00:00Arvind Krishna joins us. It is a joy to have you on our network in Arvind. And look, speed is
00:05the
00:05name of the game. This deal has completed swiftly and indeed Confluent is all about speed and data
00:11analysis. How is that important in this age of AI? Yeah. So Karan, it's great to be here with you
00:17and on Bloomberg. So just look at what Confluent does. Moving data in real time so that it gets
00:24available both for the enterprise, for analytics, but more importantly for AI agents. And doing
00:31that in a way that is the most capable product in the world is why it is so exciting to
00:37get
00:37it done. And your point on speed, I think the regulatory environment is definitely friendlier
00:43where we got this done in just under four months, whereas it used to take a lot longer a few
00:50years back. If regulatory environment is friendlier, should you be doing more of it? Should
00:55there be more M&A, particularly with some beaten up overall valuations of software companies
01:00at the moment? I'll just say, watch the space. Oh, watch the space. Okay. But where would you
01:05want to add on in this moment? I mean, what would make sense to be adding to your portfolio?
01:09So we are very focused hybrid cloud and AI and the intersection. So if you look at Confluent,
01:16some of the data is in cloud, some of the data is in SaaS properties, some of the data
01:20is on premise. An AI agent needs to get hold of it wherever. So that's the hybrid piece combined
01:26with the AI piece. Our sweet spot is going to be hybrid cloud, AI, automation, other areas where we
01:34are very, very focused on M&A activities, as well as organic development. If you look at what we have
01:40done around our WatsonX products, what we have done around our mainframe modernization, these are all
01:46things that we have built organically. But then we supplement it with targeted acquisitions, where
01:53the multiple makes sense, where it fits our strategy, and where we can normally increase the growth rate
01:59of the target property, which we certainly hope to do in the Confluent case.
02:03Let's talk about WatsonX a little bit. Let's look more at partnerships, because of course,
02:08that's what was announced, a little bit more of a deeper partnership with NVIDIA yesterday,
02:12helps your stock. We're seeing expanded collaboration. Again, this is about faster
02:15data analysis, but cheaper, more effective. How does that help you seal more deals?
02:21So in that one, actually, the work we're doing together with NVIDIA was a five times speed up. So
02:27five times, not 5%, not a little amount, but five times. So there, we began to leverage the NVIDIA
02:34GPUs together with some other CUDF software, combining it with our WatsonX data. And the
02:42example we used was our client Nestle, where together, we managed to get that speed up across
02:48their massive amounts of data. And that really is important. In that case, combining some of the
02:54technologies we work on also in open source with the Presto data engine, NVIDIA, and the example at
03:01Nestle, but then we are very excited. We're going to do more work on that, and then take it into
03:06the
03:06market and take it out to hundreds of clients from there.
03:09You're offering these tools, but you're also helping companies like Nestle just embed AI, make sure
03:14they're using it the most effective manner possible. When your consultants go in, how much does a Nestle
03:18want to use your offerings? But those of Anthropic, of OpenAI, of others, how much do you see that as
03:25a
03:25competitive force or a competitive threat?
03:27Look, our goal has always been that we want to help our clients integrate the best capabilities from
03:33where they come. And the word integration comes in here because to some extent, we are model agnostic.
03:39We believe that our clients are going to use the frontier models from all three or four of the main
03:45providers. They're going to use open source models, and they're also going to use models from us,
03:50though we tend to make much smaller models. But helping them really get value, and I really believe
03:552026 is the year when enterprises are going to be focused and obsessive on ROI or what they're doing
04:01on AI. And so if you take Nestle, how do you take cost out of procurement? How do you take
04:06cost out of
04:06HR, the finance function, procurement, all of these things? And that's where our consultants and our
04:13technologies come together to help them do that using both technologies from IBM, but also in the case of
04:18Nestle, from many other of the partners and the software vendors that they use.
04:22So do you think the investor base, and just more broadly, your customers as well, see that AI isn't a
04:30double-edged sword for you, but actually that competitive threat that sunk the shares so much
04:33on February the 23rd was a misunderstanding of what Anthropic could do with COBOL?
04:38I'm going to be much more straightforward. I believe that the investors and parts of the media
04:45misunderstood what that blog said and did. I actually am convinced AI is the tailwind for us.
04:51It's not a mixed half headwind, half tailwind. AI is a tailwind for us in terms of how our technology
04:57will get adopted, in terms of what we do with our clients. And by the way, we had put our
05:02tools to
05:03convert and modernize COBOL two and a half years ago. So there was no new news in that blog, but
05:09they
05:09certainly can get a lot of attention. You talk though, what also gets attention is being able to become
05:14more productive, take costs out of the business using AI. A lot of people see that as the costs of
05:19people. We're seeing block remove 40 percent of their employee base. You hear about time, headline
05:25after headline. Is that what's going to happen to the labor force? So I'll take ourselves. We've been very
05:30public that we have taken four and a half billion dollars out of the enterprise with a mixture of AI
05:36and
05:37automation. That's a hard number. Now, over three billion of that got reinvested into the business in more R&D,
05:45in
05:45more sales, in more marketing, in more delivery, in more what we would call client engineering, where
05:50clients love the fact that our engineers can work with them to help them solve problems. So there is a
05:56reshaping of the workforce, but there isn't a net decrease. And to put our money where our mouth is, we
06:03will do two to three times as much entry level, meaning college hires this year as we did last year.
06:09Wow. So you're still seeing that demand for that level of technical prowess coming straight out of
06:15university. But I'm interested, you have said in the past, look, you see job displacement, your term
06:20and phrase of five to 10 percent. Is that done now at IBM, do you think? I think that we're
06:25probably
06:26halfway down that total journey. And I still stick to my total numbers in that range. And I think we're
06:32halfway through that journey. But our total employment has remained roughly stable. So to the point that I make
06:38about how there will be more opportunities, but you've got to be willing to upscale yourself and retrain
06:43yourself. What's not stable is the geopolitical environment. How is that affecting you? How is that affecting the
06:50consulting part of the business in particular? Look, let's acknowledge there's a lot of pain and disruption that's
06:56going on in the Middle East right now. I think we should feel a lot of sympathy for our people.
07:02We have thousands of
07:03employees in that region. The vast majority, I think, are reasonably stable and are getting their
07:09work done. There are about 20 percent of the people in the Middle East who are who are disrupted, who
07:15are
07:15not able to get to their clients, who are not able to get the work done. I think within a
07:22quarter is going to be
07:24very, very minor because most of this work is much more long cycle than short cycle as in days or
07:29weeks. If this goes on for many more months, then I do think that we will take a slight headwind.
07:36But in the
07:37end of the day, our consulting business in the Middle East is only a few single digit percentages of our
07:42total
07:42business. So it won't impact IBM's top line and bottom line, but it is going to be impactful to some
07:48of our
07:49people and some of our clients who are in the region. Thanks for that nuance, Arvind. And I'm interested just
07:54on the
07:54nuance of total sales growth in consulting. Will that come back to growth, do you think, in the near
07:59term? I think when I was talking in January and also in October, I felt that the second half of
08:08this year
08:08will be a lot better on consulting growth than right now. We kind of said that we could see an
08:14inflection
08:14coming where the first half of last year was negative. The second half of the year turned kind of
08:20flattish. I think that we'll see continued improvement in that inflection to maybe slight
08:25growth in the first part of the year. But true growth, I think, is still out in the second half
08:31of
08:31the year. True growth, true excitement about quantum always seems to be in the second half of the decade
08:36or the millennium. But you really are putting your money where your mouth is when it comes to quantum
08:41computing becoming, you know, really tangibly useful by the end of the decade. You want to have a fault
08:46tolerant quantum supercomputer by 2029. How on track are we there? How on track are we to integrating
08:51quantum within the data center? So in all three of your questions, Caroline, I think we are
08:57completely on track with what we had said. We said that we're going to have better qubits, better
09:04quantum processors. And our processor at the end of last year is now being used, whether it's for
09:11medical work at Cleveland Clinic, whether it's for bond pricing at HSBC. Those are real use cases that
09:19are coming out. We expect to see error correction between this year and next year demonstrated and
09:25out there. So that'll be another big plus. We're investing very heavily in how we bring together
09:31our normal computers and quantum computers. We put out a roadmap for quantum centric supercomputing
09:37that has gotten a lot of attention or partnerships with academia, with institutions. I'm so excited
09:45by what we're doing in Illinois, but also at RPI with MIT and in other places. We have done some
09:52work
09:53with the national labs. This is all tells us how tangible it is and you can feel it. People are
09:58now
09:58realizing this is not science fiction. This is not engineering to get through the next two to three years.
Comments