theCUBE + NYSE Wired: AI & Retail Trailblazers | Alex Bouzari, DataDirect Networks

Clips
More from theCUBE + NYSE Wired: AI & Retail Trailblazers

Alex Bouzari

CEO

DDN

play_circle_outline Blackstone to Invest $300 Million in DDN for AI Acceleration and Cloud-Enabled Data Intelligence

play_circle_outline Unlocking Success: AI's Role in Expanding Customer Base and Enhancing Product Capabilities in the Digital Era

play_circle_outline Accelerating AI Success: Collaboration between NVIDIA and DDN for Data Processing and Insight Delivery

play_circle_outline The Impact of Companies Staying Private: Shifting Landscape in the Business World

play_circle_outline Maximizing Growth Through AI Transformation: A Deep Dive into Company Profitability and Capital Infusion Strategies

Info
Transcript

Alex Bouzari, DataDirect Networks

Alex Bouzari

CEO DDN

Alex Bouzari, CEO of DataDirect Networks, received a $300 million investment from Blackstone to expand DDN's presence in the AI industry. Bouzari highlighted the importance of data intelligence for AI success and the need for a flexible data platform to support AI applications. DDN aims to broaden its customer base across industries like life sciences, manufacturing, energy, autonomous driving, and financial services by partnering with Blackstone. The collaboration with NVIDIA focuses on hardware, software acceleration, and data storage for AI applications li... Read more

explore Keep Exploring

What were the reasons behind DDN's decision to consider partnering with Blackstone for investment in AI? add

What are the two main goals for expanding the capabilities of the product portfolio within the enterprise sector, and how is Blackstone helping achieve these goals? add

What are the key components needed for AI to add value to organizations? add

What are the advantages and disadvantages of being a public company in terms of making strategic moves that tie into DDN as opposed to staying private, especially in light of the changing landscape of organizations staying private and the company's own profitability? add

What was the main purpose for the company's recent decision to take on outside capital? add

bolt Powered by CUBE AI

Alex Bouzari, DataDirect Networks

search

Dave Vellante

>> Harvey, welcome back to the NYSE Wired and the Cubes Media Week coverage. We're here covering NRF, but we have a special guest right now, Alex Bouzari is the CEO and founder, co-founder of DataDirect Networks, DDN. Alex, great to see you. We just saw you at Super Compute. Wonderful to have you in studio here.

Alex Bouzari

>> Thank you so much, David. It is great to see you, and again, congratulations on the expansion of what you guys are doing. You're doing amazing things.

Dave Vellante

>> Thank you very-

Alex Bouzari

>> Just keep expanding.

Dave Vellante

>> Thank you very much, and we're looking forward to having you down here at the New York Stock Exchange next time you're in town. But big news, we saw a hit just after Super Compute Blackstone to make a $300 million investment in AI data company. They said a company called DDN. Of course we know who DDN is. It's interesting, you don't need the money, but-

Alex Bouzari

>> No, we don't.

Dave Vellante

>> Why did you take the money? Tell us about the logic behind that.

Alex Bouzari

>> Sure. It's simple. You've seen the announcements, you've seen everything that's going on in AI. AI is broadening, it's accelerating, it's really permeating the enterprise now. Lots of people have been coming, knocking on our doors, asking us to become part of DDN, to invest in the DDN to share in our growth. When Blackstone came to us, we took notice because Blackstone has a history of really helping pivot and accelerate industry transformations. They had already been invested into AI, they had done a very large investment into CoreWeave. They came to us and they said, "Hey guys, we think that in order for AI to accelerate, two things are needed, cloud enablement, So that's what we're doing on the data center side, but also data intelligence. Clearly, in order for these AI models to accelerate, to grow and enable enterprises to be successful, we need data. We need intelligence from data, and we think you DDN are the right ones. We can amplify your reach and we can accelerate your reach." It wasn't at all about the money. Actually, we spent quite some time discussing what is the minimum amount of money that would invest in order for us to engage. It was a reverse negotiation. They said, "We've never seen that before. You guys are doing it in reverse."

Dave Vellante

>> That's amazing.

Alex Bouzari

>> It's really that it's accelerating things for us.

Dave Vellante

>> I love it. Well, congratulations and great headlines and great news. You sort of laid out why Blackstone, of course, they've made some other data center investments as well. You know David Floyer, our analyst emeritus, he and I just did a piece last week. He was sort of sizing the data center market, the All-in. It's several hundred billion when you look at it, but it's growing at an amazing rate. It's around 15% CAGR over the next 10 years. We're talking about a market that historically has grown low single digits. All of a sudden it's been supercharged because of AI, and every part of the stack is changing. I wonder if you could talk about, they said in the journal article anyway, you're going to use this to expand your customer base. How are you thinking about expanding your customer base? What does that actually mean?

Alex Bouzari

>> Sure. It's really two things. The first one is expanding the capabilities of our product portfolio to make them more enabling for the enterprise. Clearly all of these expansions into data centers, whether it is the physical build out of data centers or the transformation of existing data centers into being AI enabled or the cloud, all of this has to tie into providing real value to enterprises. For us, the go-to-market part of it is reaching out across industries that are being enabled by AI, meaning life sciences, manufacturing, the energy sector, autonomous driving, financial services, and being able to access the C-suite within these organizations. So Blackstone, well, given their size and their reach and their footprint globally, is helping us achieve that. It's really that, it's accelerating our growth. Because we have the technology that is highly enabling to the enterprise in AI. We are accelerating the training of the models. We are providing better, faster insight. So it's reaching out and getting our voice out there, amplifying our voice, that is Blackstone is helping us with.

Dave Vellante

>> Really, I don't even... This is not a pivot for DDN, this is a line extension. Your history in the high performance computing market, those markets in AI have come together. So as you say, it's really about the go to market, the routes to market. You don't typically sell to the C-suite, you typically sell to scientists, and so that is new. Are there things, however, in the product that you feel like you have to do to get into the enterprise? Because much of the stuff in enterprise is in cloud, you mentioned before, the cloud affinity. I wonder if you could discuss that a little bit?

Alex Bouzari

>> Sure. From a product standpoint, fundamentally, what do you need in order for AI to be successful? I think Jensen from NVIDIA says it very, very well, you need an acceleration and an amplification of the models. Well, that has a cost associated with it, so you have to lower the cost. I think we have that by providing far higher levels of efficiency. That's just table stakes. We've been doing this for two decades. We've been doing it for 70 of the 100 largest, fastest supercomputers in the world, so the pivot into AI, it's a natural. Now, what do you need for the enterprise? I think you need two additional things. One, you need enterprise enablement. Enterprises do not want to disrupt their existing infrastructures, so it has to plug and play into their existing infrastructures with no pain, with no effort. We're developing layers that make that easier. The second thing is there's been a trend of enterprises to move into cloud. We seeing that with AI and with Agentic AI, the ability to bring in data that is proprietary to a company and train the models to get better insight. What will have to happen for AI infrastructures, you will need a hybrid model, where you need cloud enablement and you need on-premise data center enablement. So we're developing additional capabilities in our product so that we can deliver the AI value, the AI acceleration, both on-premise and in the cloud, in multi-cloud. That's what we believe will be required for enterprises to be really successful and really benefit from the AI revolution that is taking place right now.

Dave Vellante

>> Thank you. That's super helpful. I want to dig into that a little bit, because as I mentioned, we just sized , it was astounding when we dug into it. The other finding that we came up with, Alex, is if you go back to 2022, 2023 even, the AI, what Jensen calls the accelerated computing, we call it extreme parallel computing, EPC, same thing, but that was single digits in terms of the market share within the data center. Just this past year and into '25, we're talking about 25%. So from single digits to 25% almost overnight. That is going to dominate, it's going to basically become 90% of the market by the end of the decade or early part of the decade. It's the shift from general purpose computing to accelerated computing. It's on and it's off. Now, we didn't dig in to each of the individual layers of the stack, but you must be seeing this in your business, it's got to be a huge tailwind, the storage layer is obviously a key part of that because of data. I wonder if you could talk about the momentum that you're seeing specifically around the shift toward accelerated computing.

Alex Bouzari

>> Absolutely. As you pointed out, Dave, it's all about the data. In order for AI to add value to organizations, you really need two things. You need to ingest the data, and the data is text, it's images, it's video, it's audio. All of this needs to come in and it needs to come in from outside the organization as well as inside the organization. Now, the additional layer, which is really fueling this growth, and that's one of the reasons why the CAGR is so high in data center growth, is you need to add synthetic data to it. That's the agentic side of it. That's the Omniverse, that's digital twins. Let's say I'm a pharmaceutical company. I'm doing a new drug development. Instead of trying to branch out and try 10 or 20 or 30 different avenues, which will be very, very costly and very time-consuming, I'm going to spin up an Omniverse digital twin where I'm going to try all these scenarios and then see which ones have the highest promise of success and that's what I will focus on. That is creating huge amounts of data that needs to be processed at very, very high speed. That's where DDN really comes in. We help with that transition, the Omniverse, which Jensen spoke of at CES, it is the existing data. It is synthetic data. It is merging all of this together in order to deliver insight that will make companies and organizations more successful. Data is front and center to all of this.

Dave Vellante

>> It's interesting, you talk about Omniverse, there's NIMS, there's NeMo, there's COSMOS, all this software stack that Jensen is building. I remember I was talking to Dave Itzicheria, as recently as this past summer. We were kind of riffing on NVIDIA and its momentum. The conversation at the time was a comparison back to the.com. Is NVIDIA Cisco or is it Google? The kind of conclusion that we came up with was it's both the Intel and the Microsoft, it's got the hardware and the software. All those software components that you just talked about, it really is an amazing portfolio that Jensen is putting together. I think the point being, now you're talking about a whole range of applications, which is very difficult to forecast. Who knows if our forecasts are right? If anything, they're probably conservative because now we're talking about digital twins, digital representations of businesses in real time. What role does DDN play in that vision?

Alex Bouzari

>> Again, you're right on. In order for organizations to really benefit, and that's why NVIDIA is in such an amazing position, because they're solving two problems. One, they're significantly accelerating lowering the cost and lowering the power of what needs to be deployed in the data center. That's the GPUs, that's the accelerated GPUs, that's the ability to rev the silicon and the underlying hardware. That is drastically lowering the cost and increasing the speed. The other part of it is making the models run much faster and providing much better insight. NVIDIA is solving the problem from a processing and GPU standpoint with all of the software around it. We're solving the problem from a data standpoint. You need both. You need the NVIDIA part coupled with the DDN part. They're doing the processing and the insight, large language models, training, digital twin inference. We are doing it from the data side. Because all of this acceleration can only bring value to enterprises if the data is being analyzed and results in insight. We saw the exact same thing in high-performance computing back in the day. Organizations came to and said, "DDN, we have these super computers. They're running super-fast. We have the models, we have the simulations, but the data isn't keeping up with it." Well, we have developed technology that makes the data run as fast as anything NVIDIA is throwing at us. Whether it's the Cosmos infrastructure, which we're will foster an explosion in growth, or the digital twins spinning up all these alternatives to really analyze what is going on. All of this requires an acceleration of data, and that's what we're doing. We're accelerating data by orders of magnitude in order to deliver true value to enterprises alongside NVIDIA.

Dave Vellante

>> The parallels between what happened in HPC and what's happening now in AI are interesting to note. I want to ask you about something you talked about earlier, the hybrid. Of course, a lot of the experimentation and early work in AI has occurred in the cloud, but as we know, much of the data, if not most of the data lives on-prem. Jamie Dimon doesn't necessarily want to put everything into the cloud, he'd like to bring things, he'd like to bring the AI to the data. At the same time, the cloud has the tooling, the software ecosystems in a very robust portfolio of capabilities. I'm interested in your thinking. Now you're participating in Colossus, so a lot of the infrastructure players, they're selling either to Neoclouds, or if you're lucky enough to have earned the business in the Memphis Colossus Center, that's a nice tailwind to your business. But there's a big opportunity in the enterprise as we were talking about earlier. But you've got to have... I'm asking the question around, what other tooling do you need to replicate what's happening in the cloud, which is again, a lot of experimentation to bring that on-prem missing, and how are you filling those gaps?

Alex Bouzari

>> What you really need, well, there was a big shift, there were organizations which said, "Well, the old way of doing things on-prem doesn't really work, you need the flexibility and the versatility of the cloud." Then enterprises started shifting into the cloud. With Ai, what we're seeing is, you really need a combination. You cannot just say, "I will do everything in the cloud," or, "I will do everything on-prem." You need to bring the two together. You need a data platform, you need an AI data platform that can handle both, and that can handle both with great levels of flexibility. You need to be able to have your data in your data center, the proprietary data, which you don't want anybody else to have access to. You analyze, you process, you spin up some digital twins, and then you tie that into the other data, which is out there, which you will put into the cloud, and you combine the two together. Just as in order for AI to be successful, you need the models and the training of the models, and then you need to tie into inference and insight. Today you might be doing 90% of the cycles that you're utilizing into training. Tomorrow, you will flip it and 90% will be in inference. The flexibility of the data model is extremely important, and that's what we've really spent a decade developing. A very, very flexible, very powerful, very scalable and highly efficient model that can operate in the cloud, on-prem, for training and models as well as inference. You have to bring it all together. That's what is required. That's why Jensen is doing all the things that he's doing. That's why he's casting such a wide net. Because for AI to deliver true value for the ROI to pencil out for enterprises, you need all of it. We're doing the same thing on the data side. We're doing all of it.

Dave Vellante

>> I'm interested in maybe double clicking on that. Thinking about the entire stack, you've got clustered systems in compute, you've got ultra low latency networks, you've got OSs that are AI optimized, you've got the data stack that's transforming. People in storage talk about globally distributed file systems, you just talked about the need to have flexibility. Are there other salient attributes of, let's call it accelerated storage, that people should be aware of and or specifically where DDN has an advantage?

Alex Bouzari

>> Sure. In order for AI to work, you really need to tie into and boost the AI frameworks and the accelerated AI. The PyTorch, the tensor flows and tensor stores. You have to be able to tie into the NVIDIA ecosystem. NVIDIA has spent more than a decade developing the CUDA ecosystem with millions of developers. Well, they released CUDA OBJ, CUDA object, which gives you the ability to tap into it and all of a sudden deliver value across industries and across use cases. What we've developed is really this flexible architecture, which by the way, we started working on it in earnest in 2017 when NVIDIA came to us and said, "Hey, let's stand up," what back then was called the super pod, "Let's stand up a system which demonstrates the capabilities of our GPUs at massive scale, but we need the data storage." That's when we said, "Oh, in order for this to really play out, you have to tie into the AI frameworks, you have to tie into the CUDA. You have to leverage what these millions of developers are doing out there." So we developed this very, very complex framework, millions of lines of code just to do that. That I believe, is providing significant value and will accelerate the pivot of the global economy into AI enablement and for the ROI to pencil out. At the end of the day, the ROI has to pencil out for organizations, otherwise it's very slow.

Dave Vellante

>> Well, I want to go back to 2017. Was that a no-brainer for you and Paul and the team?

Alex Bouzari

>> Not at all. It wasn't obvious at all. Because we walked away from that first interaction and we said, "Well, wait a minute. The traditional way," which you were just describing, Dave, the parallel file system, all that, "Is good. But eventually, if AI really becomes massive and broadens across all industries, you need a different approach." But that's when we started developing our new technology called Infinia, and we said we have to tie in straight into the upper layers, because the work is happening in the upper layers and the real value of AI, the acceleration that is required is orders of magnitude, and you cannot do that just with a parallel file system. I think for the next few years it's still okay, but this Infinia technology that we've developed is taking it to another level the same way Jensen, Cosmos and all of the infrastructure around it are taking things to another level.

Dave Vellante

>> You had to make a bet though, that wasn't an obvious -

Alex Bouzari

>> Yeah, wasn't obvious. It was not obvious.

Dave Vellante

>> Then ChatGPT happens and the whole world goes AI crazy, so that's been awesome. I want to ask you-

Alex Bouzari

>> Jensen has a history of seeing things years and years ahead. So you go, "Okay, if he's seeing it like this, that's probably not really a bet, it's an educated guess."

Dave Vellante

>> Yeah, the guy's got street cred. I want to ask you, the articles that came out on the news with Blackstone and the investment, it speculated about IPO, I want to ask you about that. But specifically I want to ask you about two companies, Anthropic and Databricks are now with 60 billion valuations. A lot of people saying, "Why go public?" You've got cash, you've got a strong balance sheet. Same question to you. Is it something that you aspire to, to IPO?

Alex Bouzari

>> Not really though. Look, for us, it's all about customer delight. How do we delight our customers? Delighting our customers means delivering the highest value in their AI transformation. Being a public company, being a private company, I'm not really sure that has an impact on that. The reason we look at that is, is being a public company easier in terms of getting large organizations to make strategic moves that tie into DDN. But with what's going on over the last several years, as you pointed out, more and more organizations at very large scale, billions of dollars of revenue are staying private, so that need has gone away. I think the landscape is shifting. The need to go public in order to have better visibility is no longer there. In our case, we don't need the money. The company has been profitable. We've had double-digit to the bottom line for more than a decade. We're looking at all this, how do we best serve our customers and how do we accelerate our growth and the value we deliver to enterprises who are going through their AI transformation?

Dave Vellante

>> Well, it's a remarkable story that you guys have built. I think I'm correct, you guys started in the middle of the dot-com boom where you very easily could have raised some money with just having a website, but you chose not to. I think this is the first outside capital you've taken. Is that correct?

Alex Bouzari

>> We did a very small round at the very beginning, this is the first meaningful capital infusion. But again, the main purpose for it was really to accelerate our growth. It's not about the money. The money is good because our engineering team is like, "Oh, great, so now we can pull in all these feature developments, we can hire more engineers. It's great, it's great, it's great." But nonetheless, it was really about, how do we deliver the DDN value to enterprises with a louder voice? And that's what Blackstone is helping us do.

Dave Vellante

>> Amazing story. Alex Bouzari, thanks so much for your time. It's always a pleasure seeing you and having you on theCUBE.

Alex Bouzari

>> Thank you so much, Dave. Really appreciate it.

Dave Vellante

>> All right, we'll talk soon. All right, keep it right there. More action. Dave Vellante and John Furrier from NYSE Wired in theCUBE community. We'll be right back on Media Week.