM and your customers? Is one, having both accelerate or advance both platforms? How do you think about that? I think that these are going to be reinforcing to each other. Let's take a very large AI model. You're likely going to run that on the public cloud because very few people are going to stand up and say, "Look, I'm willing to put a couple of billion dollars into just buying infrastructure and running it." Okay? You might go forward and say, "I'm now going to write and either find or build an open source or go to somebody and get a much smaller model that I'm going to use for my regulatory compliance." There's a lot of people in here, my guess is few of you do worry about things like KYC and AML and filling out a regulatory compliance report to whichever is your regulator. After I reached 17 once, I stopped counting. There's a whole bunch of them. That probably you're going to run inside. You might train it on data that runs on a public cloud, and then you might want to run it inside because you don't really want the real data and the real questions going anywhere else. That's a perfect mix of a public and then private world. This is very much going to be the way I think that this is going to carry on. Many people say, "Oh, you're saying that's not" No, no. It's all an and world. I think the world of computing has always been an and world, and it kind of plays out that way. I have the best example. I talk to some of our teams who build models, and these developers come up to me and say, "Oh, but I'm building it on my MacBook." I kind of look at them and say, "Oh, yeah, 10, 20 billion parameter model, I can create it on my MacBook and do a run overnight. It runs well." That's not going to handle a room full of people. It doesn't have that much. For that one individual, it's great. They say, "I might run it in my enterprise." Okay, what do I need? That is why you can see the results of whether it's my friends at Dell or HP. Why they're selling GPU cards in servers, and those are taking off. Those are not being bought by the cloud folks. Those are being bought by enterprises and small businesses. You have the GPU infrastructure of the clouds itself going crazy because that's the big ones. I think they reinforce each other, and it's a spectrum that then builds on each other. As people get more comfortable, they'll tend to use all of these. You get a flywheel effect going. The reason why you have so much services work on AI is because the singular large models are not sufficient for most people. They're needing to tune things. They're needing to build their own skills. They're building them in-house. They're building them then on the different infrastructures. That, I think, is how this begins to get deployed. That is when the value comes. The value, by the way, is when deployment happens. Value doesn't come when you train things and when you start. That's super helpful. Help us understand the strategic rationale behind your most recent announcement, HashiCorp, the acquisition you announced a few weeks ago. How does that fit into the overall strategy at IBM? Yeah. We presume that our clients are hybrid. That means there are two or more infrastructures. Could be two or three public clouds, could be a public cloud and private, could be two public and private. I'll take an example. It's not maybe such a random example. They would love to use, let me take one example. They'd love to use Red Hat OpenShift, or they'd love to use MongoDB as a service across those infrastructures. How do you do that? If you let your teams do it differently in each one, then how you consume them, you'll have to get into all of the questions and differences, and we know there'll be subtle differences, and that just adds cost. If you say, "No, no. I'm going to use HashiCorp on each of those," that is what does the underlying infrastructure provisioning and gives you then a common way to do these services. Means suddenly everybody who consumes them doesn't need to worry where they are. HashiCorp becomes that glue that kind of isolates out and abstracts what you're doing at the infrastructure level from the service level. I think that's a great proposition. We then get into all of the patching, all of the security, all of the keys. How do you make sure that your keys are being renewed and not being left out there for eight years at a time? I'm not going to say why I said that, you should all think about that. The Vault product from HashiCorp is the best answer in the world to making sure that your secrets and keys are not kind of floating around in the wild and on spreadsheets. We know what happens when that happens. I think that those are great complementary adds to what we had with Red Hat as a portfolio into IBM. I think that this is something that's going to have a tremendous amount of traction with all of our clients to helping them manage infrastructure as code. There is a side impact on all of this. With the automation, that means you need fewer people. That means it can be done with tools as opposed to people. Relatedly, you talk a lot about partnerships and an ecosystem to both, that are essential for AI as well as hybrid cloud growth. Can you share a little bit how your partner ecosystem remains kind of a core part of that, and also how your core strength, both across software and consulting services, give you a unique advantage as you look toward positioning yourself in the market? Yeah. I'll begin with maybe a bit of a consulting lens and then move to a software lens. If I look at consulting, why do clients come to somebody? Because they lack the skills in-house, largely. If you look at that, then we have picked our now seven strategic partners. Amongst them, they're about 40% of our consulting business. You say, like when you say, "This is pretty straightforward." AWS, Azure, SAP, Oracle, Salesforce, Adobe, Palo Alto. As clients want to deploy these, what do we have to do? We have to build very deep skills. We've got to build assets. We've got to build methodologies. If you want to hire us to do work on those, then we can come in and reduce your risk of doing it, because we've done it many times before. We have the teams who are experts all around the globe. We can give you the optimized labor footprint of doing all that. The first part is a really critical one. We really have very, very deep expertise. We can do it at much lower risk and less time than if we took an alternate approach. It's not as simple as that. If you're in with SAP, may the client then trust us to also do application modernization. Sure. May they also trust us to do analytics. You get an explosion of projects going on. By the way, about 40% of our business now comes directly from those partners, and then you can imagine another 10, 20 comes from the expansion that comes from having those. We're quite selective. That's why I named seven only. On AI, on the other hand, we are saying we're picking an open strategy. We are going to build our own models. By the way, ours are out under an Apache license. People can do with them as they want. We also partner with others who have that same approach. We partner with Meta, we partner with Mistral, with Hugging Face, and there's others who are also going to give us the same licenses. We recently announced relationships with the government in Spain, in Saudi Arabia, in Japan, where we're sort of doing the same thing. That kind of opens up an ecosystem, much as people have discovered in open source. Open models, I think, is also going to be a similar ecosystem that helps get velocity and allow us to go forward. Of course, the ecosystem of services partners is critical to getting software deployed. We work with the TCSs, the Wipros, the Infosys, the Tech Mahindras, the Accentures, because that is kind of the world we play in. We find that thinking of it as a zero-sum game is the wrong approach. I kind of think of it as, let's see if we can work together and expand the pie. A smaller angle of a much bigger pie is a much bigger total piece of pie. I think of it that way. Can you have a win-win-win? Kind of goes to your and, not or. Correct. I'm watchful of time. There's one topic that I wanted to hit before we open it up to the audience, and that's quantum. Many of the audience heard about it, not sure exactly what that means. Maybe talk through what's IBM doing in that field, and what are some of the things that quantum will enable us to do that is not practical to do today? Quantum is a different kind of computing. I think everything we've done so far has been based on ones and zeros, it's in the digital world, all the way back to the late '40s to about now. Quantum is trying to take advantage of properties that let you do a different kind I'll use the word algorithm, and the purists in the quantum world will look at me and say, "It's not an algorithm." Okay, peace. It's circuits. Think algorithms. Can you solve things like factoring numbers? Can you figure out a better way for lithium-ion batteries to function so you could get 1,000 times the energy density? Could you reduce the energy required to fix nitrogen, AKA make fertilizers? Could you figure out a better material for carbon sequestration so that becomes much cheaper? If you think about those problems, they're outside the reach of today's biggest supercomputers. Not only are they outside, they're going to remain outside the reach because the problem space blows up if you live in a digital domain. If you go to a quantum domain, we know that the problems I just talked about will be addressed. The question is, when does this technology become usable enough and real enough? I would predict to you in the three to five-year timeframe, you're going to see massive. What's a way for investors to think about it? I'll bet you that 10 years ago, nobody thought, no one in this room thought that GPUs would beat CPUs in terms of market reach and market growth. Quantum is going to be similar. By the way, GPUs didn't replace CPUs, it was an addition. Quantum is going to be exactly similar. The same way as GPUs opened up a kind of computing that was hard to do or not cost-effective on CPUs, quantum will open up a set of computing that is not cost-effective to do on today's computing infrastructure. I think that's a way for you all to think about. I believe we have a couple of years advantage on others at this point. Some people will say it's more, but let's acknowledge it's a couple. There's a lot of smart people in the world. There's a lot of money that can chase a big topic. I'll say we have a couple of years advantage, we believe. We have systems on the 100 to 400 qubits that are up on our cloud. Anybody can go access them, so you don't have to take us at our word. We have built 70 actual quantum computers, and those have been released on our cloud for people to play around with. That's our track record. Others may have one or two, but they don't really keep them up and running. You need to have an actual computer. It's not enough to have something in a science lab and function. I'm excited about the opportunities. We think it's a massive market going down the road, and we're pleased with where we are in terms of progress. Fantastic. I think that there are mic runners and let me see if there are any questions to Arvind in the audience. Otherwise, Rich and I have probably 30 other questions. Any questions, anyone? All right. I see one. No, I thought I saw something back there. Okay. What do you think the first opportunities use cases are going to be for quantum? Financial risk. If instead of doing a Monte Carlo simulation once a week, you could do a pricing arbitrage model in a couple of seconds, may that be of some value to some people? One or two in the audience. I'm curious, the impact on AI that quantum will have. How do you see the relationship between the two? Look, I'm going to go out on a limb here. I'm one of those who believe that eventually that impact will be massive. Yes. I think in the near term, that impact will not be there. The reason is the size and the quality of quantum machines. If you take things that can live with some amount of noise in them, which is why I use Monte Carlo, it can live with noise in it. That'll be quicker. I think to get to real use on AI, you need much bigger machines. Much bigger, by definition, is going to be later, and they probably need to be a little bit more noise-free than noisy. The materials world can live with noise. I mean, like nature itself is noisy. Nature is not perfect. Simulation and risk are noisy inherently. I think those problems will come first to the fore, and I think AI will come, but later. What Rich is asking on AI, we know it takes massive amounts of energy to compute these models today. If quantum can do them in a few seconds, you're reducing that by three, four, five orders of magnitude. There's a massive advantage when we get there. That's the only caution I would put there. Arvind, you started the conversation by saying that you are an optimist or you're optimistic. What are some of the things that are keeping you up at night or some potential headwinds that we should be watchful as investors? Look, I think supply chain risks are real, so one should be very cautious about supply chain risk, because for many things in the world, because we have optimized and optimized and optimized, we got ourselves down to the place where there was only one supplier or only one place it comes from. That's a real risk. That does worry me. Now, I think that people do have their own economic interests at heart, so maybe that mitigates some of that risk. I think that's part of what we're seeing with people trying to build out diversified sources. The other one that worries me is always you get a backlash. If we're getting inflation and we're getting interest rates and people worry about being able to buy homes and so on, you get a political backlash. As an engineer, I tend to think of it in terms of over-damped versus under-damped systems. An over-damped system is an over-regulated. You get there, but it takes you a really long time. Under-damped means, unfortunately, you overreact. Instead of going towards some ideal answer, you shoot way over. You say, "Oh, shit, I shot way over." You come and you shoot way under. You kind of ring around the answer, and I worry about those backlashes. Oh, we need to raise taxes on everything in order to pay for all this. Oh, shit. That led to very low economic growth. Let's now go the other way. You get into this yin yang. Those two things kind of probably keep me up at night. The last is always talent. Can we get all the right talent? I'm happy. I'll touch hopefully there's wood underneath the leather, but we have the right talent right now. Talent is always going to be key in this world. Same as it is in all your worlds. Talent is going to be key to being able to keep making progress on what we do. Well, thank you for the insights, for the optimism and for, I really thought, balanced way of the market having a way to regulate itself and the view that it still is an efficient market. Greatly appreciate your time and thank you for joining us today. Thank you. Thanks. Great to see you.