Exclusive Coverage: theCUBE at IBM Think | Zeus Kerravala | ZK Research

theCUBE.net

Clips
More from Exclusive Coverage: theCUBE at IBM Think

Zeus Kerravala

Founder & Principal Analyst

ZK Research

play_circle_outline The Crucial Role of Data in AI Integration and the Evolving Significance of Data Control in Hybrid Cloud

play_circle_outline The potential for AI as a service and the value of observability and networking

play_circle_outline IBM's focus on AI and the partnership approach

play_circle_outline Unleashing IBM's Infrastructure and AI: Regaining Lost Ground while Simplifying Enterprise AI Solutions

Info

Zeus Kerravala | ZK Research

Zeus Kerravala

Founder & Principal Analyst ZK Research

search

>> Hello. Welcome to theCUBE's pop-up coverage. CUBE pop-up, that's what this is. We had a spot here, here at IBM Think. We are in the Omni Hotel right across the street from the Boston Convention Center. IBM think will be kicking off. It actually kicked off today. Now the sessions are here with the partners. Tomorrow, on Tuesday will be IBM Think's main program. I'm John Furrier with theCUBE. Your host Dave Vellante is out in Vegas for Dell Tech World. I'm here with Zeus Kerravala of ZK Research, CUBE friend, CUBE contributor. He's also got a very busy conference. This is conference season. You got here, you got Zoom, you got this.

Zeus Kerravala

>> Yeah, it is-

>> Good to see you.

Zeus Kerravala

>> Yeah, from here to the end of June, it's conference season.

>> Well thanks for coming on. I want to break down the keynote opening. Arvind was up on stage. You had a series of presentations, Dario Gill and then Rob Thomas, all the top executives. Of course they sprinkled in Partner of the Year. Everyone gets a trophy, you know how it is in enterprise tech. But the real story to me, and I just tweeted this out, is that IBM's back. The Once-Lost IBM now is back under one banner. Watson and Red Hat really together wedging in as the lead pillars. And everything wrapping around it, they got a huge portfolio, but AI is a gift. There's a platform shift happening and IBM is taking advantage of it. And we've been documenting this on theCUBE, you and I have been with Dave, and we've been researching and writing stories about. But AI is kind of a gift for IBM. They had Watson originally winning the Jeopardy thing and the NLP world. Now we're in a whole other ballgame with vectors and embeds and RAG and all kinds of foundation models. IBM's looking good, Zeus. What's your takeaway from the keynote?

Zeus Kerravala

>> Yeah. They were early out of the gates with AI, but then they kind of went to sleep and let everybody else take the mantle here for AI. And I think they are back, and it is a bit of a different IBM. When you think of IBM, I think a lot of people think of all the way back to the mainframes, the closed proprietary systems. And they worked well because they were walled gardens. And if you bought everything IBM, it would work well together. It was very clear in this keynote from Kate Woolley through Arvind, they are taking an open approach to AI. I think it was Dario Gill that said, "When you think of what open did for the internet, it's going to have the same impact for AI." And during Kate's presentation, she opened it up with the economic opportunity for AI is $4.4 trillion, of which they want the partners to share in that. And so she gave examples of some partners and some of the things they're doing. But ultimately if IBM is going to reestablish itself as the premier IT vendor, you could argue they've lost a little bit of that, it is going to be through the partnerships that they have. And they gave a lot of really good partner examples, and I think that having Partner Day lead off Think is a great way to do it.

>> And I think the open thing is a key message. The partner network could drive a lot of revenue, and that traditionally has not been an IBM strong point. I know they've had partners, rather than IBM Blue specifically ... When I say IBM Blue, I mean IBM only. Now what they have the open source mojo going on with Red Hat. They got HashiCorp on the other side. Like I said, you got it going on. Now, a couple nuanced points I want to get your reaction to because Dave Vellante brought this up in his breaking analysis. I also just posted it on the Twitter feed here that there's two things going on. One is IBM's portfolio is massive. So we look at the overall company performance from a deep tech perspective. Not every area is going to be impacted, but it is impacting a lot of the key areas. So I want to talk about that in terms of what the key pillars and the pistons of their business engine is. And the other thing that's going on with the foundation models and AI is that the gap between consumer AI and enterprise AI -

Zeus Kerravala

>> Huge....

>> is shrinking very fast. And if you look about the early days of the web, it was at least a couple year lag between consumer internet and then consumer internet in the enterprise, putting intranets aside. There's only a month lag now between consumer and internet. Now all the chips are being bought up by the big players. So ... Okay.

Zeus Kerravala

>> One of the differences though, John, and Rob brought this up in his presentation. The gap is huge because the data isn't there. And I think it was Rob Thomas in his presentation that said only ... Was it his? Or somebody said, only 1% of enterprise data is actually being used to fuel LLMs today. So that's where the big gap is. It's not the models per se, it's the data. And I think that's where companies struggle.

>> Well actually it's a very nuanced point. Let's unpack that because I'm actually saying the opposite. I'm saying that the technology gap between enterprise adoption is accelerating faster than the internet, meaning the language models don't have that end data for the enterprise for a good reason. As the power laws suggest, the long tail of data is actually an opportunity for enterprise. You're seeing that with the RAG, the Retrieval Augmentation Generation. So what's happening now is you're starting to see real production workloads that are very specific that aren't sexy projects. So the adoption on the enterprise is faster than the adoption on the internet is what I'm saying. So what that means is that that's what's happening and that's driving a lot of integration business. So if you're IBM and you have installed base, you have a built-in data business built in with the customer. So then the question is integration. So you're seeing InstructLabs, I am blown away by that, Zeus, because what Dario Gil was saying is what we've been saying on theCUBE. Models can interact with each other, you can train and you can impact a proprietary model on your data. So this means that models will integrate. That's going to be a huge win for IBM because they've been in the consulting business and integration business for years. So in comes Red Hat, in comes the generative AI hype. And guess what? IBM's not just a system integrator, they're a technology partner. So putting an ecosystem together, I just see this thing lining up beautifully.

Zeus Kerravala

>> Well, the services opportunity with AI is huge, because it is, I think easily the most complex system of technology that an enterprise has to put together. And every time we introduce new technology, we raise the complexity. So you think of cloud, hybrid cloud, edge, automation, all these things we've been through. Complexity is at a point where CIOs are telling me it's now at where it's untenable. And so you layer on AI on top of that and it makes the challenge more difficult. I think one of the big issues for businesses is how fast they can get their data into these systems. During this panel, Crawford Del Prete, the CRO of IDC said 75% of enterprise data loses its value in months and 50% in days. And so I think here's the challenge for businesses, as they generate so much data across so many different systems, can they find a way to aggregate it and make it useful from an AI perspective? And every CIO I've talked to about AI talks about data being the number one inhibitor. So if IBM and its services group and its massive partner organization can help companies get a handle on data, then yeah, they are back and I think they're back in a big way.

>> The data is the problem. And just the mind-blowing thought exercise is if you want to generate a new thing with generative AI, because it generates, it's not a pre-programmed thing, you got to have access to data. So if you're out on the edge on a device or a wearable or a machine, you need access to every single piece of data in very low latency. To make that happen, Zeus, is going to require a completely different rethinking of the data modeling, the availability of data, moving data around-

Zeus Kerravala

>> And your infrastructure....

>> and how you're going to store that. So I think what's happening now, and it's clear to us as we start to see that the old way, new way form, it's very obvious. If you can't make data highly available and highly available for applications and scale it horizontally while having that vertical expert domain specific value, proprietary value, meaning company's data, which is an intellectual property, then you can't win because there'll always be either gaps in the data or hallucinations, hallucinations from the lack of data, for instance. So is that going to be filled up by synthetic data, or will the cloud players change their egress fees? Because guess what? There's going to be a lot of data movements, Zeus.

Zeus Kerravala

>> They're not changing their egress fees, at least not anytime soon. One of the interesting aspects too, that was brought up in the keynote was all the adjacent markets that should get value creation because of AI ... and I've been looking at this for a while that when you look at the capital markets, NVIDIA has the obvious AI tailwind, but not too many other companies do. And they brought up the value of observability, the value of networking. And so when you take storage, observability, networking, put it all together, you need that to fuel AI. And I don't think investors have really given enough value to these tangential markets that will be used to fuel AI. It's really been an NVIDIA show, but it needs more than that to make AI go.

>> We'll see how that consortium goes. It's coming together to try to create a CUDA. What'd you think of Arvind's presentation? Because Arvind had talked about a lot, what struck me on that panel with Crawford was his comment about the pie being bigger, that there's a big pie for all of us to eat from. Okay, that is an important nuance because he's basically saying, "Hey, we'll open up IBM's distribution and our resources to partners because there's so much money to be made, so much service to have, that we're not afraid."

Zeus Kerravala

>> So that is the biggest change that I've seen out of IBM over the last few years is IBM is a company that always wanted the whole pie. And I think when you look at their acquisitions and the services business, to your point, you had to be the Blue partner, so you had to be committed to them. They're now willing to take a piece of the pie, because I think in the AI space, a piece of the pie is still bigger than the current pie you're playing in. And I think that's been a good change for this company. I think they've long-lived the walled garden mentality and that they seem to be moving away from that.

>> And if they can capture this slipstream and ride this wind, they can actually make up for so much lost ground on frankly, a cloud miss. SoftLayer, Bluemix, remember those days?

Zeus Kerravala

>> Yep.

>> That was IBM's old attempt at the cloud. Now they got a cloud group that's focused on hybrid. They get that with Red Hat. Red Hat comes into IBM, they got Ansible, now they got HashiCorp. They have a very formidable infrastructure play as a middleware DevOps layer. So DevSecOps is coming. Now you add generative AI into that. If they can get the data modeling right, to your point, they could be checkmate, game over.

Zeus Kerravala

>> And I do think the world is moving hybrid. I think I've talked to a lot of business leaders and CIOs about this where you look at all the macro issues with the war and even with the election creating a lot of uncertainty, businesses want to have more control over their data. And so I've seen a lot of organizations repatriate data back from public clouds to private clouds and want more of a hybrid model. Hence, that's where you see the value from HashiCorp and Red Hat and a lot of the IBM assets. It's that layer that's important more so than the underlying infrastructure.

>> Zeus, we've been in this business over 30 years. We've seen the movie many times. We are now at an inflection point. Like I said, you're starting to see the fog lift up and you can see the line forming that describes the old way and new way. What's your take on the market right now as you look at the landscape? What does IBM's key success factors have to be? What do they need to do better? And what is that line, how to determine old way versus new way?

Zeus Kerravala

>> I think there's a lot of fear still around AI. And I think for IBM, their ability to take its ecosystem, which includes all its partners, and be able to make AI consumable and simple. Crawford talked about the rise of as a service models, still growing strong. And I think this AI as a service is something that IBM could do, both from public cloud perspective but also from private and making that more of a hybrid AI model. And I think that's something they're ... I'm not going to say they're uniquely positioned to deliver, but the companies that can do that are pretty small. And for IBM, I think this just comes down to I want to see it in customer wins and they're knocking at the door right now.

>> Yeah. One of the things I like is that the software defined on the infrastructure side's got to be there. I think separating compute from storage is a great move in gen one. Now separate data from compute is another wave happening. And then I think just the infrastructure advances are going to come fast, then it's going to be a battle for the middleware. And the old school middleware, again, a generic concept, but what's going on in the middleware is control plane for data, the semantic layer, how data's going to move around, governed. If you look at watsonx.governance, from talking to folks around the ecosystem, it's the hottest product.

Zeus Kerravala

>> Yeah, John-

>> Because you got to get it right.

Zeus Kerravala

>> You could argue a lot of things you just brought up, software defined and infrastructure as a service things, largely were about cost savings, and they were, to me, really a solution looking for a problem other than cost savings. But now with AI, the problem has come to them. And so I think from that standpoint, that was almost a missing piece. I think technology, you see the hockey stick adoption when you have a perfect storm of things coming together. We had a lot of the components, we were missing that one piece. And that, to me, is AI.

>> You and I and Dave talk about baseball metaphors a lot. Hit the home run, early innings, it's a double header, it's not even spring training. Okay, gen AI, I'd say certainly in inning one or two, it's like it's happening, year two actually. But I think the way I would categorize these, and I'd love to get your reaction, is that the home run swings have to be on the consumer side. You got to hit the home run. If you're a Perplexity or an Anthropic, OpenAI, clearly those are home run plays. But in the enterprise right now, there's legit get on base small ball, hit a single, get some momentum. That's why I think the rise of RAG Retrieval Augmentation Generation and vector embeds in databases are hot because it's an easy thing to start doing. That's getting on base. So I think the enterprise is small ball right now, and I think it's major leagues. It's not minor leagues yet. So I think consumers clearly much different. It's a home run business. Not a lot of singles there. If you don't have the scale, you're out of business. What's your reaction to that?

Zeus Kerravala

>> I think the consumer space wins you the games with the home runs, let's you have some flash and sizzle, throw the bat. I do think on the enterprise space, that's where the scale comes from. And so if you want to win the season, that's where the small ball becomes more important and having more of a complete game. And so you're seeing a lot of wins right now in the consumer space, but ultimately it's been proven over time that the opportunity in enterprise and governments is so much bigger than it is in the consumer space. And so it's easy to get caught up in the consumer, I want to win that. But ultimately, if you're going to be a real business and scale, that's where enterprise becomes more important.

>> The reliability and being durable is a huge factor in the enterprise. Big deal.

Zeus Kerravala

>> Well, that's where the big contracts come from, right? You're not going to see any consumer company come out with multi-year, multi-billion dollar contracts, but you do see that in the enterprise side.

>> You can't have a ... "Oh, sorry, hallucination." Healthcare, you're dead. Stakes are high.

Zeus Kerravala

>> Well, that's where data becomes important because these are going to be built off curated private data sets. If you think of search, when search scaled, it's when LexisNexis and when Bloomberg and when Dow Jones and these companies built search into their products. They're not searching the internet, they're searching that private curated dataset. And we haven't seen that yet with gen AI. We don't see most enterprise apps with their own gen AI interfaces. There's a few of them. But that's where it's going to be built off that private curated dataset, and that'll remove a lot of the hallucinations.

>> Zeus, it's great to have you on theCUBE. Great analysis. We'll keep in touch. Enjoy the rest of the conference.

Zeus Kerravala

>> Always a pleasure, John. .

>> All right. We are theCUBE here, the pop-up CUBE here in Boston, Massachusetts for IBM Think. I'm John Furrier, your host with Zeus Kerravala here, bringing all the action. Keep it right there for more CUBE coverage from Boston.

Zeus Kerravala | ZK Research

search

>> Hello. Welcome to theCUBE's pop-up coverage. CUBE pop-up, that's what this is. We had a spot here, here at IBM Think. We are in the Omni Hotel right across the street from the Boston Convention Center. IBM think will be kicking off. It actually kicked off today. Now the sessions are here with the partners. Tomorrow, on Tuesday will be IBM Think's main program. I'm John Furrier with theCUBE. Your host Dave Vellante is out in Vegas for Dell Tech World. I'm here with Zeus Kerravala of ZK Research, CUBE friend, CUBE contributor. He's also got a very busy conference. This is conference season. You got here, you got Zoom, you got this.

Zeus Kerravala

>> Yeah, it is-

>> Good to see you.

Zeus Kerravala

>> Yeah, from here to the end of June, it's conference season.

>> Well thanks for coming on. I want to break down the keynote opening. Arvind was up on stage. You had a series of presentations, Dario Gill and then Rob Thomas, all the top executives. Of course they sprinkled in Partner of the Year. Everyone gets a trophy, you know how it is in enterprise tech. But the real story to me, and I just tweeted this out, is that IBM's back. The Once-Lost IBM now is back under one banner. Watson and Red Hat really together wedging in as the lead pillars. And everything wrapping around it, they got a huge portfolio, but AI is a gift. There's a platform shift happening and IBM is taking advantage of it. And we've been documenting this on theCUBE, you and I have been with Dave, and we've been researching and writing stories about. But AI is kind of a gift for IBM. They had Watson originally winning the Jeopardy thing and the NLP world. Now we're in a whole other ballgame with vectors and embeds and RAG and all kinds of foundation models. IBM's looking good, Zeus. What's your takeaway from the keynote?

Zeus Kerravala

>> Yeah. They were early out of the gates with AI, but then they kind of went to sleep and let everybody else take the mantle here for AI. And I think they are back, and it is a bit of a different IBM. When you think of IBM, I think a lot of people think of all the way back to the mainframes, the closed proprietary systems. And they worked well because they were walled gardens. And if you bought everything IBM, it would work well together. It was very clear in this keynote from Kate Woolley through Arvind, they are taking an open approach to AI. I think it was Dario Gill that said, "When you think of what open did for the internet, it's going to have the same impact for AI." And during Kate's presentation, she opened it up with the economic opportunity for AI is $4.4 trillion, of which they want the partners to share in that. And so she gave examples of some partners and some of the things they're doing. But ultimately if IBM is going to reestablish itself as the premier IT vendor, you could argue they've lost a little bit of that, it is going to be through the partnerships that they have. And they gave a lot of really good partner examples, and I think that having Partner Day lead off Think is a great way to do it.

>> And I think the open thing is a key message. The partner network could drive a lot of revenue, and that traditionally has not been an IBM strong point. I know they've had partners, rather than IBM Blue specifically ... When I say IBM Blue, I mean IBM only. Now what they have the open source mojo going on with Red Hat. They got HashiCorp on the other side. Like I said, you got it going on. Now, a couple nuanced points I want to get your reaction to because Dave Vellante brought this up in his breaking analysis. I also just posted it on the Twitter feed here that there's two things going on. One is IBM's portfolio is massive. So we look at the overall company performance from a deep tech perspective. Not every area is going to be impacted, but it is impacting a lot of the key areas. So I want to talk about that in terms of what the key pillars and the pistons of their business engine is. And the other thing that's going on with the foundation models and AI is that the gap between consumer AI and enterprise AI -

Zeus Kerravala

>> Huge....

>> is shrinking very fast. And if you look about the early days of the web, it was at least a couple year lag between consumer internet and then consumer internet in the enterprise, putting intranets aside. There's only a month lag now between consumer and internet. Now all the chips are being bought up by the big players. So ... Okay.

Zeus Kerravala

>> One of the differences though, John, and Rob brought this up in his presentation. The gap is huge because the data isn't there. And I think it was Rob Thomas in his presentation that said only ... Was it his? Or somebody said, only 1% of enterprise data is actually being used to fuel LLMs today. So that's where the big gap is. It's not the models per se, it's the data. And I think that's where companies struggle.

>> Well actually it's a very nuanced point. Let's unpack that because I'm actually saying the opposite. I'm saying that the technology gap between enterprise adoption is accelerating faster than the internet, meaning the language models don't have that end data for the enterprise for a good reason. As the power laws suggest, the long tail of data is actually an opportunity for enterprise. You're seeing that with the RAG, the Retrieval Augmentation Generation. So what's happening now is you're starting to see real production workloads that are very specific that aren't sexy projects. So the adoption on the enterprise is faster than the adoption on the internet is what I'm saying. So what that means is that that's what's happening and that's driving a lot of integration business. So if you're IBM and you have installed base, you have a built-in data business built in with the customer. So then the question is integration. So you're seeing InstructLabs, I am blown away by that, Zeus, because what Dario Gil was saying is what we've been saying on theCUBE. Models can interact with each other, you can train and you can impact a proprietary model on your data. So this means that models will integrate. That's going to be a huge win for IBM because they've been in the consulting business and integration business for years. So in comes Red Hat, in comes the generative AI hype. And guess what? IBM's not just a system integrator, they're a technology partner. So putting an ecosystem together, I just see this thing lining up beautifully.

Zeus Kerravala

>> Well, the services opportunity with AI is huge, because it is, I think easily the most complex system of technology that an enterprise has to put together. And every time we introduce new technology, we raise the complexity. So you think of cloud, hybrid cloud, edge, automation, all these things we've been through. Complexity is at a point where CIOs are telling me it's now at where it's untenable. And so you layer on AI on top of that and it makes the challenge more difficult. I think one of the big issues for businesses is how fast they can get their data into these systems. During this panel, Crawford Del Prete, the CRO of IDC said 75% of enterprise data loses its value in months and 50% in days. And so I think here's the challenge for businesses, as they generate so much data across so many different systems, can they find a way to aggregate it and make it useful from an AI perspective? And every CIO I've talked to about AI talks about data being the number one inhibitor. So if IBM and its services group and its massive partner organization can help companies get a handle on data, then yeah, they are back and I think they're back in a big way.

>> The data is the problem. And just the mind-blowing thought exercise is if you want to generate a new thing with generative AI, because it generates, it's not a pre-programmed thing, you got to have access to data. So if you're out on the edge on a device or a wearable or a machine, you need access to every single piece of data in very low latency. To make that happen, Zeus, is going to require a completely different rethinking of the data modeling, the availability of data, moving data around-

Zeus Kerravala

>> And your infrastructure....

>> and how you're going to store that. So I think what's happening now, and it's clear to us as we start to see that the old way, new way form, it's very obvious. If you can't make data highly available and highly available for applications and scale it horizontally while having that vertical expert domain specific value, proprietary value, meaning company's data, which is an intellectual property, then you can't win because there'll always be either gaps in the data or hallucinations, hallucinations from the lack of data, for instance. So is that going to be filled up by synthetic data, or will the cloud players change their egress fees? Because guess what? There's going to be a lot of data movements, Zeus.

Zeus Kerravala

>> They're not changing their egress fees, at least not anytime soon. One of the interesting aspects too, that was brought up in the keynote was all the adjacent markets that should get value creation because of AI ... and I've been looking at this for a while that when you look at the capital markets, NVIDIA has the obvious AI tailwind, but not too many other companies do. And they brought up the value of observability, the value of networking. And so when you take storage, observability, networking, put it all together, you need that to fuel AI. And I don't think investors have really given enough value to these tangential markets that will be used to fuel AI. It's really been an NVIDIA show, but it needs more than that to make AI go.

>> We'll see how that consortium goes. It's coming together to try to create a CUDA. What'd you think of Arvind's presentation? Because Arvind had talked about a lot, what struck me on that panel with Crawford was his comment about the pie being bigger, that there's a big pie for all of us to eat from. Okay, that is an important nuance because he's basically saying, "Hey, we'll open up IBM's distribution and our resources to partners because there's so much money to be made, so much service to have, that we're not afraid."

Zeus Kerravala

>> So that is the biggest change that I've seen out of IBM over the last few years is IBM is a company that always wanted the whole pie. And I think when you look at their acquisitions and the services business, to your point, you had to be the Blue partner, so you had to be committed to them. They're now willing to take a piece of the pie, because I think in the AI space, a piece of the pie is still bigger than the current pie you're playing in. And I think that's been a good change for this company. I think they've long-lived the walled garden mentality and that they seem to be moving away from that.

>> And if they can capture this slipstream and ride this wind, they can actually make up for so much lost ground on frankly, a cloud miss. SoftLayer, Bluemix, remember those days?

Zeus Kerravala

>> Yep.

>> That was IBM's old attempt at the cloud. Now they got a cloud group that's focused on hybrid. They get that with Red Hat. Red Hat comes into IBM, they got Ansible, now they got HashiCorp. They have a very formidable infrastructure play as a middleware DevOps layer. So DevSecOps is coming. Now you add generative AI into that. If they can get the data modeling right, to your point, they could be checkmate, game over.

Zeus Kerravala

>> And I do think the world is moving hybrid. I think I've talked to a lot of business leaders and CIOs about this where you look at all the macro issues with the war and even with the election creating a lot of uncertainty, businesses want to have more control over their data. And so I've seen a lot of organizations repatriate data back from public clouds to private clouds and want more of a hybrid model. Hence, that's where you see the value from HashiCorp and Red Hat and a lot of the IBM assets. It's that layer that's important more so than the underlying infrastructure.

>> Zeus, we've been in this business over 30 years. We've seen the movie many times. We are now at an inflection point. Like I said, you're starting to see the fog lift up and you can see the line forming that describes the old way and new way. What's your take on the market right now as you look at the landscape? What does IBM's key success factors have to be? What do they need to do better? And what is that line, how to determine old way versus new way?

Zeus Kerravala

>> I think there's a lot of fear still around AI. And I think for IBM, their ability to take its ecosystem, which includes all its partners, and be able to make AI consumable and simple. Crawford talked about the rise of as a service models, still growing strong. And I think this AI as a service is something that IBM could do, both from public cloud perspective but also from private and making that more of a hybrid AI model. And I think that's something they're ... I'm not going to say they're uniquely positioned to deliver, but the companies that can do that are pretty small. And for IBM, I think this just comes down to I want to see it in customer wins and they're knocking at the door right now.

>> Yeah. One of the things I like is that the software defined on the infrastructure side's got to be there. I think separating compute from storage is a great move in gen one. Now separate data from compute is another wave happening. And then I think just the infrastructure advances are going to come fast, then it's going to be a battle for the middleware. And the old school middleware, again, a generic concept, but what's going on in the middleware is control plane for data, the semantic layer, how data's going to move around, governed. If you look at watsonx.governance, from talking to folks around the ecosystem, it's the hottest product.

Zeus Kerravala

>> Yeah, John-

>> Because you got to get it right.

Zeus Kerravala

>> You could argue a lot of things you just brought up, software defined and infrastructure as a service things, largely were about cost savings, and they were, to me, really a solution looking for a problem other than cost savings. But now with AI, the problem has come to them. And so I think from that standpoint, that was almost a missing piece. I think technology, you see the hockey stick adoption when you have a perfect storm of things coming together. We had a lot of the components, we were missing that one piece. And that, to me, is AI.

>> You and I and Dave talk about baseball metaphors a lot. Hit the home run, early innings, it's a double header, it's not even spring training. Okay, gen AI, I'd say certainly in inning one or two, it's like it's happening, year two actually. But I think the way I would categorize these, and I'd love to get your reaction, is that the home run swings have to be on the consumer side. You got to hit the home run. If you're a Perplexity or an Anthropic, OpenAI, clearly those are home run plays. But in the enterprise right now, there's legit get on base small ball, hit a single, get some momentum. That's why I think the rise of RAG Retrieval Augmentation Generation and vector embeds in databases are hot because it's an easy thing to start doing. That's getting on base. So I think the enterprise is small ball right now, and I think it's major leagues. It's not minor leagues yet. So I think consumers clearly much different. It's a home run business. Not a lot of singles there. If you don't have the scale, you're out of business. What's your reaction to that?

Zeus Kerravala

>> I think the consumer space wins you the games with the home runs, let's you have some flash and sizzle, throw the bat. I do think on the enterprise space, that's where the scale comes from. And so if you want to win the season, that's where the small ball becomes more important and having more of a complete game. And so you're seeing a lot of wins right now in the consumer space, but ultimately it's been proven over time that the opportunity in enterprise and governments is so much bigger than it is in the consumer space. And so it's easy to get caught up in the consumer, I want to win that. But ultimately, if you're going to be a real business and scale, that's where enterprise becomes more important.

>> The reliability and being durable is a huge factor in the enterprise. Big deal.

Zeus Kerravala

>> Well, that's where the big contracts come from, right? You're not going to see any consumer company come out with multi-year, multi-billion dollar contracts, but you do see that in the enterprise side.

>> You can't have a ... "Oh, sorry, hallucination." Healthcare, you're dead. Stakes are high.

Zeus Kerravala

>> Well, that's where data becomes important because these are going to be built off curated private data sets. If you think of search, when search scaled, it's when LexisNexis and when Bloomberg and when Dow Jones and these companies built search into their products. They're not searching the internet, they're searching that private curated dataset. And we haven't seen that yet with gen AI. We don't see most enterprise apps with their own gen AI interfaces. There's a few of them. But that's where it's going to be built off that private curated dataset, and that'll remove a lot of the hallucinations.

>> Zeus, it's great to have you on theCUBE. Great analysis. We'll keep in touch. Enjoy the rest of the conference.

Zeus Kerravala

>> Always a pleasure, John. .

>> All right. We are theCUBE here, the pop-up CUBE here in Boston, Massachusetts for IBM Think. I'm John Furrier, your host with Zeus Kerravala here, bringing all the action. Keep it right there for more CUBE coverage from Boston.