 Hello, welcome back to SuperCloud 3. This is a keynote conversation here back in Palo Alto, California, I'm John Furrier, host of theCUBE with David Flynn, CEO of Hammerspace, CUBE alumni, entrepreneur. Some big news launching here at SuperCloud. David, great to see you on theCUBE. Thanks for coming on. Good to see you too, thanks. So before we get started, I'll talk about SuperCloud 3, the dynamics of security, data, AI, all things going on at SuperCloud, powering the next generation apps, this next cycle's coming. You guys got some big news with Hammerspace, as the CEO. That's right. CUBE here, exclusive news. Exclusive news, just announced a major funding round, 56 million in funding from an A-list set of investors. So a tough market right now for funding. What was some of the things that we're talked about? What was the validation? What was the key factor in closing this round? Yeah, the key here has to do with proof of the product market fit through the customer base. So the likes of Blue Origin, Jeff Bezos own company, who's going to know Cloud better than that, that Blue Origin is using Hammerspace as their data plane for Hybrid Cloud. That's a pretty big endorsement. So it's really just seeing the momentum and especially how it's impacting the next data cycle with AI, ML, advanced analytics. You guys done a great job with SuperCloud before our last event and been a big part of the community as this has been growing. This been a big focus on things below and below, below SuperCloud and above SuperCloud. So super computing, seeing a lot of advances at the physical layer, GPUs, and then above SuperCloud layer, you start to see super apps emerge. So you get the super stack, I call it. And it's all revolving around data. So whether you're talking about security, data matches or whatever, the role of data is key. And that has become the hottest topic because AI is data and all the value you're seeing with AI is from the data revolution that's going on. Again, another inflection point on data. You call it the next data cycle. I like this term, explain what you mean by this next data cycle. The next data cycle is really defined in how things have radically shifted. We went from a world with data warehouses, maybe cloud hosted data warehouses, highly structured data, a lot of painstaking maintenance of it, to a world now where unstructured data rules, the massive amounts of data that you have. And that's because with AI and ML, you don't need it to be pre-structured for you. So a lot more value can be extracted from much larger quantities of data. We view this as just a new data cycle. It's the next data cycle. You know, David, we've known each other now for over a decade. You've been a great successful entrepreneur, multiple success and exits. But you've been in the storage data game for a long time. In the data for a long time. So you've seen many cycles. Absolutely. This cycle, why is it different? Because, again, storage, you store data, it goes from point A, it gets stored at point B. Now you have data in flight. Okay, you're seeing cloud, you're seeing AI. Where are we now? Tell us how you see this, because this is unique time in history. That's right. We've finally reached the tipping point where we have to move from a world where data is something that is stored, copied and merged to a world where data simply exists and is orchestrated transparently. So moving from store copy merge to orchestrated is a radical change. And what it allows you to do is to have data be present everywhere that you need it with high performance, local access. So you can feed it into AI training and inference. But have it at any site that you need it at. Servers, storage and networking never go away. They just abstract it away as more functionality complexly comes in. As storage goes into super mode, super cloud mode here. What has to change? You mentioned orchestration, what's different and what has to be different going forward. The first thing is that data is a platform thing. And yet data has been a mirage that is rendered by the infrastructure. And so infrastructure storage systems have been the captors of data. So you have an inversion between platform and infrastructure. So with data orchestration, you finally have data that can transcend the infrastructure is not a captive to the infrastructure. What are some of the technologies involved in this? Because you're seeing everyone's getting behind this next wave. I think everyone now sees AI as kind of like it is the future path, they see AI. They can connect the dots. They can see the bridge that they have to cross. Absolutely. What are those technologies? Well, obviously GPUs, other specialized processing for being able to work with those large amounts of data. The unsung hero behind it though is the data platform and the ability to orchestrate the data to where you have the GPU clusters, whether that's rented in the cloud or on-prem. So I think some of the key things are the GPUs, AI and ML, the models and such that we work with for training and the data orchestration piece for making the data. Explain what is the Hammerspace data orchestration product solution technology? What does it mean? What is it? Fundamentally, yeah. And define it and then talk about the benefits. So fundamentally it's two things. It's a file system that has finished the progression that file systems have been on since the beginning. File systems used to be part of the OS and then they were put on the network as an appliance, network appliance and then they were made to scale out. Now what we're talking about is a file system that can span everything. It can sit on top of any form of storage, third-party storage, vendor neutral and can span across whole data centers. So that first thing, some talk about that as a global namespace, right? But it's a high performance and it's a new class of file system. It's a true parallel file system that gives you the kind of scale. So that's the number one piece. And the second part is that that then allows you to take the data movement where you place and move data can now be done behind the file system to where it can now be fully automated and is non-disruptive to the use. So data orchestration is what happens when you take data management and put it behind the file system instead of out front. Give me an example of this in motion. I mean, I shouldn't say that in motion, because data is in motion. Give me an example of this in- In practice. In practice, a use case, give an example. In practice, there's two key motivators. Number one is to have a workforce that can be anywhere in the world and be able to interact with the data. And number two is to have the compute be anywhere in the world. Be able to take data that's collected at one site and be able to do the ML training at another site to have researchers at other locations be able to work with it. And each of them be able to have local high performance access to that data from different computing centers. You know, I love the AI world and the data analytics, how it's in a collision course that has collided. It happened. It was almost the chat GPT woke everyone up saying, wow, that's magic. But it really kind of educated the masses, which is great. What is the AI impact here? Because when you have data that's fundamental for AI, you have to have it available, highly available, and highly available, have high availability and make it highly available, two different things. So you have the training and the inference of the two kind of nov people are talking about that are costly right now. Well, and the key to that is that the old mantra of move the compute to your data, you know, because of data gravity and because data, a platform layer thing is a prisoner to storage, an infrastructure layer thing, that layering inversion, because of that, the mantra was move the compute to the data. But that simply won't work anymore when you need tens of thousands of GPUs. When you want to do things in mass, you have to be able, I mean data after all is the thing that's digital. Why would we be saying moving the compute to the data? The compute is what you have to have racked and stacked and powered and cooled on massive scale. It's the thing that's tied to the physical world. The data's digital. So we have to be able to orchestrate the data so that it can be, at one point in time, local to this location and at other points in time, local to another. Okay, that goes against some of the things that I've been hearing over the years, like move the compute to the data. You mentioned that. Yeah, yeah. It's expensive to move data. What's the cost factor here? Is that a bottleneck? Is that a feature or a bug? I mean, this is a- So the key to that and the magic part in data orchestration is a push-pull model to where data can be pushed through policy and can be pulled on demand. And key to both of those is that it's done granularly and non-disruptively. So you see a file system, you access files. It's that simple. But through policy, you can have the data pushed so that it's already local. You don't have to wait for it. But because it's behind the file system, you can also have pull, something goes to access it, it goes and grabs it. So you get the best of both worlds and it boils down to the, by the way, it's the same design methodology that's used in Kubernetes. And that's to do automated orchestration of very fine-grain, lightweight, encapsulated units. I was going to bring up Kubernetes and it's funny that you brought that up because Kubernetes solves that orchestration problem. They're orchestrating compute. Exactly. In the cloud native. And that opened up a unification model for the cloud native surge microservices. And essentially now that's rendering itself and shifting left, developers are in charge. So the parallels between, you know, orchestrating containerized microservices and orchestrating data are very powerful. One is you take the smallest, most granular unit, the microservice or the individual file. And the second is you have the lightest weight encapsulation, a container, right? And it's orchestrated independently. And the same thing is true here. The ability to move data at the file granularity without disrupting the access. So very similar principles, whether you're orchestrating compute or orchestrating data. I want to get you to react. I love this conversation because it's right in line with what we've been reporting and thinking. The rise of super cloud is the new IT is what I'm calling it. Because if the super cloud is IT, the developers are ultimately in charge. Because the Kubernetes has unleashed the productivity of developers. Yeah, it caused some new concerns, software supply chains being addressed. Okay. Now you got data going down that same path. I've been saying data developer is the super cloud, the new IT. So you could say so, because what it's really letting you do is to work at the platform layer and no longer have to care about the infrastructure layer, right? And that's how it empowers those folks is they can self-service with orchestrated, containerized microservices, with orchestrated unstructured data at scale. Now you don't have to think about all of the logistics of the infrastructure. I was saying to Dave Vellante, we were riffing on this topic on our podcast every Friday by the way, check it out, is IT guys are going to be in the guardrail business. You hear that term a lot, set up the guardrails. In AI you hear that, you know, set up the guardrails. Set up the playpen, make a playpen. Get everyone set up, bounce around the guardrails, don't go off the track. Mainly because of one fear, security fears also, AI and the economy. That's just kind of how things are going. This guardrails is an operational impact. And so I can see data orchestration be part of super cloud. How does it connect to super cloud in your opinion as you go down this data orchestration, next data cycle emerging, how does that intersect with super cloud? Well, the way I think of super cloud really is it's a software defined cloud, a cloud of clouds. You know, a cloud that's, I would say, as it was always intended to be, the ability to run anything anywhere without having to give a hoot about the infrastructure concerns. But you can't do that when data is subordinated to the storage infrastructure. So that data has to be something whose existence is now transcendent of the infrastructure. And that is what orchestration does. It allows data to exist independent of any storage and outlive any storage because now it's, your data is perpetually in motion. Simply through objective, you can swipe a mouse and the data is present wherever you need it. You can commission new infrastructure. You can decommission old infrastructure, move data to it. And all of that is done behind the scenes, behind the file system. It's like a new abstraction layer. And I like this idea of smart data management or intelligent data management. That's my word, you kind of saying that. But that's what orchestration is. It's really bringing this ability to scale horizontally, but yet give the vertical domain AI the right data at the right time in the apps. Exactly, exactly. But I don't think people understand how radical a departure it is from what we know in the past. Give an example. Data management has always been storage, store, copy and merge, right? In one form or another way that's making backups or restoration of backups, whether it's replicating between sites, whatever. It's always been a function of store it, copy it and merge it. And if you ask people if there's another way, there is no other way. It doesn't seem like there's another way other than that. But in an orchestrated environment, now the data moves behind the data presentation layer so you don't even see it when it moves. It allows data to be transcended of it and have a continuous accessibility even while the data is pushed and pulled across different infrastructure. That's a great point. I want to ask you one more thing to add to that because I think that's a good description. You mentioned platforms earlier. When you're in an orchestration model, you got to have an enabling benefit versus say a siloed benefit, which is an operational construct. Okay, a siloed database. Yeah, yeah. Store stuff, call the guy, galley, get some data, merge it, whatever. When you talk about a platform, that's an enablement. You got to have an enablement, enabling benefit. Well, I'd say one of the most profound things that this enables is it enables a new form of collaboration between companies whose products are data. More and more data is the product, right? And it too has a supply chain and that typically has been, again, store, copy, and merge. But in a world where you have data orchestration, you can literally share the same global file system. And vendor A and their contract vendors are all simply working from within the same file system, having the data present in their own infrastructure, serving their high performance applications, but with it being the same data. I was on Threads, which is the new Facebook meta app that's competing against Twitter, and I made a comment, you'll appreciate this. I said, hey, everyone, old IG, young guns out there. I'm the old guy. I've been around many cycles. And I said, how many cycles have you been in? And then everyone goes, what's a cycle? Meaning the cycle in the industry. And I think Dave Vellante calls it ways of innovation or a bunch of points. This next data cycle is an interesting concept because it does highlight how to think differently around what's coming next and what you have now. How does HammerSpace see that evolve for customers because I think that's a key super cloud enabler. If you do it right, you get benefits. If you do it wrong, you don't. Yeah, I think in the next data cycle, it's a big shift from small amounts of structured data to the entirety of people's data needing to be kept online. So forget active archival. It's everything there, right? And the need to have that in different locations so that you can do training so that you can do inference. And so it will be a radical departure from what we know of for those reasons. And you know, this video is actually data. We're going to convert it into the cloud. It'll do general AI, get the clips. We also bump it into our small language model, I guess, compared to the other ones. But we get the data. Everything will be fundamentally digitized. It's data. Every company's impacted by this next wave. And notice how little you have to think about the structure of it. Because the structure is understood now by the AI side of it. So you got a big wave. It's a big inflection point, probably bigger than anything that we've seen. At least I've been saying publicly that I've never seen anything this massive and this impactful, generationally, going back to the PC days when I was younger, saying, well, that's a complete platform shift from the way things were before. Certainly that was a cycle and that was a major inflection point than the web. Now this, I would even consider social media smaller compared to this wave coming with AI. Oh, absolutely. How are you guys looking at this? You got great investors, you got 56 million in cash. How are you going to use the funds? What's your plan? What's your strategy? First, we are super excited about the investors being led by Prosperity 7. That's actually an Evergreen fund from Saudi Aramco. Very innovative, especially in the AI space. And then Kathy Wood with ARC Investments. You know, the thing that's common about all of these is their belief in the technical founder, CEO, which fits me well. You've done well with that. But also that they invest in companies that are disruptive innovators that stand to disrupt the status quo. And in this case, how data has been traditionally stored and managed. But then also the thing that I like is that they are all long-term shareholders. Kathy is mainly playing in the public markets. So this is her crossover fund, her interval fund. And so, you know, the same is true with our other investors like Peer88 and so forth. So every single one of these investors are not the traditional venture capitalist who have to distribute to their LPs as soon as a company- So they're in the long game here. They're in the long game. And that's because we are, you know, getting to an IPO for us is really just getting to the starting gate, right? And we don't want to have to worry about shareholder turnover. So raising 56 million from these very long-term investors, that's very exciting for us. And you've been there, done that with the IPO, track record and another race around the track as they say, another cycle for David Flynn. We saw that with Fusion.io, but this one is even more impactful. And it's because I don't think people understand the impact that moving from traditional storage, copy and merge to a world where data is orchestrated. The TAM here is trillions. I mean, it's not even so much small, I mean, bigger than the Fusion.io space, which was a large market there. That was still massive. This is like every company. And you'll remember how solid state was new when we introduced Fusion.io and people didn't really understand what you were going to do with that kind of performance horsepower and so forth. Well, the same is true here. If we're successful, then people are going to look back in five years and they're going to, 10 years maybe, shouldn't take us that long. And they're going to go, I mean, what on earth did we survive in a day when you had to do manual janitorial work to shuffle data off between these storage systems? I mean, with the acceleration of value we were seeing with AI, it could be five years, I would say 10 to 15, maybe on other cycles, but this next data cycle is going to be highly accelerated because the value, time to value will be faster if done right. So I think there's going to be a lot of kind of broken down cars on the side of the road. You're going to see people speeding through the freeway. You're going to start to see the winners. I mean, it's going to be very clear. Yeah. And the beauty about data orchestration is that for the first time, you can actually use AI to do the pushing of data to where you're going to need it in advance. So you can use AI to be predictive about how data gets placed and moved across a complex infrastructure environment where you have computing assets in many different locations. I think people have not repivoting their business around data, they're going to be extinct. David Flynn, thanks for coming on theCUBE. Thank you so much. Congratulations on breaking the news here on SuperCloud 3. Thanks for helping us. SuperCloud 3 keynote conversation with David Flynn, the CEO of Hammer Space, big player in SuperCloud market, enabling the VAPs and increasing the performance, of course, breaking news here on theCUBE. I'm John Furrier. We'll be right back with more SuperCloud 3 after this short break.