On this episode of the Futurum Tech Webcast – Interview Series I am joined by Renen Hallak, Found and CEO of VAST Data for an exciting conversation about the future of IT and data, and how it is impacting the way we work, live and communicate. It was a fascinating conversation from the leader of this great up-and-coming company.
The Future of IT and Data
In our conversation we discussed the following:
- A quick overview of VAST Data
- An exploration in the high-level shifts we are seeing in enterprise tech
- The opportunities in the data and storage landscape
- How VAST Data works to help organizations overcome the way they typically view storage resources
- Real-world customer success stories that solidify VAST’s position in the market
VAST Data is making a name for itself and it will be interesting to continue to follow their journey over the next few months. I expect we will continue to hear great things from the company. If you’d like to learn more about VAST and what they do, check out their website or watch the full episode below.
Watch my interview with Renen here (and subscribe to our YouTube channel while you’re there):
Or listen to my interview with Renen on your favorite streaming platform here:
Don’t Miss An Episode – Subscribe Below:
Disclaimer: The Futurum Tech Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we do not ask that you treat us as such.
More insights from Futurum Research:
Daniel Newman: Hey, everybody. Welcome to the Futurum Tech Podcast and Futurum Tech TV. I’m your host, Daniel Newman, principal analyst and founding partner at Futurum Research. Excited about this Futurum Tech podcast, webcast, and TV edition of our interview series. And today, I have Renen Hallak, CEO of VAST Data, and he’s going to join me here to talk a little bit about the future of IT, the future of data, the way it is impacting the way we work, the way we live, the way we communicate, and this is a very exciting and up-and-coming company that I think everybody out there is going to have a lot of fun hearing about and learning about. Renen Hallak, welcome to the Futurum Tech Podcast Interview Series.
Renen Hallak: Yeah, thank you for having me.
Daniel Newman: I am really excited to have you here. I’m excited to be in a studio, by the way.
Renen Hallak: Yeah, me too.
Daniel Newman: We’re live, we’re in person. I mean, the people in our community that have been following Futurum in our podcasts, they’ve been watching us do these over Zoom for almost 18 months now. So of course, we’re following all the protocols, we’re being smart, we’re being safe, but we have the chance to sit down live, and I just think there’s so much more energy that comes out of that.
So over the pandemic, I’ve had the chance to get to know you a little bit, your company a little bit, and hopefully, in the show, we’ll get to do that. And then we’re going to talk a lot about some of these just seismic shifts that are going on in the marketplace, but let’s just set the table. First of all, a quick introduction. Tell us about yourself, a little bit about your background, and how you landed at VAST, starting this company.
Renen Hallak: My background is in complexity theory, cryptography, computer science, and have been in the startup world for a while now. My last startup was a company called XtremIO. I was VP of engineering at XtremIO, and that company built a block device, an all-flash block system that was very successful at the time. We were acquired by EMC very early on and were able to hit a $1.2 billion run rate within that first year of sales.
But ultimately, what I realized was that the data space is exploding to much larger levels than what those systems were able to handle, and that inspired us to start a new company, VAST, based on a whole new technology and a whole new architecture for the way we think things should be done.
Daniel Newman: It’s great to hear that. I’m going to want to get deeper into what inspired you to start that, because of course, it sounds like you had a great exit, you watched the company flourish, but most entrepreneurs that I know, that’s, a lot of times, the signal to go do something else. I’m sure that’s got something to do with it.
Let’s back up a second though because another source of inspiration for a lot of entrepreneurs is identifying changes, shifts, new innovations that are going to come into the market. Just give me some broad things that you’re seeing in tech. What are some of the high-level shifts and changes that you’ve noticed in enterprise tech?
Renen Hallak: Yeah, I think the main change that we’re seeing is around data, and data is changing both in its form and in the access patterns. 10 years ago, it was all about rows and columns and transactional databases and the beginning of analytics. We’re now seeing more natural data, pictures, video, genomes. Things that computers really didn’t have any business analyzing before are now, through new algorithms and applications, such as AI and machine learning and deep learning, are becoming much more interesting. And that is also transforming the access pattern because these data sets are so much larger than the ones that we had before, but they need very fast access in order to run these training and inference algorithms.
Daniel Newman: Yeah, I’m glad you pointed those out. On our show, we talk a lot about chips and SAS, we always say, and I think there’s a really interesting parallel path that is taking place in terms of architectural developments at the semiconductor level. The world heard so much about the shortage and what’s going on, and some of that is sensational, some of that’s very real. We’re seeing, of course, a Moore’s law slowing down, which is forcing different kinds of innovation, whether that’s packaging, chiplets. We’re seeing us get down to, what, now, single-digit nanometers, three.
And I guess, as you say this, I’m thinking, all this stuff is so tightly aligned. We’re seeing some of the most innovative semiconductor companies, like the NVIDIAs of the world, building these DPUs to now offload different parts of the processing so that networking and security and data and storage and AI are all being done. And it’s creating a huge opportunity though, to rethink the whole data and storage landscape, right? And I mean, that’s really something that, at least when I was listening to your story early on, that I wanted to hear a little bit more from you though, is it seems like part of the inspiration of this whole thing is that you observed that the way it’s been done might not be the way the market needs to think about storage going forward.
Renen Hallak: Yes, so I think there are two parts to it, the applications that we already discussed, and the hardware elements that are enabling factors. And so, VAST is 100%, a software company, but we definitely use the latest and greatest in hardware technology in order to build new architectures that allow us to solve these age-old problems that have existed for a long time.
Daniel Newman: Yeah, because if I’m not mistaken though, there’s a lot of limitations in tiered storage. That’s something we’ve observed, right? We’re hearing more about this universal data, this trend to going towards… Look, in tiered storage, it had its place, and it’s not to say that it’s going to just flip the switch, and overnight, it’s just going to go away. But what we have now is, well, we’ve got this vast world of data, structured data, unstructured data, real-time data that’s being created. You’ve got the edge, you’ve got IoT. We’ve got user-generated content at scale. And by the way, so much more of the data isn’t in our data centers anymore, right? It’s being created in real time. At this moment, as we speak, we’re creating data, and the systems weren’t really designed for that.
Renen Hallak: They weren’t. So the tiering of storage and the tiering of data is a consequence of limitation. You used to have very fast systems that were very expensive and very small, and then very large systems that were cost-effective, but you couldn’t really access data off of an archive. And so, it fit the world of us human beings being the ones reading the information because you could say, “This is new, it’s hot, this is relatively new, it’s warm, and this is a month old. No one’s ever going to read it again.”
But with these new applications and algorithms, you really want fast access to the entirety of the dataset, and so the whole concept of tiering is now obsolete, and that’s the problem statement. The nice thing about the last few years is that these new hardware pieces and other technologies are enabling us to build a brand-new architecture that allows us to break those age-old trade-offs between price and performance and capacity and resilience, and to build one system that really does collapse that pyramid of storage tiers.
Daniel Newman: So VAST, you guys, at least as I’ve listened and observed, really don’t want to be associated with storage, at least as it’s understood today. At the same time, you are a software company that’s addressing that tiered storage problem that we just reviewed. There’s some fundamental procurement, buying, CIO behaviors that, of course, you have to be changing because this is a behavioral shift. You’re a software company shifting the way that entire enterprises need to think about how they procure, manage, utilize, leverage data. How do you overcome that? How do you overcome the folks that have become so used to buying storage in a certain way and say, “We don’t want to talk about storage. Let’s talk about something new”?
Renen Hallak: Yeah, so I think the CIOs of the world and the IT administrators are realizing that they’ve been buying in the wrong way. They’ve been buying software as an uplift to hardware, and they’ve been overpaying as a result, and they’ve been buying things that they don’t really need as a result. What we did from a business perspective is, in the same way that we’ve dis-aggregated capacity from performance and state from logic on the technology, we’ve dis-aggregated hardware from software. And so now, our customers, just like a cloud provider, are able to purchase their hardware at cost or leverage somebody else’s hardware as infrastructure and put our software on top and license exactly what they need from the software perspective.
Daniel Newman: But for all those companies out there that have made mega investments in NAS. This has got to be causing a little bit of indigestion, right? We’ve spent all this money. We want to use what we’ve bought. That’s probably, by the way, one of the biggest challenges with technical debt in any part of the stack, right? We bought this big, expensive ERP. We’re going to use it. And it’s like, well, there’s a much better mouse trap, and it is costing you productivity and revenue, and it’s costing you innovation. But at the same time, we’re committed, right? This is a human behavioral issue, by the way. How do you unwind that? How do you get them to move faster, because like I said, there has to be some sort of chart or something you have to create that says, this is the time value that’s being lost?
Renen Hallak: Well, it’s a lot easier than that, in fact, because as these organizations are moving to these new applications and workloads, they’re hitting a wall, and they notice when they hit that wall because the legacy NAS systems are all inadequate in providing the benefits that are required for these new applications. And so, whenever a customer starts using GPUs or any type of advanced analytics or AI, they very quickly realize that the shared-nothing architectures don’t provide what they need for that. They don’t provide the scale, they don’t provide the performance, they don’t provide the combination of scale and performance in the way that our shared-everything architecture does.
And so, a customer would start out by purchasing VAST for that workload, that doesn’t work well on the legacy NAS. And then very quickly thereafter, they realize, “This is actually faster than my tier one,” and they would start putting tier one workloads on us, like VMs and transactional databases. And then they realize, “This is actually more affordable and easier to use than my backup system,” and they would start directing backups to us, and that’s where we realized the true vision of universal storage. When there aren’t any trade-offs anymore, you don’t need all of that complexity.
Daniel Newman: So you’ve read the book Consumption Economics, right?
Renen Hallak: I have not.
Daniel Newman: Okay. Well, make a note. You just talked about something called land and expand. The VAST strategy has been land-based upon addressing the most irritating problem related to cold storage and not being able to quickly access data. A lot of this is being driven by the desire to use machine learning and AI and having data that’s just not accessible. And it starts there, and like you said, then there’s things that are being used for pretty much real-time or semi-real-time that you are able to, through software and the innovation that VAST is building, to do better than what their current investments are able to support.
So I mentioned the AI thing, and so as I learned more about the company, that’s what comes to mind. This is the killer app for VAST. But it’s got to be more than that because if it’s just that alone, it’s going to create a lot of competition, it’s going to bring a lot of entrance into the market. You’re building something that really wants to address storage across the entire enterprise, not just for AI and ML, right?
Renen Hallak: So what we realized was that AI and ML are the hardest problems to solve, and so those are the workloads where you keep data forever, and so it grows into very large amounts. Those are the workloads where you need to read it over and over and over again, many times, in a random access fashion, and so we’ve built the system to be agnostic to the workload, to the sides. It gets much better as it gets bigger, and there’s really nothing to manage, and so it’s very, very easy to use. And once you’ve solved that hard problem, you realize that you’ve, as a by-product, solved all of the other data and storage problems that existed in a way that you can now have a single platform that gets used for everything, and storage is not something that you need to think about anymore.
Daniel Newman: Okay, yeah. I agree with that. I agree that when you solve the hardest problem, by nature, you probably have the propensity, as long as the economics work, meaning that you’re not paying exponentially more to do the easier problems that have cheaper technologies, that already exist to solve.
So the other area I think you run into is, there are some very, very big, very successful, resource-laden companies out there that are in stores and aren’t just going to… They’re not just going to jump out and say, “Hey, VAST, take the market from us,” right? I mean, you’re going to have to disrupt them and it’s going to have to be material enough that you’re able to pick off business.
And of course, you’ve gotten to this point, you’ve got the valuation you’ve built because the investor community believes in you, you’re starting to win business, customers are believing in you, but how do you differentiate from what other companies… because I got to imagine universal storage isn’t… it’s not going to start and end with VAST. All of these companies that have been making their living for a long time are going to find a way to compete in this space.
Renen Hallak: Yes, so we’ve, so far, deployed several exabytes of effective capacity, of effective data being stored on VAST, and the reason is, in one word, the architecture. All of those legacy NAS systems, whether it’s EMC’s Isilon, or Pure Storage’s FlashBlade, they are based on a concept where every node in the system has a specific piece of the pie that it’s responsible for, and those nodes need to communicate often in order to serve application requests. That, inherently, was a very good thing 20 years ago when it was invented because it allowed us to go past dual-controller architectures and scale up. But now, we’re seeing the limitations of it because you get diminishing returns the more you scale those systems.
What we’ve done is, we’ve flipped it on its head and said, “Shared nothing is a thing of the past. We want to do shared everything.” Every node in our system has access to the entirety of the namespace through a protocol called NVMe over Fabrics and accessing drives on the other side of the network that are both flash and storage class memory. What that allows us to do is break those fundamental trade-offs, and what it prohibits our competitors from doing is gaining those same benefits without starting from scratch. And so, large organizations have a tendency to be very good at iterating on what they have, not be very good at starting something from nothing, and so far, we’ve seen them try to compete by advancing the state of the art of shared nothing. We have not seen any of them make that switch to shared everything just yet.
Daniel Newman: Yeah, of course. When they see it work, resources will be put in play, but ideally, for VAST, you’ll be at a point where you’ve got enough established foundation and roots in the market that doesn’t cause any real turbulence for you. And of course, if they can’t build it bigger and better, you become a very attractive potential acquisition, which may or may not be your route, but I’m sure, as a great CEO, you would be paying attention to every and all different routes that could help return value to both your customers and your shareholders.
So see, you mentioned that you’re starting to get into the market, you’ve found an identity in terms of what gets that customer excited and wins that piece of business for you. Can you share even maybe an example or two, generically or specifically, about a customer win and what drove that win versus others in the market?
Renen Hallak: Sure, so we see a lot of success wherever there’s a combination of a lot of data and the requirement to access it very quickly, over and over again, and so that ranges from AI and analytics workloads. For example, in life science, we do a lot of genomics and medical imaging and brain research, in the financial space, a lot of backtesting of algorithms in terms of going over trade data and understanding how new algorithms need to trade going forward.
And so, just as an example on that, if a hedge fund before could backtest on one week’s worth of data or two weeks’ worth, because that’s what they could fit on that top tier that they had before, now they can do the entirety of their history 10 years back, 15 years back. And instead of taking four or five hours in being an overnight batch process where the quants need to come in the morning and see what the results are, if it’s good or not, now they can do it in minutes, and so it becomes interactive.
In a completely different space, you can see a lot of media and entertainment news for our system. For example, a large animation company told us that they could not have developed their latest movie without VAST, and they showed us movements in hair and shadows, and they said, “Our animators have been asking for these types of abilities for years now and we couldn’t give it to them because of an infrastructure bottleneck. Now that we have infinite access to all of the information, they can do whatever they want.”
Daniel Newman: Yeah, those are great examples. And of course, as they become more public and some of these wins can get more specific, I’d love for you to come back and share those with me. I’m hearing a lot about regulated industries, so it sounds like the life sciences, it sounds like financials. And those, of course, are industries that tend to have mega data sets that are constantly growing at exponential rates, and it sounds like you’re in a really good position to address that.
I guess, my question now is, it seems like you’ve planted into the market, you’ve raised multiple rounds of capital, you have customers starting to buy from you, and you feel like you have a story that’s winning, by the way, and a winning story is one of the biggest keys to a startup reaching its ultimate potential, but where do you go now? Is it all about nailing this product? Is it about expanding and disrupting other parts of the stack? What’s next for VAST Data?
Renen Hallak: Yeah, so the story we’ve had for a while, and I’m always paranoid, and so I feel like the story’s good, but the product won’t match. And when customers start buying, I feel like, okay, we have some very good salespeople, but they won’t buy again. And then in the SAS space, as you mentioned, chips and SAS, we have a metrics called net dollar retention, which basically says how customers stay with you over time, and a hundred percent, means you’re not losing any customers. Our net dollar retention has consistently been above 300%, and so customers love this and are expanding their systems very, very quickly, and that, in turn, makes them very good partners of ours.
And our customers are at the bleeding edge of building out AI practices and building really scalable AI solutions, and what we’ve found is that we’ve solved the storage problem for them now that they have infinite access to all of their information, but a lot of them are struggling with piecing together the rest of the parts that are required, whether it’s compute or analysis of metadata or building out a global namespace across geographies. All of those parts are other elements that are required to build out this next-generation stack, and we are in a unique position, thanks to our customers, to be seeing that firsthand and to be building on top of this new architecture in a way that couldn’t be done before.
Daniel Newman: I like that answer a lot. And what I actually like about the company, if I was an outsider, looking in, as I am, is that you’re in a… AI is still a pretty nascent state. A lot of people think it’s developed, like 5G. It’s like, no, we’re still very early in terms of this market. A lot of what we think of as AI is really just applied ML or applied analytics, and the work that’s going to have to be done in every industry, the ones you mentioned, the highly regulated, all the way down to things like retail or smart cities, it’s exponential. The volumes of data being created are going to be exponential, which means these problems that these companies are going to be trying to solve for both critical mission type of work, as well as just everyday interactions that we’re going to have, is going to require more and more data, so being in the right place at the right time is always a good thing for a startup, Renen.
Renen Hallak: And we’d like to continue to ride that wave that’s obviously a lot bigger than we are, and we think it’s the biggest wave that has ever been in tech. I saw a chart the other day that shows the amount of market cap built out by the internet is something like $13 trillion, and the amount of market cap expected to be built by this new wave of deep learning in AI is more than $30 trillion, and we believe that. We believe that once we figure out this part of having computers, not just do what we tell them, but actually learn and think and come up with creative ideas, will enable us to solve a lot of the other problems that humanity is facing.
Daniel Newman: It’s very exciting to be part of something, to solve such a big problem, and it is also going to be such a part of the solution to solving so many of these global issues we have, from the next large virus or pandemic to how we can give more equity and equality. All these different things that the world are thinking about, data, technology, are going to help them.
Renen Hallak, it’s great having you here on the Futurum Tech podcast. I hope we can bring you back soon, maybe six months from now, a year from now. I want to hear more, and I’m sure our audience will as well. Renen Hallak, VAST Data CEO. I expect we’ll be hearing a lot more from this company. Check out the show notes. I will share some links so you can learn more about the company, more about Renen.
Also, continue to stay with us. Hit that subscribe button and join us for future episodes of our show, as well as the interview series for this episode. It’s time to say goodbye. We’ll see you later. Thanks for tuning in.