Bloomberg Audio Studios, Podcasts, radio News.
Romain brings us a very important conversation right now against the backdrop of this market. Romain Bostik at the Mazuho Tech conference underway right now in New York with the CEO of IBM. This is live on Bloomberg. Romain, take it away.
Thank you, Joe.
I am here with the CEO of IBM, Arvind Krishna, here at the Miszoo Technology conference in Lower Manhattan.
Great to have you here.
Pleasure to be here with you. Roman.
Obviously, all the talk is about AI. We talk about digital AI, we talk about biological AI, physical AI. Everyone's already looking around the corner, including you, to quantum, which of course many people think is going to be the next big thing, something that IBM has been working on for years. Two years ago, you laid out a relatively ambitious target that we would actually see quantum advantage using IBM hardware by the end of twenty twenty six, for halfway through the year.
Are we going to get that? Absolutely?
If I look at what some of our partners like the Creative and Clinic are getting done, what the national labs like Okage are getting done, I can already see the early signs of what they're getting, the progress they've made.
When we talk about what that progress could mean not only for the world of business and finance, but as an application through our society.
Is healthcare the first place that gets it.
I think that materials or maybe simple molecules are the very first place. Maybe magnetic materials as alternates for some of these things that are very hard to get. I'll use the word rare as a bit of a fun over there. I think lubricants to maybe get more effective energy, maybe better ev batteries, and then small molecule drugs are
there in the first set. But I also want to be equally excited about what we could do about modeling weather perhaps or modeling flows to improve how chemical plants work, maybe aerodynamics to improve how vehicles and airplanes work. These are really really exciting use cases for us.
I am curious.
So I mean, when we talk about actually starting to see some sort of proof of life on quantum advantage this year, your own head of research has said that that process might be a little bit more elongated. Is there a disconnect between your ambitions and what he's seen from the scientific perspective?
No.
So I'll take an example of what's happening on molecule properties. So last summer June of twenty five, we could maybe understand a five at of molecule you react with. That's not very exciting. You could do that on a piece of paper if you're smart enough.
Okay, I'm not smart enough, but I trust here.
If Yeah, last November, that team could do three hundred at of molecule. Okay, you'll say that's great progress, five to three hundred, but you could probably do that on a normal supercomputer. So okay, that's not one of an advantage. This April they did twelve thousand atoms. They did a half of a protein called tripsin. I think they're on their way to about thirty thousand items already. So five to three hundred, same amount of time, three hundred to twelve thousand.
One more month, twelve to twenty one thousand.
It gives you the rate of improvement, some of it from the hardware, some of it in algorithmic techniques that.
People are learning of what can work.
By the way, at this size, you can't actually do it.
Any other way.
Okay, Well, we talk about the commitment that you've made to this. You've been invested or committed to invest about ten billion dollars in this, which I assume is a tip of the iceberg. You got some pretty big backing from the federal government giving you one billion dollars, not that you need it, but certainly a vote of confidence.
When you start to see.
The government getting involved, obviously you have a lot of I guess quasi competitors in this space also getting involved.
Does that accelerate this pace?
That's what I believe the government investment is for. So if you look at it, it's not so big that it's controlling thing. However, it's a sign of endorsement, and it is a signal of them wanting to get speed, meaning, IBM, you're going to do this. If we commit to do this, will you do it faster? Will you open it up to everybody else? So it's open to the ecosystem, And we said yes, because we actually believe an ecosystem is important.
And if government investment gets us there a year or two or three before we might otherwise, both of those are worth it. And that is what is important about that.
The one billion dollar investment from the Trump administration is specifically to build up a chip boundary.
If you are a quantum chip foundary.
Quantum foundry that we have committed will be open to all comers that are acceptable to the United States. So the three dozen odd companies that are here, we are happy to give them our best technology, the same as we use for ourselves at that level, whether they all get access to it, and you say.
All commers as to what's been approved, this is the idea that this is about furthering the US's aims to make sure that they have an advantage with regards to what gets made, and I would presume so that it doesn't get in the hands of the countries that we don't want to ask.
Well, I think that is easy.
I mean, like there have always been restrictions and advanced technologies that you cannot sell it or provide it to certain places.
I think that is fine.
But in a new area, it's really important to be first, both for the economic competitiveness long term, but also for national security. I think for those two reasons it's impetative that we get there first.
Do you think the US will be there first.
On quantum Absolutely, You're.
Not concerned about some of the reports that we've heard about some of the progress that we've already seen in China.
I'm paid to be paranoid, so I wouldn't say I'm not concerned. However, from everything I can see, I think we are ahead a head by a decade. Probably not, but are we a head by a couple of years, That would be my guess.
I do want to talk about just AI in general. At your think conference last month, you talked a lot about this idea of the progress company are making with regards to the deployment of AI. More importantly, some of the progress they aren't making. You've talked about this idea, or IBM has certainly talked about this idea of the limited ROI that we've.
Seen so far.
You say that the companies that are getting this right aren't necessarily the ones that are just deploying more models, but they're finding ways to take the models that they already have and use them better.
Explain that.
I'll give you an example.
So you can certainly improve productivity of individuals. We'll call that productivity apps. That's where both economists and CFOs get very concerned. But where's my bottom line on that? Is it leading to less expense or more revenue. Where is that? And that's hard to find. But let me give you a second example, and you'll say, okay, I can see
that we are pretty acquisitive. Anybody who does MNA, you often bring the entity in and actually, normally your profits might decrease for the first year or two until you take all of the costs and get the cost s energy.
Using AI tools, now we can.
Probably add ten points of profit s energy on day one because the amount of time it used to take to move contracts over, to do all of the sales automation, to do all of the revenue forecasting, all of that now using AI can be shrunk down to literally a few weeks. If you can do that, I cannot take that ten percent. I'm not saying it's ninety percent, but ten percent and invest it instead in marketing and R and D and innovation to drive the revenue growth even higher.
That's an example of what I mean by use the AI to really fundamentally change how you do business. Has got to become part of your operating model of the company.
Are you taking that medicine yourself?
Internally?
We are.
In the last acquisition we did, we actually managed to take out about sixty percent of the GNA on day one.
What's the message though, that you have not only for your partners but also internally for your own workforce with the guards to how this actually not only helps, but maybe the idea that maybe it requires a smaller headcount.
It's not necessarily going to be a smaller headcount. So a lot of people talk about this. So let me give you an example on early on college hires. We tripled our hiring in twenty twenty six compared to twenty twenty five. Okay, is the entry level for entry level positions tripled?
Absolute number tripled.
So you look at me and say why, Because my belief is if these tools make people who are coming in out of college more productive quicker, that means we are going to be advantaged by having more capacity to build things at a cheaper cost point, which means we
can go get market shared. Will the shape of the workforce change, absolutely, but I think there will be more people in the value driving functions, be IT R and D, be IT sales, be IT marketing, and the creative activities and some of what I would call classic back office will decrease because you don't need so much there, So the shape of the workforce changes. Total employment is not decreasing.
I want to talk about security.
Obviously, we know the potential opportunities when it comes to AI, quantum and everything in between. You've invested about so far a commitment of about five billion dollars towards something called Project Lightwell, the idea that it gives your partners a little bit more visibility about flaws in that sort of AI chain, if you will, for my sort of uninformed framing of it. But talk about why that's necessary, and more importantly, is five billion going to be enough.
It's hard to tell whether any amount is enough here. I think, look, this is like asking why are cyber criminals existing, and the old really sudden joke call as well, I draw banks because that's where the money is. Well, if money's inside the cyber infrastructure, they're going to come. Our investment in Light twelve. I think a lot of people can point out the problems and point out vulnerabilities. Our investment in light Trail is actually about fixing them.
So we are saying that not just open source that we are good at, but if you have open source and the source code is known, we can use these same tools that I used to find vulnerabilities to also create the fix and give it back to the community.
That is really what light Trail is about.
Of course, in order to do that, you need cooperation around where people are finding things, how critical are they, where are they in the critical infrastructure? So that is the whole project. But it's actually about us coming up with an antidote, not just the problem.
This I can just tell it by look at interface. This is probably an exciting time for you. IBM is obviously a storied company when it comes to computing, and we are clearly at the inflection point of something new, maybe something better. You've gotten a big bump in your stock price, largely because of at least the government backing
or at least the government support. Is there a sense here when we talk about the potential for an acceleration in revenue growth the widening of profit margins, that that is something that is imminent or is that something still long term?
Well depends on what you call long term.
I kind of like in what we're doing on quantum and cyber to where GPUs were in twenty fifteen twenty sixteen, and each era of technology tends to go faster, so if I say QPUs plus cyber is.
Going to go fast.
I think in the next two three years being going to see incredible returns coming from these technologies into our numbers.
And with those ambitions, you have the capital you need in order to do that. Any plans to maybe go to the market to look for a little bit more.
We go to the markets all the time. We are always in the debt markets, and so far our debt investors have been very confident.
In US and we get great rates there.
We could maybe dip into other markets, but I think for what I'm pointing out over the next five years, I think we have the capital and we got the WAB withth All insight.
Arvin really appreciate you joining us.
My pleasure.
That's Arvind questionna. He is the CEO of IBM