INTERNATIONAL BUSINE

IBM
Temps Différé Nyse - 22:00:02 03/02/2023
136.94 USD +0.40%

Transcript : International Business Machines Corporation Presents at RBC Capital Markets Technology, Internet, Media and Telecommunications Conference 2022, Nov-16-2022 11:45 AM

16/11/2022 | 17:45

Presenter Speech
Unknown Attendee (Attendees)

Good afternoon. Please welcome President and CEO, Royal Bank of Canada, Dave McKay; and Chairman and CEO, IBM, Arvind Krishna.

Presenter Speech
David McKay (Analysts)

Right. Well, hopefully, everyone is enjoying lunch, and thanks so much for joining us Today, it's so good to be back in person and see everybody after doing this digitally for a couple of years. So great to be back for me in the city. And Arvind, I know you live here. You travel a lot. We talked about your travels. I am joined by the Chairman and CEO of IBM, Arvind Krishna. And Arvind, thanks so much for taking the time to sit down with us today. I just watched a great episode on CNBC where you told the story. We're going to have a little bit more time to get into the strategy and details of everything you've done, but congratulations on your first 2 years of CEO. You've achieved so much.

Presenter Speech
Arvind Krishna (Executives)

Dave, thank you. Look, it's always good to be here, as you said, with a crowd in person not just over a video link. But it's also always a lot of fun to talk to you and get your insights.

Question
David McKay (Analysts)

Yes. We'll be able to share experiences, particularly on the macro side. You've been with IBM 30 years, and leading up to CEO as you were, I think, Senior Vice President of cloud for sure and then cognitive -- computing cognitive services, software. They had a chance to kind of build a and curate under your leadership, some of the most important technologies from blockchain to AI to cloud, hybrid cloud, that today, as CEO, you're trying to marshal the organization forward to really execute on that.

So it's going to be a great conversation because you've led the path for IBM, and now in the seat of CEO, have a chance to bring all that to life. But maybe we should start with a macro context, because a lot going on in the world that impact your clients. Obviously, you get a chance -- you just got back from the Middle East, as we talked about just before we came on stage -- a chance to really get a feel for the world. Can you give us your view of the macro context and how it affects the consulting business, the hardware business and your customers in general?

Answer
Arvind Krishna (Executives)

Yes. Happy to do that, Dave. And look, for this audience, I'm going to sound remarkably optimistic. So let me preface it with that as opposed to you wondering. Because I would have begin with, here's all the doom and gloom. We have high interest rates. We have inflation. We have foreign exchange rates going crazy. We have supply chain issues. We have cyber issues. We have demographic issues. There's not enough skilled labor. I kind of say, wait, I mean, like how do you worry about all that?

But let's take the flip side. The flip side is in most nations, we have full employment or as high employment as we ever had. Household balance sheets are very, very healthy. So you've got to take those and you say, this is different. So there is a lot of economic uncertainty, but it is different as in there are going to be a lot of winners that come out of this. There are nations that are going to emerge winners. There are companies that are going to emerge as winners, and there are a lot of individuals who will come out as winners. But you have to sort of worry about all this.

So why am I such an optimist? Well, maybe it's because of my industry in tech. Because when you have high interest rates and you have demographic issues on labor, what's the -- I'll say maybe the only deflationary force is technology. So I see this as most companies are actually not trying to cut down a lot on technology spending. They're saying, how should I use technology to really come out stronger because the COVID period also taught us that you can actually deal with your clients remotely, not for everything, but for a lot of things. And if you can deal with them remotely, how can we come out stronger to fundamentally change our companies to scale better and serve even more people around the globe?

And that's why I'm maybe an optimist on this, Dave, and that's what I see. I mean you said the Middle East. There's almost no talk about inflation or interest rates there. Yes, they're subject to the same interest rate pressure as all of us. Maybe there isn't inflation that much because energy is a big driver of inflation, not the only, but a big driver. And this is -- they are big producers of energy. That piece is kind of -- makes them actually, I guess, have an advantage, not a disadvantage.

Question
David McKay (Analysts)

But they're trying to pivot their economy the same way Western democracies and Western companies are still trying to think through this. And to your point, there's so much liquidity in the system, consumer liquidity, business liquidity, that when we talk to CEOs and we participate in these various CEO roundtables around the world, CEOs feel that's a good shock absorber under the con of, yes, inflation is going to collapse demand, but this liquidity will help us exit and absorb the impact. And therefore, most CEOs are continuing to invest through this uncertainty, to your point, particularly on technology, to alleviate the shortage of workers, to pivot to new opportunities.

And one of the opportunities I wanted to ask you about, one of the themes that we hear from CEOs is I will not cut my climate pivot strategy. And therefore, as you see that as a constant source of investment from your partners and your clients, how do you see IBM taking advantage of that opportunity and that investment?

Answer
Arvind Krishna (Executives)

So Dave, it's a great point. One, I think until about 2, maybe 3 years ago, it was more and less action. That clearly seemed to have flipped. Whether it flipped in 2019 or 2020 or 2021 doesn't really matter. Today, it's something that's dead serious. So if I look upon it, most of our clients are saying, look, I need to have a correct base line. I need to know what is my real scope 1 and scope 2. And yes, I got it. The price is really going to be in scope 3, which is really the impact of their own products and services downstream. But you got to begin first by measuring yourself.

And to me, that's a data problem. How do you actually do that without data? How do you not connect to all of your systems? And so that's a place where we think we'll have an incredible advantage. That's what we kind of do for a living. We integrate data. We integrate across different systems, and we can give them that. We invested in it. We did some acquisitions. We've been growing our own technology. We stood up our own consulting teams. So we think of that as a big opportunity.

But then, Dave, I'll turn it around, and I kind of say a little bit tongue in cheek always. This is a place where purpose and profit go absolutely hand in hand. So if you look across Europe, energy prices are anywhere from double to triple, double roughly for the retail people, triple or more for businesses. So if we can come to you and say we could help you save maybe 30% of your energy bill, is that purpose, or is it profit? It's kind of both.

Question
David McKay (Analysts)

Both, together.

Answer
Arvind Krishna (Executives)

And I think that for most countries, other than a few who are lucky enough to be energy surplus, this is going to go hand in hand. And some of it is going to be driven by shortages on the supply side. Some of it is going to be driven by government policy. And so there's a huge opportunity to sort of make people much more efficient. I think, by the way, that 30% number, I would tell is one that most industries can go get done, including those in the oil and gas industry because you can say how about stopping methane leakage. That's one kind. How about turning off the lights when there's nobody there? How about reducing the cooling temperature in a hot country when nobody is in the office? Or how about reducing the heat for a cold country?

You can go on and on in all these elements. And so sensors, which are cheap, IoT technologies, data integration, AI, it sort of all comes together in a nice bundle and is a great opportunity, I think.

Question
David McKay (Analysts)

I think it's so exciting for you and for technology because we're in a world of optimization. And you just introduced a completely new variable into every optimization routine in every industry vertical. In the entire economy, we have to optimize for GHG output and minimize for it ultimately.

Answer
Arvind Krishna (Executives)

Right. Before we get to fully circular, which is the next step.

Question
David McKay (Analysts)

Until we get to fully circular. So computing power, data, AI are all really important components in helping society pick the right path and minimize and optimize the variables. So really exciting.

So as your first 2 years as CEO, as I said off the top, you had the chance to invest and curate and acquire Red Hat and kind of bring these technologies to the forefront of IBM capabilities. But now as CEO, you got to deliver them to clients. So can you talk a little bit about the journey of organizing IBM to face clients, curating these technologies, bring them to market, particularly hybrid cloud, obviously, in AI? You made a lot of progress. So maybe just talk about the journey -- the client-facing journey you've been on.

Answer
Arvind Krishna (Executives)

Yes, happy to do that. And I think many of you may not realize how big a change we've made over the last, let's call it, 3 years. So we brought Red Hat into the company. And for a company that has made his mark on 110 years and, call it, proprietary technologies, to have the largest open source technology in-house is a big change. So that's let's call it, 15,000, 20,000 people who come in around their organization.

We then took the managed services organization called Kyndryl and took 90,000 people out of the company. And I'll hazard to say that for the vast majority of our clients, they didn't really feel that we did that. So I give a lot of kudos to our team for getting both those pieces done. Now we got those done, and I kind of turn around and say, but our clients want to interact with us very differently in the past. The days of doing only a financial sale, as in walking in and talking about pricing and long-term contracts, is over. Everybody wants to get value. It all comes down to how do I get value quickly?

So our clients want to interact with us by saying, come show it to me in my environment. Don't tell me. Show me that this works. And so we have got to pivot the whole front of the house into being much more co-creators with our clients, demonstrating the value. And as consumption increases, then we get a return that goes along with that. And so as you do that, you got to change that side of the house. And I would say we have kind of done that as well.

As we do that, you got to work a lot better with partners. So do we compete with Microsoft? Or do we partner with Microsoft? How about if I tell you that our book of business with them is now over $1 billion on the backlog? And then if I look at Amazon, our book of business with them is even bigger than that. Or SAP, with whom we have worked really well in the past, or Salesforce or Adobe or ServiceNow.

So you can go on on this and say we've always had a lot of benefit for our clients by working in the ecosystem, which our clients use. And so big pivot on the portfolio. Red Hat is probably the best example of that. Big portfolio, deciding what not to do. So Kyndryl was that. A change in how we approach our clients. Change in really -- be a really good partner. And I call that sort of a win-win-win. It's got to be a win for the client. It's got to be a win for the partner. And it's got to be a win for us. And as long as we can do that, we'll keep going. And we sort of committed that with these partnerships, at least the top few, we'll get them well north of $1 billion each. And we're on our way there, at least half of them already, and the other half will get there.

Question
David McKay (Analysts)

Just as we're all leaders in this room, is there one part of the journey that was particularly hard? Because we've all tried to create reorganizational change. It never goes exactly as planned. There's always things that you learn about your employee base or your client base. Is there one thing that you can share or you feel comfortable sharing?

Answer
Arvind Krishna (Executives)

Sure, Dave. Look, these journeys are never a straight line. I can make them sound like a straight line, but they're always kind of 2 steps forward, 1 step back. I think the cultural shift of how we approach our clients is probably the toughest. And I will probably tell you we're probably only halfway through that, so it requires a lot of grit to make sure we get fully through that.

So why is it hard? So we used to have one individual who represented IBM to our clients. So this is all the annuity business, the transactional business, the new value. So it's really hard. I call that -- there's a few of them who are super humans who can do it really well, but it's impossible for the vast majority. So they tend to then lean on the people under them. So then that means their nature is more to be facilitators than to be deep subject matter experts on anything. Then they have to always call that person in. So we get the reputation of, well, you guys always walk with an army of people to sort of talk to us about what's there. And then you'll say, "Well, I need to come back to you in a week with somebody else."

So if you strip away a lot of that and you say, "Look, I really want you to focus on what brings value," that lets you change the nature, but it changes dramatically what to do every day. That's hard for people. So all these phrases [indiscernible] in business, right? Who moved my cheese, maybe the most sort of cynical way of putting it, but change is hard. And so sort of sticking with the change program, and then you get from what do you want me to measure you on? We want to measure you on getting much more value for the client, not just on the total book of revenue because a lot of that may be done by somebody 5 years ago. We want to measure you on how delighted is the client. We want to measure you on how are you guys co-creating together. And so we want to measure that NPS back from the client not from what our team say because if you get too financially driven, revenue is a proxy for NPS, and we all know that's not true.

I'd much rather sort of -- and say, yes, of course, there isn't a return from the client, but we don't have a healthy business. But it's not sufficient. We want to lean in and say are we doing things that are getting our clients much better prepared for the future. And if we can get that associated with us, that's a wonderful place to be in for the next decade. So I will tell you that that's the toughest changed in there.

Question
David McKay (Analysts)

I always like to say and tell the team, if it was easy, there's no moat and someone will copy it. If it's hard and you stay at it and you're successful, then you create a moat and create sustainable differential value. So it's just -- you got to keep at it. So that's just great.

So let's pivot to 1 of your big 3 strategies, and that's obviously hybrid cloud. And you organize hybrid cloud differently by industry vertical, but maybe just how you think of a hybrid cloud and the advantage of how you're trying to create a moat for IBM around your industry strategies.

Answer
Arvind Krishna (Executives)

So both parts. First, we made the observation about 4, 5 years ago. We didn't think that most clients would use only one public cloud. We think that many will use a couple, maybe more. You then get into the added complexity of sovereignty where some clients will say, look, the data can't really leave our national boundaries. That's not just true by the way for small countries. That's even true for many countries in Europe, like France and Germany, et cetera. You then go to -- there are some workloads that are going to require things that are local because of latency, if you're using AI to thwart, let's say, shop lifting, thinking about what some of the retailers are talking about this morning.

Question
David McKay (Analysts)

This morning in the Target results. Shoplifting was one of the -- shrinkage was one of the drivers of their performance.

Answer
Arvind Krishna (Executives)

Right. But then if you get into using AI in factories and industrialization, maybe even eventually in branch offices and so on to replace some of the labor, then you're going to turn into I need local processing. So there's going to be edge cloud. So if you say the environment straddles multiple public and edge and local and sovereignty, is it a place for somebody who has the right architecture and capabilities to help our clients go across that world? And that was the first, I would say, macro observation, and then you got to go operationalize it.

So we began from there and said, "All right, what's the best technology to do this on? And we came -- and we're pretty convinced that open source Linux and Kubernetes and containers is the right platform to go do it. When we said this 5 years ago, a lot of people looked at us like we had 2 heads. What are you guys talking about? I mean like there's this little thing on the side. I'll turn around and say jokingly, even to Dave, even your team [indiscernible] I said, yes, maybe, but they squinted and said, maybe, maybe not. I'd say today, even they would agree with the thesis so far before we get to any [indiscernible].

Question
David McKay (Analysts)

Absolutely. We don't want to be held hostage on development.

Answer
Arvind Krishna (Executives)

Exactly So then you say, how do you win there? So you can win if you can help people straddle these and say 1 team can go across these environments. And we're not going to tie you down. The days I think of tying people in with proprietary APIs and proprietary code are done. You can still win on the basis of engineering excellence and how easy is it and how convenient is it and does it have the biggest ecosystem and provide expertise so that you help your clients on their journeys. So we want to go down that path.

So that's where we're going on hybrid cloud. And that's why you see. Then Microsoft and Amazon are not competitors. They're actually partners. Great. Go along with that. By the way, both of them offer our Red Hat portfolio on their cloud catalogs. So you can actually consume it from whatever contract you have with them. And they like it because it creates more consumption on their underlying infrastructure. And we like it because it lets our clients experience that in conjunction with the other capabilities that they provide.

Then we said, all right, but there is an added space. Many clients have begun with sovereignty. But if I look at financial services, as -- probably to the chagrin a little bit, the regulators are really tough on that industry. Is that fair beyond almost any other industry?

Question
David McKay (Analysts)

Yes, we see our share of regulatory cloud audits, yes.

Answer
Arvind Krishna (Executives)

So then you say, can we help on that. Can we begin to use technology to begin to cut the owners down to say, can we provide a lot of controls? So the regulators can say, "All right, if you're using it this way, then it is actually much more acceptable." And these are always going to be around not just is it safe. We got that. But I'll say a lot of people can begin to do that. But it always begins, hey, somebody can break through. Now what do you do?

Or can you do the same controls across multiple environments? And remember, until about 5 years ago, this is not even a concern because when you're sitting in your own data center, by definition, everything inherits the controls of that data center. So now that you're sitting on somebody else's data center and they look at you and say, "Well, who else can get access?" That's a much tougher question to answer. Or who can change the firmware or who can change what? Or if something goes bad, what are you going to do to isolate this and stop it?

In your data center, it was easy. I pulled the network out and I'm kind of done. It's isolated. But not going to be so easy if it's across somebody else's hyperscale data centers. So can we provide these controls in a way? And we got some success there. So we've got clients like BNPP in France, Caixa in Spain, Citi here in New York that are beginning to do this. I won't mention those who are in the room and on the stage. So that's the way to get going. We do a lot.

SAP is used by a lot of clients for really critical workloads, so we want to be amongst the best at servicing SAP and cloud. I look at VMware, which is a great workload. And whether or not you're going to move away or with it, it's going to be there for the next decade or 2. So we want to be one of the best at servicing VMware on the cloud.

I begin to look at certain workloads like simulation and pricing. And maybe down the road, how do we begin to add quantum into the cloud as well as a specialized workload? So you kind of put that together and you say, "What is that opportunity? That's $1 trillion." And when people are maybe 1/4 of the way towards it, we think that's a great opportunity based on open-source standards, based on this observation of what the cloud environments are going to be like, and that's where we're kind of headed to.

Question
David McKay (Analysts)

From my seat in the financial services, as you're looking at strategy, I just appreciate the fundamental strategic pivot you made for IBM. 20 years ago, IBM wouldn't have made that decision. They said we have to own our end-to-end platform and build walls and compete. You've pivoted to a strategy of being the integrator, enabler and a capability set for the ecosystem. That's a fundamental strategic shift. It's impressive.

Answer
Arvind Krishna (Executives)

It's a massive shift. When you have a history of everybody's sort of pride is in, we invented magnetic tape, and then we did it ourselves. And by the way, all the way from, I think, making the materials to selling the end product and the solution and, in most cases, even servicing it for the next 10 years, to now saying, look, maybe 10% to 20% of the people who write the open source work for us in some form. The other 80% are really all over the world and going down that path. But doing the pivot and succeeding on doing it, Dave, that lets you tap into that innovation. And I think that's a big, big plus.

Question
David McKay (Analysts)

But your addressable market is in trillions now. So you made a pivot to a much broader addressable market with a different strategy. So just as I look at CEOs, I'm just impressed by how you've done that. You mentioned the partners, I think, was really important. Anything else on hybrid cloud that I missed? As you cover the partners that we wouldn't have expected to see, Salesforce is another big partner linking that, I think it's part of the addressable market.

Answer
Arvind Krishna (Executives)

So just talking about the partners. So if you think you're competing in the pure public cloud market, then most of the people I mentioned would be competitors. And we got to realize that -- I think you're hinting at this, Dave. Strategy is not formed only in corporate headquarters. Yes, that's an essential piece because capital allocation and training and education and who you hire goes in there. But in the end, the people on the ground have to be able to operationalize it and do it in front.

So if we say we are going to compete. No, they are not very good because they compete with us. It's really hard for the person on the ground to then say, but I want to partner with them because now that's opportunistic. So to me, you've got to align your strategy in such a way that gets embodied in the frontline. That was perhaps an even bigger change. So we said, "No, we are not going to compete with these folks." Okay, that's a strategic point. But then you say, we are actually going to create opportunity, consulting team, team who deploys software, our software properties. We are going to go enable them to run on top of AWS and Azure, and we're going to make it first class. By the way, Red Hat, first proof point along that. Red Hat on AWS, the ROSA or Red Hat on top of Azure ARO, they're in their cloud catalog. Their sales team is paid on it. There's a clear motivation for the Red Hat sales team, the IBM sales team and the Microsoft or AWS sales teams to collaborate because it's a victory for all of them.

So this is the harder part. They have to plumb it and they turn on and say, wait, but I'm giving away margin. Yes, but I'd much rather, to your point, get a bigger pie, and a smaller slice of a much bigger pie is a much better place to be in than fighting over a small pie and who gets the biggest slice of that, and getting that motion going.

Question
David McKay (Analysts)

But you need to be a more flexible organization, adaptable, to see how this ecosystem is evolving and not fully [indiscernible].

Answer
Arvind Krishna (Executives)

You've got to say, hey, both sides win. That's the conversation to go have. But look, the proof is there. When -- if I look back at early 2020, our pipeline and book of business with either of those 2 partners that I mentioned was in the tens of millions, maybe. At our scale, that's hard to even, to be honest, measure. Today, the pipeline, not the running book of business, but the pipeline is measured in the billions for each of them. And then you can say, how about the actual ARR to use that word or recurring revenue. For one of them, it's over $1 billion, for one of them, it will get there pretty quick. That's a big change from when you can barely measure it. So there is a return in there for all of us.

And that means we have more, I think, trained certifications. I think on AWS than any of their other partners, from being in a place where there were none to that place in 2 years is the speed to your point on the nimbleness. So that's all a lining strategy. You got to operationalize it so it hits the ground.

Question
David McKay (Analysts)

Accessing a revenue stream that in the old strategy model, wouldn't have been there or you would need to hold that customer to yourself.

So let's pivot a little bit to the next big platform, in artificial intelligence. And I think I heard you say once, I don't know if it was in a meeting with you or I saw it online, that we're collecting as a society 2.5 quintillion bytes of data a year. And what do you do with that? Because data is a cost until you create knowledge from it, and it's still a cost really until you create value from it.

So artificial intelligence is something we've invested heavily, but it's a fraction of how you see the -- how do you help your clients access this complex where we've hired 100 PhDs. We just saw in your video, promo or Eden launch -- Eden 2, our equity trading business, we're very proud of. But you're trying to help multiple clients and multiple industries, access that data and create value from it at the end of the day. So how are you going about it?

Answer
Arvind Krishna (Executives)

And I think, by the way, the video with Eden was a great example of the value AI can bring everybody. So 2.5 quintillion bytes a day. So for those who don't know, quintillion...

Question
David McKay (Analysts)

A day, sorry.

Answer
Arvind Krishna (Executives)

Quintillion is 21 zeroes. So that just gives you a sense. So the point is whether it's 2.5 or 2 or 5, it doesn't really matter. It's so much data, there is no kind of process-based manual technique that has any chance of being able to digest and get value from that much data. So I would tell you, artificial intelligence is the only set of technologies we know which can go do that.

Okay. Great. Now what do you do with it? How do you get business value? So we turn on and say, look, we are a B2C company, helping you do translation of languages or maybe recognizing a face. People who are dealing with billions of people are going to be far better at doing that than we are. So we decided to focus our efforts in 4 areas. One, almost everybody has to interact and converse, whether with employees or with their customers. And if I think about that, that's an area that we are going to invest in very heavily.

A couple of examples there. CVS who, when they came out with a vaccine for COVID here in the U.S., suddenly realized they could get 30 million to 60 million calls of people saying, "This is my medical condition, am I eligible for a vaccine? Where can I get it? This is my health plan." You think about that. And they said, wait, we could either begin to use something like AI to do that, and they did that with us. Or maybe we should go higher, and you'd begin to think how many people do you want to hire to be able to have even a reasonable queue if you're expecting 30 million phone calls. So that's a great example.

In the days of demographic shifts, McDonald's who you would think of would not use AI, but they are into -- but they have to have a couple of people against the drive-thru, taking orders. And when you can't even hire people, even if you're willing to raise your rates from $10 to $12 to $15 to $18 to $20 an hour. But there, it's not as simple as just speech to text. That doesn't help. It's got to be against the menu because somebody says I want a happy meal has got to understand, well, at this one, here are the 3 things. And then I'm going to substitute milk for apple juice, you got to begin to do those things. And so we're about at 1,000 of their franchises now. And our whole goal is to get to tens of thousands, and that's something that is great.

The second is something we should be really good at, AI for IT. And here, we're really proud, but we are applying it in a way that we excel. So we look at what we do around DevOps, and we say, "As opposed to writing code, could you tell your answerable what you want? And could it write all the code for you?" And that is something that's out in preview now and you're going to begin to see it, knowledge extraction. And we work with EY, for example, on when they do due diligence.

Question
David McKay (Analysts)

Can you do our COBOL code for us, too?

Answer
Arvind Krishna (Executives)

That will be coming.

Question
David McKay (Analysts)

I started as the COBOL programmer, actually.

Answer
Arvind Krishna (Executives)

That would be coming. COBOL is actually...

Question
David McKay (Analysts)

I'm not sure we're going to find anybody to recode some of that COBOL language inside financial institutions...

Answer
Arvind Krishna (Executives)

Maybe COBOL [indiscernible]. And knowledge extraction is a third area that I'll mention. So applying it in these enterprise areas, so really improving enterprise business processes is, I think, -- and to sort of turn it into -- we think that AI could produce about $16 trillion of productivity by the end of the decade. By the end of the decade, the world is going to be, what, $120 trillion in total. So something which is 10% to 12% is pretty amazing.

Question
David McKay (Analysts)

That is an incredible opportunity. So as we think about bias in the machines or perception of bias in machines, as we think about the evolving space of confidential computing, and part of extracting value from data is protecting the privacy and confidentiality of emerging data sets. How do you think about that space? And then how do you think about ethical AI at the same time? So 2 different questions that I'm linking in.

Answer
Arvind Krishna (Executives)

First, let's talk a little bit about confidential computing. So you have AI data sets. AI data sets could be public. They could be completely inside your boundaries. Some could come from partners. Everybody gets worried about correctly you don't really want the data to escape out into the wild. So confidential computing, which is sort of a broad term that covers everything from making sure your data stays encrypted, to how do you begin to share data, is it using in-memory encryption, is it using technologies on a cloud, is it using blockchain for probably the most extreme cases, is going to be critical.

Take the simple example of encryption. So if we encrypt something, but the administrator has a copy of the key, in the end, that's about as safe as you are because they could maybe get bribed, maybe inadvertently, maybe maliciously, give that key to somebody else. So to me, the simple thing is, to us, a client should own their own key. Meaning even if we are running the confidential compute for a client, we should not be able to look at the data if they have encrypted it. So those are simple examples of why we think this is so, so important.

And if you get into government and into financial data that is critical to keep a country working, believe me, they'll all begin to ask these questions. So these are inhibitors maybe to cloud and AI getting advanced, and we need to get rid of that inhibitor, and I think confidential computing gives you one way forward.

To your point on ethics, AI is wonderful. It can learn -- let's make it really simple. It's like a child. Children observe things and then they can repeat them. And like in humans, they observe certain behaviors, and they can copy their behavior for good or for bad. AI is wonderful. It can, with great exactitude, learn from existing data and repeat it with perfection.

Well, if you happen to have a bias in your prior data, AI is going to be wonderful. It will repeat it with great speed and great [indiscernible]

Question
David McKay (Analysts)

Great predictability, yes.

Answer
Arvind Krishna (Executives)

So like many things, I would never sort of blame the technology. Now the question is, how do you correct for it? And the reason ethics is the right word because you can't take it upon yourself to correct for things that you as the programmer or a company want to correct for, which ones are allowed to be corrected for, which ones can you watch for.

Example, in mortgages, should there be a difference based on gender? Probably not. How do you begin to correct for it? So then you get into, okay, how do I check my model, how do I make sure my model doesn't drift. And checking the model is easy. You might tell it the gender and you change -- keep everything to constant, but just change the gender if you get a different answer. That's a problem. I'm being rather simplistic, but you got to get an idea. So all the techniques we all know in software engineering, of testing, et cetera.

Question
David McKay (Analysts)

We're all engineers, and if we can actually perform, we'll solve for this.

Answer
Arvind Krishna (Executives)

I would tell you. So -- but checking for the ethics is really important. But if you go above a [ beta ] level about 2 years ago, we haven't been looking at it, but it kind of pushed us over the edge. Facial recognition is an example where we know that AI can be biased, not so much as pure bias but a bias based on skin tone and ethnicity, et cetera, maybe based on the training sets, maybe based on the physics, maybe based on lighting, that's our speculation. All we know is that there is a difference in accuracy.

And so as a consequence, we pulled all facial recognition from the market. By the way, after our action, many others did also. But that's an example of where you come to the conclusion, well, we can do all the corrections, but it's not going to get done quick enough. So both sides of it. I think we should apply it into how we build it, but we should also probably make business decisions based on it, even if it hurts you in the short term.

Question
David McKay (Analysts)

How are you concerned -- or how bring society with us on this journey and make sure that society doesn't turn against the perceived bias to perceive negatives of artificial intelligence because it's got a bit of a damaged brand because even media doesn't understand it, so they pick on it with a bit of a negative bias versus the positives that can do. Any thoughts about how we create transparency, or how do you make sure society feels comfortable with us using these technologies?

Answer
Arvind Krishna (Executives)

Look, I think for one, I think that we should always apply things in areas where there is a productivity price to be garnered, but maybe the consequences are not irreversible. So you got to begin with what the Europeans would call Level 1. You apply it in areas that are somewhat easy. Level 2 is maybe they're harder. There is an application, but it's still reversible. Maybe some kinds of I'll call them small financial decisions, maybe inside an enterprise processes and so on.

Before you're ready for the third and fourth, which is making decisions that have pretty dramatic consequences on people, so if you want to get there, now you've got to be much more transparent. We use these data sets. We are willing to demonstrate the checks we use. We are willing to demonstrate that we make sure there isn't a drift or a bias in what's happening. And so I think that creates transparency, and that will create trust. So I think that there are ways to go, and I would always say, begin with the first 2 levels before you get to the third and the fourth.

Question
David McKay (Analysts)

Okay. That's a great idea. I like that. Third platform, less developed than the first 2, so quantum computing. So as you look at the nascency of quantum computing, the power, how do you compare it where it's in its evolution to, say, blockchain, artificial intelligence and its opportunity set to change the world and to monetize that value. So how do you put it on a scale with the other 2?

Answer
Arvind Krishna (Executives)

So if you say cloud or public cloud was 2006, day 1, just to pick -- you can debate plus/minus 5 years, and you see where it has come to today over now 16 years, multiple hundreds of billions of dollars of revenue and probably trillions of value for those who use it. If you look at AI, probably 10 years behind that journey. So you would say it's maybe 5 to 10 years in, probably tens of billions in revenue, hundreds of billions in value.

I personally believe that if I go out 10 to 15 years, AI will cross cloud in terms of total value and revenue because it's going to infuse everything. I think quantum will be bigger than cloud 10 years out. So if I go to the middle of the next decade, so why do I say that? The sets of problems it can solve, you cannot solve using today's computing. Applying classical to something that is growing so big sounds kind of funny, but I'll use the word classical for today's computing.

So problems on risk, problems on arbitrage, problems on lightweight materials, maybe carbon sequestration on a problem that is worth trillions, if you can solve problems worth trillions, that means you're going to create revenue in the hundreds of billions. That's why I'm sort of so excited about it because we don't have any other way to solve those problems, right? There's no other way to go ahead and solve.

And so how far away are we? I think we're somewhere 3 to 5 years from commercial advantage, but probably 5 to 10 from the full deployment that I just spoke about. That's massive. That's exciting to be able to go get that done.

Question
David McKay (Analysts)

It's a huge opportunity. Is it also a threat? I mean, some of the -- we're experimenting in the world of fraud with quantum computing, but also I get warned by a number of experts, I don't know if it was at IBM or not, about it's going to break down our security algorithms much more quickly and that it's -- so is it a threat as well as an opportunity? Do we have to worry about playing defense as well with this capability?

Answer
Arvind Krishna (Executives)

Maybe I'd say offense, not defense. So quantum computing, if it gets anywhere near the scales I just talked about, will break today's encryption. So even if we have put stuff on tape and it's encrypted, somebody could get hold of it 5, 7 years from now and decrypt it. That's dangerous because there is data we all have and we don't really want out there, okay? So that's the threat.

Now in this tier in the United States, which is a standard body run by the government, has put out 4 not yet full standards, but what they're calling proposed standards that allow you to put out what is called quantum safe encryption. I would strongly urge everybody, be offensive for your most critical data, not for everything. Start using those algorithms now.

Pure sort of truth in advertising, we helped write 3 of the 4, not all 4, but I'll use the word help because these are always large sets of people from academia and companies who write these things. So we began working on quantum safe encryption because we realize this danger. And so it's appropriate to go create an alternate.

And by the way, it's no harder than to do this encryption. It's just a different kind. And so I would strongly urge -- and that's why I say offense. So yes, it's a threat. But if you got to get good about it, and for example, on our mainframe systems, we already put out quantum safe encryption as an option. So anybody who uses these for the retail banking or insurance or claims can turn on the quantum safe encryption and be reasonably comfortable that they are getting much better protection.

Question
David McKay (Analysts)

Right. No brute force attack by a super power computer. I got a -- which will happen. No, it's coming at us. I talk to my Chief Technology Officer about it frequently as well. I got a bunch of questions. Where should I finish here? We only have a couple of minutes. I'll try to get this one in and if there's more. But it's the future of work.

So you've talked a lot about consulting services, creating value, how your organization interface with the client and brings these new technologies to create value. But it requires us to adapt to a very different profile of work, different capabilities, as you talked about. How do you think about this hybrid world? What do you need from your customers, and what do you look for from your own organization in this new world? Because we're all struggling with it as leaders trying to find this new place of innovation, engagement and performance.

Answer
Arvind Krishna (Executives)

Yes. David, as I think about it, at the end of the day, we're all corporations, companies, whatever you want to use. So our goal has to be to provide great service to our clients. So if I look at our clients, and I back up to what we started with in the macro, they all want to use technology to help improve their own businesses. And today, it comes down to what advantage can we create. So I'd love to use the word co-create. How do we sit down with you to co-create the value? That means we've got to put some people in with you because I think creativity, really quick decision-making, serendipity gets driven with being together.

Okay. Once you get beyond that, then you say, but I also want to very quickly prototype or see how this works. You've now got to bring skills in from all over the globe because in the current demographics, I kind of have tongue and cheek, unless somebody can find something the size of China, maybe there is a hidden country in the Pacific Ocean. I kind of doubt it. So you're not going to find another ...

Question
David McKay (Analysts)

Bunch of people on the beach looking to work, right?

Answer
Arvind Krishna (Executives)

So that means you're going to have to tap into skills from all over. So we have got to be sort of flexible. You got to have the people who can have that business conversation and technology conversation in person, but tap into others. That means we are in this mode of we all use the word hybrid. But it's hybrid as in there are people all over the globe, they can come together to help solve a client's problem, but there are people who are sitting together and working. And the client themselves may be across multiple locations.

And so I think that having people who are flexible, but who are really motivated by helping solve our clients' issues but being deep enough both technically and business-wise to go do that and then have enough thrust to leverage their people everywhere and having then the hiring engines for having people put those skills, the leadership capabilities of who can work well in these environments and not being tied into just multiyear contracts, but saying, look, we win on the basis of the value we can provide, I think is where the world of work is going.

By the way, that leads to every one of you. That means the skills that are needed change every 3, 4, 5, 6 years. They don't stay constant. And that, I think, is the big uncertainty that is just part of this whole period that everyone faces. How do you keep evolving your own skills every few years?

Question
David McKay (Analysts)

That's a great perspective. We've run out of time. I've got about 4 more questions. Thank you. Thank you. You're a thought leader among CEOs. You've always been a thought leader, as I think you were recognized as 1 of the top 25 technologists that's going to change the world early in your career, and you're continuing that as CEO. You're leading a firm that has always been at the forefront of leading our society towards leveraging technology to the references you gave and building out those first computers. You've done a magnificent job in trying to reimagine the role you're playing in future technologies. And we're all cheering you on.

Thank you so much for spending time with us today and articulating how you're pivoting help clients like us and clients in the room really adapt and leverage these incredible new capabilities. Please give a warm round of applause to Arvind Krishna, CEO of IBM.

Answer
Arvind Krishna (Executives)

Great. Thank you. Thank you for the opportunity.

Question
David McKay (Analysts)

Thank you so much. Thank you, everybody.

© S&P Capital IQ 2023
Copier lien
Toute l'actualité sur INTERNATIONAL BUSINESS MACHINES CORPORATION
03/02
01/02
01/02
01/02
01/02