 Welcome back to Los Angeles. The Cube is live. It feels so good to say that. I'm going to say that again. The Cube is live in Los Angeles. We are at KubeCon CloudNativeCon 21, Lisa Martin with Dave Nicholson. We're talking to StormForge next. Cool name, right? We're going to get to the bottom of that. Please welcome Matt Probo, the founder and CEO of StormForge and Tom Ellery, the SVP of revenue. StormForge, guys, welcome to the program. Thanks for having us. So Matt, StormForge, you have to stay at like that. I feel like, do you guys wear stormtrooper outfits on Halloween? Sometimes stormtrooper. The colors are black. We hit anvils from time to time. I thought they saw lightning with it. There may or may not be a heavy metal band that might be infringing on our name. It's all good. That's where we come from. So you started the company in 2015. Talk to me about the genesis of the company. What were some of the gaps in the market that you saw that said, we got to come in here and solve this? Yeah, so I was fortunate to always know. I think when you start a company, sometimes you know exactly the set of problems that you want to go after and potentially why you might be uniquely set up to solve it. What we knew at the beginning was we had a number of really talented data scientists. I was frustrated by the buzzwords around AI, machine learning, went under the hood. There's really a lot of vaporware. And so at the outset, really, the point was build something real at the core. Connect that to a set of problems that could drive value. And when we looked at really the beginnings of Kubernetes and containerization five, six years ago at its genesis, we saw just a bunch of opportunity for machine learning to play the right kind of role if we could build it correctly. And so at the outset, it was what's going on? Why are people moving workloads over to containers in the first place? And because of the flexibility and the portability around Kubernetes, we then ran into quickly its complexity. And within that complexity was really the foundation to set up the company and the solution for a set of problems uniquely and most beneficially solved by using machine learning. And so when we sort of brought that together and designed out some ideas, we did what any founder with a product background would do. We went and talked to a bunch of potential users and kind of tried to validate the problems themselves and got a really positive response. So Tom, from a business perspective, what attracted you to this? Well, initially I wasn't attracted, just I'll say that. Just from a startup standpoint, so I've been in the industry for 30 years. I've done six or seven pre-IPO companies. I was exiting a private company. I did not want to go do another startup company. But being in the largest enterprise companies for the last 20 years, you see Kubernetes like wildfire in these places. And you knew there was a huge amount of complexity and sophistication when they deployed it. So I started talking to Matt early on, he explained what they were doing and how unique the offering was around machine learning. I already knew the problems that customers had at scale with Kubernetes. So it was for me, I said, all right, I'm going to take one more run at this with Matt. I think we're in a great position to differentiate ourselves. So that was really the launchpad for me, was really the technology and the market space. Those two things in combination are very exciting for us as a business. And a couple bottles of amazing wine and a number of dinners, that helps as well. Behind the scenes. That definitely helps. That twisted Tom's arm. Yeah, yeah, yeah, yeah. Matt, tell us, just really kind of get into the technology. What does it do? How does it help facilitate the Kubernetes environment for customers? Absolutely, so when organizations start moving workloads over to Kubernetes and get their applications up and running, there's a number of amazing organizations, whether it's through cloud providers or otherwise, that sort of solve that day one problem, those challenges. And as I was mentioning, they move because of flexibility. And so developers love it and it starts to create a great experience. But there's these set of expectations. And let me just, where typically are these moving from? What are the top three environments these are moving out of? Yeah, I mean, of course, non-containerized environments more generally. They could be coming from bare metal environment, they could be coming from kind of a VM driven environment. So when you look back at kind of the growth and genesis of VMs, you see a lot of parallels to what we're seeing now with containerization. And so as you move, it's exciting. And then you get smacked in the face with the complexity. For all of the knobs that are able to be turned within a Kubernetes environment, it gives developers a lot of flexibility. These knobs, as you turn them, you have no visibility into the impact on the application itself. And so often, organizations are becoming more agile, shipping code more quickly. But then all of a sudden, the cloud bill comes and they've over provisioned by 80, 90%. They didn't need nearly as many resources. And so what we do is we help understand the unique goals and requirements for each of the applications that are running in Kubernetes. And we have machine learning capabilities that can predict very accurately what organizations will need from a resource standpoint in order to meet their goals, not just from a cost standpoint, but also from a performance standpoint. And so we allow organizations to typically save, usually between 40 and 60% off their cloud bill and usually increase performance between 30 and 50%. Historically, developers had to choose between cost and performance. And their world view on the application environment was very limited to a small set of what we would call parameters or metrics that they could choose from. And machine learning allows that world to just be blown open. And not many humans are sophisticated in the way we think about multi-dimensional math to be able to make those kinds of predictions. You're talking about billions and billions of combinations, not just in a static environment, but on an ongoing basis. And so our technology sits in the middle of all that chaos and allows organizations just to reap a whole lot of benefits that they otherwise may not ever find. Those numbers that you mentioned were big from a cost savings perspective and a performance increase perspective, which is so critical these days as in the last 18 months, we've seen so much change. We've seen massive pivots from companies in every industry to survive, first of all, and then to be able to thrive and be able to iterate quickly enough to develop new products and services and get them to market to be competitive. Yep, yeah, exactly. Yeah, sorry. I mean, the thing that's interesting, there was an article by Adresin Horowitz, I don't know if you've taken another cloud paradox. So we actually, if you start looking at that, great example would be some of these cloud companies that are growing like astronomical rates. Snowflakes, like phenomenal what they're doing. But go look at their cogs and what it's doing also. It's growing almost proportionally as the revenue's growing. So you need to be able to solve that problem in a way that is sophisticated enough with machine learning algorithms that people don't have to be in the loop to do it and that the math can prove out the solution as you go out and scale your environments. And a lot of companies now are all transitioning over to SaaS based platforms and they're gonna start running into these problems as they go to scale. And those are the areas that we're really focused and concentrating on as an organization. As the leader of sales, talk to me about the voice of the customer. What are, you've been there six months or so? Yes. We heard about the wine and the dinners is an obvious sway. We haven't done a lot of that over the last 18 months. You'll have to make up for lost time, Matt. As soon as he closes more business. Oh, there we go. We got that on camera. You know how that works, right? Sorry, Tom, we got that on camera. But talk to me about the voice of the customer. There's been three market spaces that we've had some really good success in. So I talked about a SaaS marketplace. So there's a company that does Drupal and Matt knows very well up in Boston, Aquia. And they have, every customer is a unique Snowflake customer. So they need to optimize each of their customers in order to ensure the cost as well as performance for that customer on their site works appropriately. So that's one example of a SaaS based company that where we can go in and help them optimize without humans doing the optimization and the math and the machine learning from StormForge doing that. So that's an area. The other area that we've seen some really good traction in the candidates with GSIs. So part of our go-to-market model is with GSIs. So if you think about what a GSI does, a lot of times customers are struggling either initially deploying Kubernetes or putting it in for 12, 18 months and realizing we're starting to scale. We got all kinds of performance issues. How do I solve that? A lot of these people go to the Accentures, the Cognizance and other ones and start flying their Ninjas in to kind of solve the problem. So we're getting a lot of traction with them because they're using our tool as a way to help solve the customer's problems and they're in the largest enterprise customers as possible. So if I'm hearing what you're saying correctly, you're saying that when I deploy serverless applications, I may in fact get a bill for servers that are being used. Is that what you're telling us? There in fact may be a bill for what was coined as serverless that is very difficult to understand by the way. That's crazy talk, Matt. Connect back, yeah. But absolutely we deal with that all the time. It's a painful process from time to time. But have you seen the statistics that's going on with how people, I mean there was huge inertia from every CIO that you had of a cloud strategy in place. Everyone ran out and had a cloud strategy in place and then they started deploying on Kubernetes. Now they're realizing, oh wow, we can run it but it's costing us more than it ever cost us on-prem and the operational complexity associated with that. So there's not enough people in the industry to help solve that problem, especially at the grassroots. That's where you need sophisticated solutions like storm forging machine learning to help solve this at scale problem in a way that humans can never solve. And I would just add to that that the same humans managing the Kubernetes application environments today are likely the same humans that were managing it in a VM world. So there's a huge skills gap. I love what Kastin announced at KubeCon this year around their learning environment where it's free. Come learn Kubernetes in this and we need more of that. There's an enormous skills gap and the problems are complex enough in and of themselves. But when we have, when you add that to the skills gap it presents a lot of challenges for organizations. What are some of the ways in which you think that gap can start to be made smaller? Yeah, I mean, I think as more workloads get moved over, over, you know, over time, you see, you see more and more people becoming comfortable in an environment where scale is a part of what they have to manage and take care of. I love what the Linux Foundation and the CNCF are doing around Kubernetes certifications, you know, more and more training. I think you're gonna see training, you know, availability for more and more developers and practitioners be adopted more widely. You know, and I think that, you know, as the tool chain itself hardens within a CICD world, in a containerized world, as that hardens, you're gonna start seeing more and more individuals who are comfortable across all these different tools. If you look at the CNCF landscape, I mean, today compared to four or five years ago, it's growing like crazy. And so, but there's also consolidation taking place within the tools and people have an opportunity to learn and gain expertise within those, which is very marketable by the way. Absolutely. My employees often show me their LinkedIn profiles. They remind you of that. And remind me of how much they're getting recruited. But they've been loyal, so it's been fantastic. There are so many parallels when you look at VM and virtualization and what's happening with Kubernetes. Obviously, all the abstractions and stuff. But there was this whole concept of VM sprawl, you know, maybe 10 years in. If you think about the Kubernetes environment, that is exponentially bigger problem because how many they're spinning up versus how many you spun up in VM. So those things ultimately need to be solved. It's not just gonna be solved with people. It needs to be solved with sophisticated software. That's the only way you're gonna solve a problem at scale like that. No matter how many people you have in the industry, it's just never gonna solve the problem. So when you're in customer conversations, Tom, what do you say are the top three differentiators that really set StoreCorge apart? Well, so the first one is we're very focused on Kubernetes only. So that's all we do is just Kubernetes environment. So we understand not just the applications that run in Kubernetes, but we understand the underlying architectures and techniques, which we think is really important from a solution standpoint. So you're a specialist? We are absolutely specialists. The other area is obviously our machine learning and the sophistication of our machine learning. And Matt said this really well early on. I mean, the buzzwords are all out there. You can read them all over the place for the last five to seven year AI and ML. And a lot of them are very hollow. But our whole foundation was based on machine learning and PhDs from Harvard. That's where we came out of from a technology background. So we were solving more, we weren't just solving the Kubernetes problems, we were solving machine learning problems. And so that's another really big area of differential for us. And I think the ability to actually scale and not just deal with small problems, but very large problems, because our focus is the Fortune 2000 companies. And most of them have been deploying like financial services and stuff, Kubernetes for three, four or five years. And so they have had scale challenges that they're trying to solve. Yeah. Lisa and I talk about this concept of machine learning and looking under the covers and trying to find out, is the machine really learning? Is it really learning? Or is it people are telling the machine, you need to do this if you see that? Or is the machine actually making those correlations and doing something intelligently? So can you give us an example of something that is actually happening that's intelligent? Well, so the if this, then that problem is actually a huge source of my original frustration for starting the company. Because you tag AI as a buzzword onto a lot of stuff. And we see that growing like crazy. And so I literally at the beginning said, if we can't actually build something real that solves problems, we're going to hang it up. And as Tom said, we came out of Harvard and there was a challenge initially of, are we just going to build like a really amazing algorithm that's so heavy it can never be productized or commercialized? And it really should have just stayed in academia. And I will say a couple of things. One is I do not believe that black box AI is a thing. We believe in what we would call human augmented AI. So we want to empower practitioners and developers into the process instead of automate them out. We just want to give them the information and we want to save time for them and make their lives easier. But there's a kill switch on the technology. They can intervene at any point in time. They can direct the technology as they see fit. And what's really, really interesting is because their worldview of this application environment gets opened up by all the predictions and all of the learning that actually is taking place. And because that worldview is open, they then get into a tinkering or experimental mindset with the technology. And they start thinking about all these other scenarios that they never were able to explore previously with the application. And so the machine learning itself is on an ongoing basis understanding changes in traffic, understanding in changes in workloads for the application or demand. If you thought about surge pricing for Uber because of a big game that took place. And that change in peaks and valleys and demand, our technology not only understands those reactively, but it starts to build models and predict proactively in advance of the events that are going to take place on what kind of resources need to be allocated and why. That's the other piece around it. Often solutions are giving you a little bit of the what, but they certainly are not giving you any explanation of the why. So the holy grail really, like in our world, is kind of truly explainable AI, which we're not there yet. Nobody's there yet. But human augmented AI with actual intelligence that's taking place that also is relevant to business outcomes is pretty exciting. So that's why we try to operate. Very exciting. Thanks for joining us talking to us about StormForge. I feel like we need some StormForge t-shirts. What do you think? I think we can get you that out of it. Can you make that happen? All right. I'm not even asking for the bottle of wine. I'm just a t-shirt. We'll wrap it in a bottle of wine. Oh, I like that idea. That's a good idea. Matt and Tom, thank you so much for joining us. Exciting company, congratulations on your success. And we look forward to seeing what great things are to come from StormForge. Awesome, thanks so much. Thanks so much for the time. Thank you. Our pleasure. For Dave Nicholson, I'm Lisa Martin. We are live in Los Angeles. We're at the Cube, covering KubeCon and CloudNativeCon 21. Stick around. Dave and I will be right back with our next guest.