This is Bloomberg business Week inside from the reporters and editors who bring you America's most trusted business magazine, plus global business finance and tech news. The Bloomberg Business Week Podcast with Carol Messer and Tim Stenebeck from Bloomberg Radio.

This is Bloomberg business Week inside from the reporters and editors who bring you America's most trusted business magazine, plus global business finance and tech news. The Bloomberg Business Week Podcast with Carol Messer and Tim Stenebeck from Bloomberg Radio.

Hi, everyone, Welcome to the weekend edition of Bloomberg Business Week. We're gonna just kind of cut to the chase here because this week has really been so much about what is one of the world's largest market cap companies. It's the chip maker at the heart of the artificial intelligence boom. We are talking about Nvidia, which wrapped up earning season for US and gave a revenue forecast that fell short of some of the most optimistic estimates, stoking concerns that its explosive growth is waning more in the report as it was coming out.

Hi, everyone, Welcome to the weekend edition of Bloomberg Business Week. We're gonna just kind of cut to the chase here because this week has really been so much about what is one of the world's largest market cap companies. It's the chip maker at the heart of the artificial intelligence boom. We are talking about Nvidia, which wrapped up earning season for US and gave a revenue forecast that fell short of some of the most optimistic estimates, stoking concerns that its explosive growth is waning more in the report as it was coming out.

In just a moment, also this hour, Dell and AI, we speak to the CTO and chief AI officer at the Computer Giant about what's coming.

In just a moment, also this hour, Dell and AI, we speak to the CTO and chief AI officer at the Computer Giant about what's coming.

Up and later on. Rather than large language models of AI h, we talked simply about learning another language with the CEO of Babbel. All of that to come. We begin with the most anticipated earnings report from this past week, and Vidia, the chip maker, failed to live up to investor hopes with its latest results this past Wednesday, despite beating for the past quarter, but nonetheless, it delivered an underwhelming forecast in news of production snags with its hyped Blackwell chips.

Up and later on. Rather than large language models of AI h, we talked simply about learning another language with the CEO of Babbel. All of that to come. We begin with the most anticipated earnings report from this past week, and Vidia, the chip maker, failed to live up to investor hopes with its latest results this past Wednesday, despite beating for the past quarter, but nonetheless, it delivered an underwhelming forecast in news of production snags with its hyped Blackwell chips.

The company's quarterly report met or beat analyst estimates on nearly every measure, but in Vidia investors have grown accustomed to blowout quarters and the latest numbers didn't qualify. After the company's earnings call, CEO Jensen Wong sat down exclusively with Bloomberg TV's Ed Lovelow to talk about the results and how AI is impacting how every layer of computing is done.

The company's quarterly report met or beat analyst estimates on nearly every measure, but in Vidia investors have grown accustomed to blowout quarters and the latest numbers didn't qualify. After the company's earnings call, CEO Jensen Wong sat down exclusively with Bloomberg TV's Ed Lovelow to talk about the results and how AI is impacting how every layer of computing is done.

I think the market wanted more on Blackwell. They wanted more specifics, and I'm trying to go through all of the call and the transcript, it seems like a very clearly this was a production issue and not a fundamental design issue with Blackwell. But the deployment in the real world, what does that look like tangibly and is there sort of delay in the timeline of that deployment and thus revenue from that product.

I think the market wanted more on Blackwell. They wanted more specifics, and I'm trying to go through all of the call and the transcript, it seems like a very clearly this was a production issue and not a fundamental design issue with Blackwell. But the deployment in the real world, what does that look like tangibly and is there sort of delay in the timeline of that deployment and thus revenue from that product.

I let's see, that's just the fact that I was so clear and it wasn't clear enough.

I let's see, that's just the fact that I was so clear and it wasn't clear enough.

Kind of tripped me up there right away.

Kind of tripped me up there right away.

And so so let's see, we made a mass change to improve the yield. Functionality of Blackwell is wonderful. We're sampling Blackwell all over the world today. We show people giving tours to people of the Blackwall systems that we have up and running.

And so so let's see, we made a mass change to improve the yield. Functionality of Blackwell is wonderful. We're sampling Blackwell all over the world today. We show people giving tours to people of the Blackwall systems that we have up and running.

You could find.

You could find.

Pictures of Blackwall systems all over the web. We have started volume production. Volume production will ship in Q four. Q four, we will have billions of dollars of Blackwell revenues, and.

Pictures of Blackwall systems all over the web. We have started volume production. Volume production will ship in Q four. Q four, we will have billions of dollars of Blackwell revenues, and.

We were ramped from there. We were ramped from there.

We were ramped from there. We were ramped from there.

The demand for Blackwell far exceeds its supply, of course in the beginning, because the demand is so great. But we're going to have lots and lots of supply and we will be able to ramp. Starting in Q four, we have billions of dollars of revenues, and we'll ramp from there into Q one, into Q two and two next year. We're going to have a great next year as well.

The demand for Blackwell far exceeds its supply, of course in the beginning, because the demand is so great. But we're going to have lots and lots of supply and we will be able to ramp. Starting in Q four, we have billions of dollars of revenues, and we'll ramp from there into Q one, into Q two and two next year. We're going to have a great next year as well.

Jensen, what is the demand for accelerated computing beyond the hyperscalers and meta.

Jensen, what is the demand for accelerated computing beyond the hyperscalers and meta.

Hyperscalers represent about forty five percent of our total data center business.

Hyperscalers represent about forty five percent of our total data center business.

We're relatively diversified today.

We're relatively diversified today.

We have Hyperscalers, we have Internet service providers, we have Sovereign AIS, we have industries enterprises, so it's fairly fairly diversified. A site outside of Hyperscalers is the other fifty five percent.

We have Hyperscalers, we have Internet service providers, we have Sovereign AIS, we have industries enterprises, so it's fairly fairly diversified. A site outside of Hyperscalers is the other fifty five percent.

Now, the application.

Now, the application.

Use across all of that, all of that data center starts with accelerated computing. Accelerated computing does everything, of course, from well the models the things that we know about, which is generative AI, and that gets most of the attention.

Use across all of that, all of that data center starts with accelerated computing. Accelerated computing does everything, of course, from well the models the things that we know about, which is generative AI, and that gets most of the attention.

But at the core we also.

But at the core we also.

Do database processing, pre and post processing of data before you use it for generative AI, trans coding, scientific simulations, computer graphics of course, image processing of course. And so there's tons of applications that people use are accelerated computing for, and one of them is generative AI.

Do database processing, pre and post processing of data before you use it for generative AI, trans coding, scientific simulations, computer graphics of course, image processing of course. And so there's tons of applications that people use are accelerated computing for, and one of them is generative AI.

And so let's see what else can I say? I think that's the.

And so let's see what else can I say? I think that's the.

Lever jump in Jensen. Please on Sovereign AI. You and I've talked about that before, and it was so interesting to hear something behind it. In this fiscal year there will be low double digit I think you said billions of dollars in Sovereign AI sales. But to the lay person. What does that mean. It means deals with specific governments, if so, where.

Lever jump in Jensen. Please on Sovereign AI. You and I've talked about that before, and it was so interesting to hear something behind it. In this fiscal year there will be low double digit I think you said billions of dollars in Sovereign AI sales. But to the lay person. What does that mean. It means deals with specific governments, if so, where.

It's not necessarily sometimes it deals with particular regional service provider that's been funded by the government. And oftentimes that's the case in a case of in the case of Japan, for example, the Japanese government came out and offered.

It's not necessarily sometimes it deals with particular regional service provider that's been funded by the government. And oftentimes that's the case in a case of in the case of Japan, for example, the Japanese government came out and offered.

Subsidies of a.

Subsidies of a.

Couple of billion dollars I think for several different internet companies and telcos to be able to fund their AI infrastructure. India has a sovereign AI initiative going and they're building their AI infrastructure. Canada, the UK, France, Italy, I'm missing some body, Singapore, Malaysia. You know, a large number of countries are subsidizing their regional data centers so that they could become able to build out their AI infrastructure. They recognize that their countries knowledge, their countries data digital data is also their natural resource, not just the land they're sitting on, not just the air above them.

Couple of billion dollars I think for several different internet companies and telcos to be able to fund their AI infrastructure. India has a sovereign AI initiative going and they're building their AI infrastructure. Canada, the UK, France, Italy, I'm missing some body, Singapore, Malaysia. You know, a large number of countries are subsidizing their regional data centers so that they could become able to build out their AI infrastructure. They recognize that their countries knowledge, their countries data digital data is also their natural resource, not just the land they're sitting on, not just the air above them.

But they realize now that their their digital knowledge is part of their.

But they realize now that their their digital knowledge is part of their.

Natural and national resource and they had to harvest that and process that and transform it into their national digital intelligence. And so this is a this is what we call sovereign AI. You could imagine almost every single country in the world will eventually recognize this and build out their AI infrastructure.

Natural and national resource and they had to harvest that and process that and transform it into their national digital intelligence. And so this is a this is what we call sovereign AI. You could imagine almost every single country in the world will eventually recognize this and build out their AI infrastructure.

Benson, you use the word resource, and that makes me think about the energy requirements here. I think the cool it's about how the next generation models will have many orders of magnitude greater compute needs. But how will the energy needs increase And what is the advantage you feel in Nvidia has in that sense, Well.

Benson, you use the word resource, and that makes me think about the energy requirements here. I think the cool it's about how the next generation models will have many orders of magnitude greater compute needs. But how will the energy needs increase And what is the advantage you feel in Nvidia has in that sense, Well.

The most important thing that we do is increase the performance of and increase the performance and efficiency of our next generation. So Blackwell is many times more performance than Hopper at the same level of power used, and so that's energy efficiency, more performance with the same amount of power or same performance at a lower power And that's number one. And the second is using luca cooling. We support air cool we support air cooling, we support liquor cooling, but liqual cooling is a lot more energy efficient, and so so the combination of that, all of that, you're going to get.

The most important thing that we do is increase the performance of and increase the performance and efficiency of our next generation. So Blackwell is many times more performance than Hopper at the same level of power used, and so that's energy efficiency, more performance with the same amount of power or same performance at a lower power And that's number one. And the second is using luca cooling. We support air cool we support air cooling, we support liquor cooling, but liqual cooling is a lot more energy efficient, and so so the combination of that, all of that, you're going to get.

A pretty large, pretty large step up.

A pretty large, pretty large step up.

But the important thing to also realize is that AI doesn't really care where it goes to school, and so increasingly going to see AI be trained somewhere else, have that model come back and be used near the population or even running on your PC or your phone. And so we're going to train large models, but the goal is not to run the large models necessarily all the time. You can, you can surely do that for some of the premium services and the very high value AIS, but it's very likely that these large models would then help to train and teach smaller models.

But the important thing to also realize is that AI doesn't really care where it goes to school, and so increasingly going to see AI be trained somewhere else, have that model come back and be used near the population or even running on your PC or your phone. And so we're going to train large models, but the goal is not to run the large models necessarily all the time. You can, you can surely do that for some of the premium services and the very high value AIS, but it's very likely that these large models would then help to train and teach smaller models.

And what we'll end up doing is have one.

And what we'll end up doing is have one.

Large, few large models that are able to train a whole bunch of small models and they run everywhere.

Large, few large models that are able to train a whole bunch of small models and they run everywhere.

Jensen, you explain clearly that demand to build generative AI product on models or even at the GPU level is greater than current supply in black Vel's case in particular, explain the supply dynamics to me for your products and whether you see an improvement sequentially quarter on quarter or at some point by the end of fiscal year into next year.

Jensen, you explain clearly that demand to build generative AI product on models or even at the GPU level is greater than current supply in black Vel's case in particular, explain the supply dynamics to me for your products and whether you see an improvement sequentially quarter on quarter or at some point by the end of fiscal year into next year.

Well, the fact that.

Well, the fact that.

We're growing would suggest that our supply is improving, and our supply chain is quite large, one of the largest supply chains in the world. We have incredible partners and they're doing a great job supporting us in our growth. As you know, We're one of the fastest growing technology companies in history, and none of that would have been possible without very strong demand but also very strong supply. We're expecting Q three to have more supply than Q two, We're expecting Q four to have more supply than Q three, and we're expecting Q one to have more supply than Q four. And so I think our supply, our supply condition going into next year will be in will be a large improvement over this last year with respect to demand. Blackwell is just such a leap and there's several things that are happening, you know, just the Foundation model makers themselves. The size of the Foundation models are growing from hundreds of billions parameters to trillions of parameters. They're also learning more languages. Instead of just learning human language, they're learning the language of images and sounds and videos, and they're even learning the language of three D graphics. And whenever they are able to learn these languages, they can understand what they see, but they can also generate what they're asked to generate. And so they're learning the language of proteins and chemicals and physics. You know, it could be fluids, and it could be particle physics, and so they're learning all kinds of different languages. They learn the meaning of what we call modalities, but basically learning the languages.

We're growing would suggest that our supply is improving, and our supply chain is quite large, one of the largest supply chains in the world. We have incredible partners and they're doing a great job supporting us in our growth. As you know, We're one of the fastest growing technology companies in history, and none of that would have been possible without very strong demand but also very strong supply. We're expecting Q three to have more supply than Q two, We're expecting Q four to have more supply than Q three, and we're expecting Q one to have more supply than Q four. And so I think our supply, our supply condition going into next year will be in will be a large improvement over this last year with respect to demand. Blackwell is just such a leap and there's several things that are happening, you know, just the Foundation model makers themselves. The size of the Foundation models are growing from hundreds of billions parameters to trillions of parameters. They're also learning more languages. Instead of just learning human language, they're learning the language of images and sounds and videos, and they're even learning the language of three D graphics. And whenever they are able to learn these languages, they can understand what they see, but they can also generate what they're asked to generate. And so they're learning the language of proteins and chemicals and physics. You know, it could be fluids, and it could be particle physics, and so they're learning all kinds of different languages. They learn the meaning of what we call modalities, but basically learning the languages.

And so these models are.

And so these models are.

Growing in size, they're learning from more data, and there are more model makers then there was a year ago. And so the number of model makers have grown substantially because of all these different modalities. And so that's just one, just the frontier model. The foundation model makers themselves haven't really grown tremendously. And then the generative AI mark it has really diversified, you know, beyond the Internet service makers to startups and now enterprises are jumping in. Different countries are jumping in, so the demand has really grown.

Growing in size, they're learning from more data, and there are more model makers then there was a year ago. And so the number of model makers have grown substantially because of all these different modalities. And so that's just one, just the frontier model. The foundation model makers themselves haven't really grown tremendously. And then the generative AI mark it has really diversified, you know, beyond the Internet service makers to startups and now enterprises are jumping in. Different countries are jumping in, so the demand has really grown.

Jensen, I'm sorry to cut you off. I will lose your time soon. You've also diversified. And when I said to our audience you were coming on, I got so many questions. Probably the most common one is what is in Nvidia. We talked about you as a systems vendor, but so many points on in Nvidia GPU cloud, and I want to ask, finally, do you have plans to become literally a cloud compute provider?

Jensen, I'm sorry to cut you off. I will lose your time soon. You've also diversified. And when I said to our audience you were coming on, I got so many questions. Probably the most common one is what is in Nvidia. We talked about you as a systems vendor, but so many points on in Nvidia GPU cloud, and I want to ask, finally, do you have plans to become literally a cloud compute provider?

No.

No.

Our GPU cloud was designed to be the best version of Nvidia cloud that's built within each cloud. Nvidio DGX cloud is built inside GCP, inside Azure, inside AWS, inside OCI, and so we build our clouds within theirs so that we can implement.

Our GPU cloud was designed to be the best version of Nvidia cloud that's built within each cloud. Nvidio DGX cloud is built inside GCP, inside Azure, inside AWS, inside OCI, and so we build our clouds within theirs so that we can implement.

Our best version of our cloud.

Our best version of our cloud.

Work with them to make that cloud, that infrastructure, that AI infrastructure and video infrastructure has performance as great TCO as possible, and so that strategy has worked incredibly well.

Work with them to make that cloud, that infrastructure, that AI infrastructure and video infrastructure has performance as great TCO as possible, and so that strategy has worked incredibly well.

That's in Nvidia CEO Jensen Wong speaking exclusively with Bloomberg Technology co host ed Ludbow. Here The full conversation on the Bloomberg Talks podcast feed, available everywhere you download your podcasts.

That's in Nvidia CEO Jensen Wong speaking exclusively with Bloomberg Technology co host ed Ludbow. Here The full conversation on the Bloomberg Talks podcast feed, available everywhere you download your podcasts.

Coming up, another tech company writing the AI wave, Dell recheck in with the CTO and Chief AI Officer at Dell Technologies. Is that on the other side, you're.

Coming up, another tech company writing the AI wave, Dell recheck in with the CTO and Chief AI Officer at Dell Technologies. Is that on the other side, you're.

Listening to Bloomberg Business Week. This is Bloomberg.

Listening to Bloomberg Business Week. This is Bloomberg.

You're listening to the Bloomberg Business Week podcast. Catch us live weekday afternoons from two to five pm Easter listen on Applecarplay and then brout Auto with a Bloomberg Business act or watch us live on YouTube.

You're listening to the Bloomberg Business Week podcast. Catch us live weekday afternoons from two to five pm Easter listen on Applecarplay and then brout Auto with a Bloomberg Business act or watch us live on YouTube.

So a few weeks ago, Dell announced they are getting leaner cutting jobs as part of a reorganization of its sales team to focus on a new group centered on AI products and services. You may recall earlier this year that CEO Michael Dell talked about aipcs being pretty standard in twenty twenty five.

So a few weeks ago, Dell announced they are getting leaner cutting jobs as part of a reorganization of its sales team to focus on a new group centered on AI products and services. You may recall earlier this year that CEO Michael Dell talked about aipcs being pretty standard in twenty twenty five.

We wanted an update about what Dell is doing for that, Bloomberg News reporter Katie Graifeld and I turned to John Rose. He's CTO and Chief AI Officer at Dell Technologies. We spoke to him before the release of Dell's latest earnings report.

We wanted an update about what Dell is doing for that, Bloomberg News reporter Katie Graifeld and I turned to John Rose. He's CTO and Chief AI Officer at Dell Technologies. We spoke to him before the release of Dell's latest earnings report.

CTOs have been around forever, but chief AI officers, at least at Dell and really in federal government and a few other places started about a year ago.

CTOs have been around forever, but chief AI officers, at least at Dell and really in federal government and a few other places started about a year ago.

You know.

You know.

It was a recognition that this is complex, it's bigger than an IT project. It has to have somebody caring and feeding it because it was a lot of governance and technology coming together. So, you know, for us, it was a year ago that we put the position in place.

It was a recognition that this is complex, it's bigger than an IT project. It has to have somebody caring and feeding it because it was a lot of governance and technology coming together. So, you know, for us, it was a year ago that we put the position in place.

We've talked about the Chief AI Officer in the context of the federal government and the new regulations around federal agencies needing a chief AI officer. Do you think that every company needs a Chief AI Officer?

We've talked about the Chief AI Officer in the context of the federal government and the new regulations around federal agencies needing a chief AI officer. Do you think that every company needs a Chief AI Officer?

I don't I think some companies need chief AI officers if you're doing an at scale kind of AI transformation of a large organization, But if if you're a smaller organization, you need someone who's driving it. But do you need to change your organizational structure? For a company like Dell, we have literally hundreds of not thousands, of projects that are going to happen, and a thirty percent of our work is probably going to shift below.

I don't I think some companies need chief AI officers if you're doing an at scale kind of AI transformation of a large organization, But if if you're a smaller organization, you need someone who's driving it. But do you need to change your organizational structure? For a company like Dell, we have literally hundreds of not thousands, of projects that are going to happen, and a thirty percent of our work is probably going to shift below.

The machine line.

The machine line.

That is a gigantic effort, and having someone in the position from an executive perspective is necessary. But smaller companies are going to depend on companies helping them navigate this rather than creating organizational structures.

That is a gigantic effort, and having someone in the position from an executive perspective is necessary. But smaller companies are going to depend on companies helping them navigate this rather than creating organizational structures.

Well, I'm curious in your role how you juggle timelines because you think about the here and now feels like AI has to be in every single conversation. You have a ton of companies trying to gain a foothold or get up to speed or play catch up. But then you also have this fantastic long term opportunity that we talk about all the time that no one really knows what that holds. So I mean, in your day to day, what does it look like trying to juggle those two.

Well, I'm curious in your role how you juggle timelines because you think about the here and now feels like AI has to be in every single conversation. You have a ton of companies trying to gain a foothold or get up to speed or play catch up. But then you also have this fantastic long term opportunity that we talk about all the time that no one really knows what that holds. So I mean, in your day to day, what does it look like trying to juggle those two.

Yeah.

Yeah.

You know, as a CTO, my job is to make sure we never miss a technology inflection. So things like quantum live under me, and these are long term places. One of the interesting things about AI, though, is that the future is happening now. There is no five year journey that you're going to get to eventually. The disruptions are happening almost weekly, and so this intersection between execution and vision is more pronounced than any other technology I have ever seen. So, you know, good chief AI officer is not just an operational manager. They actually have to be extremely current with the technology and what is emerging so that they don't miss an inflection which could happen in the span of like a matter of weeks.

You know, as a CTO, my job is to make sure we never miss a technology inflection. So things like quantum live under me, and these are long term places. One of the interesting things about AI, though, is that the future is happening now. There is no five year journey that you're going to get to eventually. The disruptions are happening almost weekly, and so this intersection between execution and vision is more pronounced than any other technology I have ever seen. So, you know, good chief AI officer is not just an operational manager. They actually have to be extremely current with the technology and what is emerging so that they don't miss an inflection which could happen in the span of like a matter of weeks.

I'm trying to figure out the way to ask this, to contextualize it in the right way. But when we talk to people about AI, John, we oftentimes hear them say it's just like the Internet in the nineteen nineties. You're saying no, no, on the scale of impact it is.

I'm trying to figure out the way to ask this, to contextualize it in the right way. But when we talk to people about AI, John, we oftentimes hear them say it's just like the Internet in the nineteen nineties. You're saying no, no, on the scale of impact it is.

But I was in creating that that year.

But I was in creating that that year.

I was the CTL a bunch of networking companies in the nineties, and we had a decade and a half to kind of get there, and it happened slowly over time, it did have a profound impact. I think from a scale perspective it's similar. But the speed in which this is happening, think about it, you know, let me give you an example. It would be technical for a second. There's a term out there called retrieval augmented generation RAT, which now is in the normal of vernaculars.

I was the CTL a bunch of networking companies in the nineties, and we had a decade and a half to kind of get there, and it happened slowly over time, it did have a profound impact. I think from a scale perspective it's similar. But the speed in which this is happening, think about it, you know, let me give you an example. It would be technical for a second. There's a term out there called retrieval augmented generation RAT, which now is in the normal of vernaculars.

Adie and Univer say that to me. I say it all the time.

Adie and Univer say that to me. I say it all the time.

Yeah, we were just talking about.

Yeah, we were just talking about.

In the commercial ok so common term very important.

In the commercial ok so common term very important.

Technology allows you to take your external data and use it with a large language model without the large language model absorbing it.

Technology allows you to take your external data and use it with a large language model without the large language model absorbing it.

That's pretty important, right.

That's pretty important, right.

That term, as a discussion in our industry is about a year old. It's foundational to almost every enterprise AI story that you can imagine. The technology didn't exists three years ago. So the speed in which these things are happening, you know, it's not like we've developed the Internet, then we have ten years to figure it out and mature it and then scale it and twenty years later everybody has to be there. This is if you're not there now, or if you're not moving towards getting there, you're already late. And I haven't seen that happen in any other big technology and FI is it.

That term, as a discussion in our industry is about a year old. It's foundational to almost every enterprise AI story that you can imagine. The technology didn't exists three years ago. So the speed in which these things are happening, you know, it's not like we've developed the Internet, then we have ten years to figure it out and mature it and then scale it and twenty years later everybody has to be there. This is if you're not there now, or if you're not moving towards getting there, you're already late. And I haven't seen that happen in any other big technology and FI is it.

More steam engine like Jamie Diamond talks about Industrial revolution.

More steam engine like Jamie Diamond talks about Industrial revolution.

Yeah, it's after the steam engine existed and was commercially available. What you saw was this massive uptick you had. And it's actually a good analogy because the steam engine was a way to transfer energy, and you need to transfer energy to do all kinds of things, from producing goods and services to doing transportation, and so it became incredibly pervasive.

Yeah, it's after the steam engine existed and was commercially available. What you saw was this massive uptick you had. And it's actually a good analogy because the steam engine was a way to transfer energy, and you need to transfer energy to do all kinds of things, from producing goods and services to doing transportation, and so it became incredibly pervasive.

AI is very similar.

AI is very similar.

Anybody who tries to piginol that it only does chatbots for, you know, searching staff for it only does enterprise applications X, Y, and Z is wrong because it's a foundational technology. It's a different way to distill information into knowledge and wisdom, and you're going to apply it everywhere, much like we did when we moved to the internal combustion engine.

Anybody who tries to piginol that it only does chatbots for, you know, searching staff for it only does enterprise applications X, Y, and Z is wrong because it's a foundational technology. It's a different way to distill information into knowledge and wisdom, and you're going to apply it everywhere, much like we did when we moved to the internal combustion engine.

Well, let's talk about AI as it relates to Dell's business, because of course you've made an aggressive pivot when it comes to the AI server space. How different is the manufacturing and sales versus the traditional server manufacturing.

Well, let's talk about AI as it relates to Dell's business, because of course you've made an aggressive pivot when it comes to the AI server space. How different is the manufacturing and sales versus the traditional server manufacturing.

Yeah, first thing, maybe to Dispelomith, we didn't actually do an aggressive pivot, meaning about four or five years ago, we built our first AI service Wait before Jenai, and the reason for it was that there were deep neural networks, there were early AI systems. Nobody understood them except maybe one hundred thousand people or smart enough, and so we had been building these systems for quite a while and predicting that there would be a time where this would explode, this would be something that would be bigger. So when CHATGPD happened and suddenly Jennai was available, we had this large language model explosion. We actually already had products that were ready to go into the market, and that's one of the advantages we had as this market formed. We were actually already there. And then what we had to do is adapt a little bit. But our supply chain was in place, our service organization was in place. We had to kind of optimize who we were selling to because it was new customers, but it wasn't really a gigantic surprise. The other thing that we have now included is that there isn't significant difference between the two markets in AI that we participate in.

Yeah, first thing, maybe to Dispelomith, we didn't actually do an aggressive pivot, meaning about four or five years ago, we built our first AI service Wait before Jenai, and the reason for it was that there were deep neural networks, there were early AI systems. Nobody understood them except maybe one hundred thousand people or smart enough, and so we had been building these systems for quite a while and predicting that there would be a time where this would explode, this would be something that would be bigger. So when CHATGPD happened and suddenly Jennai was available, we had this large language model explosion. We actually already had products that were ready to go into the market, and that's one of the advantages we had as this market formed. We were actually already there. And then what we had to do is adapt a little bit. But our supply chain was in place, our service organization was in place. We had to kind of optimize who we were selling to because it was new customers, but it wasn't really a gigantic surprise. The other thing that we have now included is that there isn't significant difference between the two markets in AI that we participate in.

The first is the training market.

The first is the training market.

These large scale deals that we tend to be winning a lot of and are very interesting. And then the enterprise, which is actually not a training business, it's more of an inference business. It turns out that the products that we're building in the training site at scale transfer almost literally into the enterprise unmodified, which is a very good thing because we can address two markets with a portfolio.

These large scale deals that we tend to be winning a lot of and are very interesting. And then the enterprise, which is actually not a training business, it's more of an inference business. It turns out that the products that we're building in the training site at scale transfer almost literally into the enterprise unmodified, which is a very good thing because we can address two markets with a portfolio.

Hey, John, one thing that we're really interested in is understanding who your AI server customers are today. We did see Michael Dell's tweet on shipping too Musk's XAI initiative. How would you define your AI server customers? Is its second tier cloud that's accounting for the growth. Where are the traditional enterprise when it's coming to AI adoption.

Hey, John, one thing that we're really interested in is understanding who your AI server customers are today. We did see Michael Dell's tweet on shipping too Musk's XAI initiative. How would you define your AI server customers? Is its second tier cloud that's accounting for the growth. Where are the traditional enterprise when it's coming to AI adoption.

Well, first, there's two very distinct markets. There is the traditional enterprise, which is still an early market. It's mostly experimentation. Those experiments are now graduating and giving an example internally, we have a project that was first an experiment around our service organization, creating some that would allow people to make intelligent choices about troubleshooting a problem. We then very quickly move from a prototype to thousands of people, and we're on a path the potentially forty thousand of our service technicians using this technology and a span of literally measured in six.

Well, first, there's two very distinct markets. There is the traditional enterprise, which is still an early market. It's mostly experimentation. Those experiments are now graduating and giving an example internally, we have a project that was first an experiment around our service organization, creating some that would allow people to make intelligent choices about troubleshooting a problem. We then very quickly move from a prototype to thousands of people, and we're on a path the potentially forty thousand of our service technicians using this technology and a span of literally measured in six.

Months or so. So the enterprise is on that journey.

Months or so. So the enterprise is on that journey.

We haven't really hit scale in any meaningful way, but we have a lot of things in the pipeline as people are figuring it out. That market, i would say, is my view, is a bigger market because there are millions of enterprises and every process and every enterprise is a target for this technology. But it takes time to get to a point of stability, to transfer the knowledge and too basically get moving great future. The today market is mostly characterized by the ones you mentioned. It's companies that are building their own foundation models. They are training large scale models. That's not what enterprises do. Enterprises, you use those models to do inference tasks. The training business is growing dramatically because quite frankly, that's the underlying engine of all layout.

We haven't really hit scale in any meaningful way, but we have a lot of things in the pipeline as people are figuring it out. That market, i would say, is my view, is a bigger market because there are millions of enterprises and every process and every enterprise is a target for this technology. But it takes time to get to a point of stability, to transfer the knowledge and too basically get moving great future. The today market is mostly characterized by the ones you mentioned. It's companies that are building their own foundation models. They are training large scale models. That's not what enterprises do. Enterprises, you use those models to do inference tasks. The training business is growing dramatically because quite frankly, that's the underlying engine of all layout.

But is it only a handful of companies that are doing it.

But is it only a handful of companies that are doing it.

It's not a very big market in terms of compared to the tens of millions of enterprises that are going to use size that are going to use it. The people that are going to produce these models is not anywhere near that number. There's more than a handful of them. There are certain tiers of them, as you mentioned, there are even certain enterprise, very large enterprises, Xtra being a good example. You're talking about the Yeah, they're also training their own models. But the thing that makes them all the same is every one of those companies is building infrastructure to train large scale foundation models. Enterprises aren't. Enterprises are using those models, putting them into production, maybe fine tuning them, and so there's a very different market forming, but they're happening kind of a little out of phase, but in parallel.

It's not a very big market in terms of compared to the tens of millions of enterprises that are going to use size that are going to use it. The people that are going to produce these models is not anywhere near that number. There's more than a handful of them. There are certain tiers of them, as you mentioned, there are even certain enterprise, very large enterprises, Xtra being a good example. You're talking about the Yeah, they're also training their own models. But the thing that makes them all the same is every one of those companies is building infrastructure to train large scale foundation models. Enterprises aren't. Enterprises are using those models, putting them into production, maybe fine tuning them, and so there's a very different market forming, but they're happening kind of a little out of phase, but in parallel.

When it comes to competitive landscape, what does Dell offer that a super micro or even in Nvidia doesn't.

When it comes to competitive landscape, what does Dell offer that a super micro or even in Nvidia doesn't.

Yeah, first of all, AI is not just a server transaction. There is storage networking, but more importantly, it's especially when you go to the enterprise, it's end to end. When you're talking about deploying enterprise AI, you have to put it where your people in data are, which includes the AIPC, edges, distributed architectures. And that's why you know when training may be able to consume servers, just as a standard component, enterprise has to consume systems and more and more the training infrastructures are consuming systems and that's why our expertise and networking and storage and data center design, power cooling, all the things that we do because we're very broad come together and give us an advantage in both of those markets. More pronounced in the enterprise side.

Yeah, first of all, AI is not just a server transaction. There is storage networking, but more importantly, it's especially when you go to the enterprise, it's end to end. When you're talking about deploying enterprise AI, you have to put it where your people in data are, which includes the AIPC, edges, distributed architectures. And that's why you know when training may be able to consume servers, just as a standard component, enterprise has to consume systems and more and more the training infrastructures are consuming systems and that's why our expertise and networking and storage and data center design, power cooling, all the things that we do because we're very broad come together and give us an advantage in both of those markets. More pronounced in the enterprise side.

We talk a lot about power consumption and with power consumption comes great responsibility. But actually that's not really the quote heat generation. That's what comes with power consumption. What does Dell offer to help lessen AI's power.

We talk a lot about power consumption and with power consumption comes great responsibility. But actually that's not really the quote heat generation. That's what comes with power consumption. What does Dell offer to help lessen AI's power.

Use, Well, well, the first thing is, you know, okay, first fact is when you add a whole pile of compute, you have to use energy to do that. And yes, AI is a significant increase in the overall power consumption because we've increased the compute capacity of the world by potentially orders of magnitude. The levers we have to deal with that are quite numerous. The biggest one that's going on right now is we're moving to directly with cooling in a more pervasive way, and that is a much more efficient way of getting the energy out of these systems and basically dealing with cooling. But that requires a retrofitive infrastructure and investment, but it turns out it does work quite well. We've been doing direct liquid cooling for I think twenty years or alienware products and the gaming side have been doing this for it. You know, you can't buy one now that doesn't have direct liquid cooling. So it's a space that's important. But beyond that, there are other decisions you make.

Use, Well, well, the first thing is, you know, okay, first fact is when you add a whole pile of compute, you have to use energy to do that. And yes, AI is a significant increase in the overall power consumption because we've increased the compute capacity of the world by potentially orders of magnitude. The levers we have to deal with that are quite numerous. The biggest one that's going on right now is we're moving to directly with cooling in a more pervasive way, and that is a much more efficient way of getting the energy out of these systems and basically dealing with cooling. But that requires a retrofitive infrastructure and investment, but it turns out it does work quite well. We've been doing direct liquid cooling for I think twenty years or alienware products and the gaming side have been doing this for it. You know, you can't buy one now that doesn't have direct liquid cooling. So it's a space that's important. But beyond that, there are other decisions you make.

You know.

You know.

One of the most interesting ones that we're exploring, and this is more for the enterprise side, is once you choose to put your AI into production, there are two places and an enterprise that you can run it. You could run it in a data center, which is what people are starting with. Or there's this little thing called an AIPC emerging, which says, what if I take that model, I re quantize it. I basically make it small enough to run locally, and I move it out to the computing device onto the PC itself.

One of the most interesting ones that we're exploring, and this is more for the enterprise side, is once you choose to put your AI into production, there are two places and an enterprise that you can run it. You could run it in a data center, which is what people are starting with. Or there's this little thing called an AIPC emerging, which says, what if I take that model, I re quantize it. I basically make it small enough to run locally, and I move it out to the computing device onto the PC itself.

Doesn't that still use the energy, but it just uses it locally rather than the energy is still.

Doesn't that still use the energy, but it just uses it locally rather than the energy is still.

Being use but the energy was going to be used anyway. Those PCs one, they're distributed, so they're not sitting at one choke point on the grid. I don't need to put megawatts to a PIEC. I have AC outlets everywhere. I have batteries on them, and in fact, the total number of mips in the world sitting on PCs is probably bigger than the total amount of mips in a data center MIPS million instructions per second, the measure of compute performance. And so the reality is, you know, look at a typical AIPC today and it's a relatively small amount of NPU capacity neural processing unit capacity, but if you multiply it by a billion PCs, it's a pretty big footprint to go after.

Being use but the energy was going to be used anyway. Those PCs one, they're distributed, so they're not sitting at one choke point on the grid. I don't need to put megawatts to a PIEC. I have AC outlets everywhere. I have batteries on them, and in fact, the total number of mips in the world sitting on PCs is probably bigger than the total amount of mips in a data center MIPS million instructions per second, the measure of compute performance. And so the reality is, you know, look at a typical AIPC today and it's a relatively small amount of NPU capacity neural processing unit capacity, but if you multiply it by a billion PCs, it's a pretty big footprint to go after.

That was John Rose, chief technology officer and also chief AI officer at Dell Technologies. Bloomberg's Katy Greifeld joining there as well.

That was John Rose, chief technology officer and also chief AI officer at Dell Technologies. Bloomberg's Katy Greifeld joining there as well.

When we come back more of that conversation with Dell's John Rose, including what makes a PC an AI PC?

When we come back more of that conversation with Dell's John Rose, including what makes a PC an AI PC?

A few years ago we started looking and saying, well, could we build an optimized processor that was really very power efficient to do AI task that's called a neural processing unit. And so when we just announced whole bunch of Qualcom based aipcs, and we have Intel based aipcs, and the thing that makes them an AIPC in most cases is it has all three of those processors, and.

A few years ago we started looking and saying, well, could we build an optimized processor that was really very power efficient to do AI task that's called a neural processing unit. And so when we just announced whole bunch of Qualcom based aipcs, and we have Intel based aipcs, and the thing that makes them an AIPC in most cases is it has all three of those processors, and.

Well, I learns to speak more like humans. Quite a few humans are using AI to speak to other humans.

Well, I learns to speak more like humans. Quite a few humans are using AI to speak to other humans.

Nice in our US product or English speakers learning Spanish. We have rolled that out already and people are very engaged with it. It doesn't judge you. I mean, it gives you feedback, but you're not embarrassed in front of the AI.

Nice in our US product or English speakers learning Spanish. We have rolled that out already and people are very engaged with it. It doesn't judge you. I mean, it gives you feedback, but you're not embarrassed in front of the AI.

You know, it's very effective.

You know, it's very effective.

The CEO of the Language learning at Babble. On the other side, this is Boom.

The CEO of the Language learning at Babble. On the other side, this is Boom.

You're listening to the Bloomberg Business Week podcast. Listen live each weekday starting at two pm Eastern on Apple car Play and Android Auto with the Bloomberg Business App. You can also listen live on Amazon Alexa from our flagship New York station just Say Alexa playing Bloomberg eleven thirty.

You're listening to the Bloomberg Business Week podcast. Listen live each weekday starting at two pm Eastern on Apple car Play and Android Auto with the Bloomberg Business App. You can also listen live on Amazon Alexa from our flagship New York station just Say Alexa playing Bloomberg eleven thirty.

A couple of companies are betting that we're going to need more powerful computers to run programs at home AI. That kind of makes sense. Dell is one of those companies. As we mentioned earlier, CEO Michael Dell has talked about aipcs being pretty standard in twenty twenty five, but I feel like Tim, we're still trying to figure out what exactly does that mean exactly?

A couple of companies are betting that we're going to need more powerful computers to run programs at home AI. That kind of makes sense. Dell is one of those companies. As we mentioned earlier, CEO Michael Dell has talked about aipcs being pretty standard in twenty twenty five, but I feel like Tim, we're still trying to figure out what exactly does that mean exactly?

So what makes a PC an AIPC. It's a question that I posed to John Rose, chief technology and chief AI officer at Dell Technologies. Here's the rest of our conversation along with Bloomberg's Katie Creifeld.

So what makes a PC an AIPC. It's a question that I posed to John Rose, chief technology and chief AI officer at Dell Technologies. Here's the rest of our conversation along with Bloomberg's Katie Creifeld.

Today, the definition is that it has a neural processing unit. So technical for a second. You know, all PCs have central processing units X eighty six processors. That's what most of us run our stuff on. And then years ago we put graphic processing units GPUs, which now in video is quite popular with the in the data center. Same architecture sits on almost every PC, just smaller scale. And then a few years ago we started looking and saying, well, could we build an optimized processor that was really very power efficient to do AI task. That's called a neural processing unit. And so the latest you know, Intel processors, Qualcom processers actually have all three on the system. And we just announced whole bunch of Qualcom based aipcs, and we have Intel based aipcs, and the thing that makes them an AIPC in most cases is it has all three of those processers.

Today, the definition is that it has a neural processing unit. So technical for a second. You know, all PCs have central processing units X eighty six processors. That's what most of us run our stuff on. And then years ago we put graphic processing units GPUs, which now in video is quite popular with the in the data center. Same architecture sits on almost every PC, just smaller scale. And then a few years ago we started looking and saying, well, could we build an optimized processor that was really very power efficient to do AI task. That's called a neural processing unit. And so the latest you know, Intel processors, Qualcom processers actually have all three on the system. And we just announced whole bunch of Qualcom based aipcs, and we have Intel based aipcs, and the thing that makes them an AIPC in most cases is it has all three of those processers.

You put them all.

You put them all.

Together and you end up with an actually pretty beefy system that still is low power runs on a battery can take away with you. The trick is which we haven't done as an industry yet, is when we put the AIS into production, moving more of those workloads out onto those PCs moves them out of the data center. You've shifted the work, you've shifted the power and cooling, You've distributed it into a place that's ready for it, and that buys us a significant amount of capacity for the things that truly belong in.

Together and you end up with an actually pretty beefy system that still is low power runs on a battery can take away with you. The trick is which we haven't done as an industry yet, is when we put the AIS into production, moving more of those workloads out onto those PCs moves them out of the data center. You've shifted the work, you've shifted the power and cooling, You've distributed it into a place that's ready for it, and that buys us a significant amount of capacity for the things that truly belong in.

A data center.

A data center.

We're going to do both, but today most people only do one of them, and they're not taking advantage of the billion PCs out there that could in fact be doing this work instead of.

We're going to do both, but today most people only do one of them, and they're not taking advantage of the billion PCs out there that could in fact be doing this work instead of.

In the data center.

In the data center.

Well, I'm glad you brought us to aipcs because that's where I wanted to go, because Dell has been out front talking about aipcs really starting helping to start the conversation there. I think there's some stat out there that says that forty percent of new PCs shipped in twenty twenty five will be aipcs. Does that sound realistic to you? What are you modeling?

Well, I'm glad you brought us to aipcs because that's where I wanted to go, because Dell has been out front talking about aipcs really starting helping to start the conversation there. I think there's some stat out there that says that forty percent of new PCs shipped in twenty twenty five will be aipcs. Does that sound realistic to you? What are you modeling?

I mean, if you look at the roadmap from Intel and definitely Qualcomm and others, they're all moving in this direction. You know, most of the latest Intel chips are absolutely you know the ones that are coming in, the ones they're in the market are qualified as AAIPC is the new Qualcom chip sets that are now in the market on our products and starting to roll out definitely AI. So it's not a function of you know, wanting to do something special. It's actually becoming mainstream. There will be non AIPC chipsets out there because not everybody needs this. But in the business context, if you're going to put a copilot out on a PC, if you're going to run that troubleshooting tool for your support group, if I can do that on an AIPC, the net impact is that it will be able to do it more efficiently. The NPU just uses less power for the same task, and so you end up with a much more efficient, much more performance system. Microsoft has released a whole pile of copilot stuff that runs locally on It's called Copilot Plus and includes image replacement and prediction summarization. And they realize a long time ago that pushing a bunch of this stuff to be processed locally has a net effect from a sustainability environmental impact of getting things into a place that already has the power to eliminate the burden on some of the data center architectures.

I mean, if you look at the roadmap from Intel and definitely Qualcomm and others, they're all moving in this direction. You know, most of the latest Intel chips are absolutely you know the ones that are coming in, the ones they're in the market are qualified as AAIPC is the new Qualcom chip sets that are now in the market on our products and starting to roll out definitely AI. So it's not a function of you know, wanting to do something special. It's actually becoming mainstream. There will be non AIPC chipsets out there because not everybody needs this. But in the business context, if you're going to put a copilot out on a PC, if you're going to run that troubleshooting tool for your support group, if I can do that on an AIPC, the net impact is that it will be able to do it more efficiently. The NPU just uses less power for the same task, and so you end up with a much more efficient, much more performance system. Microsoft has released a whole pile of copilot stuff that runs locally on It's called Copilot Plus and includes image replacement and prediction summarization. And they realize a long time ago that pushing a bunch of this stuff to be processed locally has a net effect from a sustainability environmental impact of getting things into a place that already has the power to eliminate the burden on some of the data center architectures.

Okay, so what is a real use case for an AIPC. Yeah, well, I mean there's customers figured it out.

Okay, so what is a real use case for an AIPC. Yeah, well, I mean there's customers figured it out.

They are figuring it out.

They are figuring it out.

The biggest one that's very visible right now is if you run the copilot plus technologies from Microsoft, which it's their kind of end to end stack, when you move them out onto the PC one, it always works with you too. It generally tends to be more private data because the data doesn't move out into a cloud. And number three, the biggest impact is just runs more efficiently. The non Microsoft examples are already emerging. You have people like Adobe and others that are starting to think about how can I run my web app so that the processing is done locally to do background insertion. We have stuff that we built internally for our own uses around this troubleshooting tool and technology to help get a large language model powered AI engine underneath all my support technicians. Guess what, Dell support technicians rarely spend time in our data center. They're out in your data center, and they got to bring those tools with them, and they might not have network access there because of security reasons. If I can bring the AI with me, then I guarantee a high quality of service. There's a lot of those, and on and on. Member AI is horizontal. We haven't figured out all the use cases. Just ask yourself, where do you want to bring intelligence that augments a human being in real time efficiently? And that's where you need an AIPC shoulder.

The biggest one that's very visible right now is if you run the copilot plus technologies from Microsoft, which it's their kind of end to end stack, when you move them out onto the PC one, it always works with you too. It generally tends to be more private data because the data doesn't move out into a cloud. And number three, the biggest impact is just runs more efficiently. The non Microsoft examples are already emerging. You have people like Adobe and others that are starting to think about how can I run my web app so that the processing is done locally to do background insertion. We have stuff that we built internally for our own uses around this troubleshooting tool and technology to help get a large language model powered AI engine underneath all my support technicians. Guess what, Dell support technicians rarely spend time in our data center. They're out in your data center, and they got to bring those tools with them, and they might not have network access there because of security reasons. If I can bring the AI with me, then I guarantee a high quality of service. There's a lot of those, and on and on. Member AI is horizontal. We haven't figured out all the use cases. Just ask yourself, where do you want to bring intelligence that augments a human being in real time efficiently? And that's where you need an AIPC shoulder.

Yeah, just telling me what to do.

Yeah, just telling me what to do.

Can we talk about pricing here though?

Can we talk about pricing here though?

When it comes to.

When it comes to.

An AI PC, how do you price that relative to a normal PC?

An AI PC, how do you price that relative to a normal PC?

Yeah?

Yeah?

I'm not a pricing guy, so I have no idea what the you know, where the market will land. What we do believe is there's more value in an AIPC if you have this ability to do these tasks in a more efficient way and are performant way, and you can't do them on an alternative PC, Clearly that has value. And we already have lots of pricing bands within our industry. You know, there's a huge difference between an alienware you know, gaming machine and a chromebook. They are both technically PCs. One of them you can run high performance gaming and the other one you can't. And so we'll see that kind of stratification occur, but these are not things that are exotic and out of the realm. In fact, our newest products of the qualcom ones were just released. If you look at them, they're pretty mainstream pricing. These are not the old days of you spend five thousand dollars for a laptop to do something special. It's kind of within the realm of normal pricing, if you will. But it is a value play in the sense that does something that quite frankly, you know, is better than if it wasn't there.

I'm not a pricing guy, so I have no idea what the you know, where the market will land. What we do believe is there's more value in an AIPC if you have this ability to do these tasks in a more efficient way and are performant way, and you can't do them on an alternative PC, Clearly that has value. And we already have lots of pricing bands within our industry. You know, there's a huge difference between an alienware you know, gaming machine and a chromebook. They are both technically PCs. One of them you can run high performance gaming and the other one you can't. And so we'll see that kind of stratification occur, but these are not things that are exotic and out of the realm. In fact, our newest products of the qualcom ones were just released. If you look at them, they're pretty mainstream pricing. These are not the old days of you spend five thousand dollars for a laptop to do something special. It's kind of within the realm of normal pricing, if you will. But it is a value play in the sense that does something that quite frankly, you know, is better than if it wasn't there.

Hey, John, very very briefly, when you're thinking about money you're spending, where's the investment going thirty seconds?

Hey, John, very very briefly, when you're thinking about money you're spending, where's the investment going thirty seconds?

Investment in aim? Well, well, I mean significantly two things. One, obviously, from a product line perspective, we are investing heavily on building out the products and services and technologies and to enable our customers to build out their AI environment. And that's AI enabling and evolving our entire portfolio.

Investment in aim? Well, well, I mean significantly two things. One, obviously, from a product line perspective, we are investing heavily on building out the products and services and technologies and to enable our customers to build out their AI environment. And that's AI enabling and evolving our entire portfolio.

That was John Rose, chief Technology Officer and Chief AI Officer at Dell Technology. It's a big thank you to to Katie Greifeld of Bloomberg TV and Bloomberg News for joining me for that conversation. Are you ready for your PC to be an AI PC?

That was John Rose, chief Technology Officer and Chief AI Officer at Dell Technology. It's a big thank you to to Katie Greifeld of Bloomberg TV and Bloomberg News for joining me for that conversation. Are you ready for your PC to be an AI PC?

I'm still not sure. I feel like there's still so much I need to learn. I will say, I think a lot of stuff we're gonna be doing is on our phones.

I'm still not sure. I feel like there's still so much I need to learn. I will say, I think a lot of stuff we're gonna be doing is on our phones.

But I don't know.

But I don't know.

Maybe we'll see I don't know. I think about like schools too, Like when will they adopt AIPCS in schools?

Maybe we'll see I don't know. I think about like schools too, Like when will they adopt AIPCS in schools?

Yeah, many kids across the country already back in school after summer break. Can you believe it?

Yeah, many kids across the country already back in school after summer break. Can you believe it?

That is crazy? Like I just I don't know. I can feel it though, and like this city's winding down, the city's winding on.

That is crazy? Like I just I don't know. I can feel it though, and like this city's winding down, the city's winding on.

We do it a little differently during New York. Yeah, we start later after Labor Day New Jersey to.

We do it a little differently during New York. Yeah, we start later after Labor Day New Jersey to.

Wait, California, Did you grow up that way?

Wait, California, Did you grow up that way?

I don't even remember, Okay we didn't.

I don't even remember, Okay we didn't.

We were after Labor Day?

We were after Labor Day?

Well in New Jerseys.

Well in New Jerseys.

Yeah.

Yeah.

Yeah. Many of the kids, when they go back to school, or maybe they're back in school, they'll take language classes. Maybe they'll learn Spanish, Mandarin, French, or they'll do what I did. What's that?

Yeah. Many of the kids, when they go back to school, or maybe they're back in school, they'll take language classes. Maybe they'll learn Spanish, Mandarin, French, or they'll do what I did. What's that?

Nothing?

Nothing?

Take Latin? Yeah, not the most practical of languages when it comes to world Just saying, well you don't have to be in school to learn a new language. There's no shortage of apps and services out there to help you do that. Do a lingo or is that a stone? Memorize and more such as Babel. Back with us as Julie Hansen, us CEO of Babbel, she joins us from New York. Julie, good to have you with us. I mentioned a few names there, some of them publicly traded, including dual Lingo. How do you differentiate given that when you google language learning apps there's like five up there that are just sponsored results. There's a lot of competition in your space. How do you do it?

Take Latin? Yeah, not the most practical of languages when it comes to world Just saying, well you don't have to be in school to learn a new language. There's no shortage of apps and services out there to help you do that. Do a lingo or is that a stone? Memorize and more such as Babel. Back with us as Julie Hansen, us CEO of Babbel, she joins us from New York. Julie, good to have you with us. I mentioned a few names there, some of them publicly traded, including dual Lingo. How do you differentiate given that when you google language learning apps there's like five up there that are just sponsored results. There's a lot of competition in your space. How do you do it?

It is a competitive space, but I think there really are two big players in the market, and Babel is one of them. Our differentiation is all about our pedagogy, but I can use that term or you know, the kind of the way we approach the learning, the seriousness of it and the effectiveness of it. You know, we have studies that have shown that Babel is more effective than others in actually learning a language, but a lot comes down to also your personal style, the way you learn. Everyone learns differently, so we try to make sure we reflect all the different ways of learning in the app, writing, reading, speaking, of course the most important Hey, how do.

It is a competitive space, but I think there really are two big players in the market, and Babel is one of them. Our differentiation is all about our pedagogy, but I can use that term or you know, the kind of the way we approach the learning, the seriousness of it and the effectiveness of it. You know, we have studies that have shown that Babel is more effective than others in actually learning a language, but a lot comes down to also your personal style, the way you learn. Everyone learns differently, so we try to make sure we reflect all the different ways of learning in the app, writing, reading, speaking, of course the most important Hey, how do.

You measure success? Because I think about we talked with you, wasn't it around like the New Year? And we talked about how people in the New Year sometimes are like I want to learn to do this or that, I want to lose weight, I want.

You measure success? Because I think about we talked with you, wasn't it around like the New Year? And we talked about how people in the New Year sometimes are like I want to learn to do this or that, I want to lose weight, I want.

To learn a language.

To learn a language.

So I am curious, how do you guys measure success that people actually have become have learned the language and can converse and are fluent.

So I am curious, how do you guys measure success that people actually have become have learned the language and can converse and are fluent.

Even, Yeah, that is a great question, and I can assure you we spend so much time on that topic. There's really two main ways to think about it. One is against the safer the common European frame of reference, which is kind of a formal way of evaluating language abilities, and that is that's very common in Europe. Everyone knows it. But even here in the US that framework is understood and used in some circles, and so we measure people's advancement against it. But at the same time, the most important thing and what we're trying to do in battle because we know will never make you fluent, No app will make you fluent, So we're trying to help you get comfortable having a conversation. So we actually ask our users to evaluate subjectively personally their view of their progress, and that's for us like their own personal learner. Success rating is our number one way of looking at it, because at the end of the day, if you're you know, a B one on the C for scale but you can't speak, then are you really making the progress you want? Or if you're a sloppy a something, but you're you're getting out there and making conversation, like, that's really more success in many ways, Julie.

Even, Yeah, that is a great question, and I can assure you we spend so much time on that topic. There's really two main ways to think about it. One is against the safer the common European frame of reference, which is kind of a formal way of evaluating language abilities, and that is that's very common in Europe. Everyone knows it. But even here in the US that framework is understood and used in some circles, and so we measure people's advancement against it. But at the same time, the most important thing and what we're trying to do in battle because we know will never make you fluent, No app will make you fluent, So we're trying to help you get comfortable having a conversation. So we actually ask our users to evaluate subjectively personally their view of their progress, and that's for us like their own personal learner. Success rating is our number one way of looking at it, because at the end of the day, if you're you know, a B one on the C for scale but you can't speak, then are you really making the progress you want? Or if you're a sloppy a something, but you're you're getting out there and making conversation, like, that's really more success in many ways, Julie.

So those two pays you.

So those two pays you.

Said that no app will will make you fluent, but can developments with AI and perhaps maybe a conversation partner that is a bot eventually make you fluent, Like, do you think that no app will ever make you fluent or are we on a sort of pace to get to that at some point in the near future.

Said that no app will will make you fluent, but can developments with AI and perhaps maybe a conversation partner that is a bot eventually make you fluent, Like, do you think that no app will ever make you fluent or are we on a sort of pace to get to that at some point in the near future.

Well, fluence is a very high bar. Fluent is you know, live there, speak it like in.

Well, fluence is a very high bar. Fluent is you know, live there, speak it like in.

That language, right, Yeah, that's what they always say.

That language, right, Yeah, that's what they always say.

Exactly exactly.

Exactly exactly.

And when I got to that point in college when I dreamt in French like, it was incredibly thrilling.

And when I got to that point in college when I dreamt in French like, it was incredibly thrilling.

It's long gone, by the.

It's long gone, by the.

Way, I know a place where you can work on it.

Way, I know a place where you can work on it.

We completely agree with you that a chat bot can serve that purpose. In fact, in our US product for English speakers learning Spanish, we have rolled that out already and people are very engaged with it. It doesn't judge you. I mean, it gives you feedback, but you're not embarrassed in front of the AI. You know, it's very effective. We didn't, by the way, just throw an LLM in the app and say have at it. Like we had a kind of an expert train it for a considerable period of time and that's what makes it successful. But absolutely that is going to help more and more.

We completely agree with you that a chat bot can serve that purpose. In fact, in our US product for English speakers learning Spanish, we have rolled that out already and people are very engaged with it. It doesn't judge you. I mean, it gives you feedback, but you're not embarrassed in front of the AI. You know, it's very effective. We didn't, by the way, just throw an LLM in the app and say have at it. Like we had a kind of an expert train it for a considerable period of time and that's what makes it successful. But absolutely that is going to help more and more.

What's the demographics. Is it a lot of people maybe younger starting jobs or what is it? Is there some I'm always curious about, like the folks to use in your platform.

What's the demographics. Is it a lot of people maybe younger starting jobs or what is it? Is there some I'm always curious about, like the folks to use in your platform.

It is a mix. Honestly, it's not school kids. We are not aimed at that market for various reasons. In many ways, our best users are older, you know, over call it forty five, because they have time. You know, they have time in their lives to pursue this. But it's really all ages. Remember that in the US, you know, in grade school, only twenty percent of students ever take a language.