theCUBE + NYSE Wired: Zero Trust Cyber Series | Roger Cummings, PEAK:AIO

Clips
More from theCUBE + NYSE Wired: Zero Trust Cyber Series

Roger Cummings

CEO

PEAK:AIO

play_circle_outline Exploring Peak:AIO's AI-Driven Storage Solutions for Fast, Cost-Effective Performance

play_circle_outline Focus on energy efficiency to save customers up to 50-70% in storage/environmental savings.

Info
Transcript

Roger Cummings, PEAK:AIO

Roger Cummings

CEO PEAK:AIO

PEAK:AIO, a software-defined storage solution for AI workloads, was founded by CEO Roger Cummings in 2021. It offers simplicity in setting up RAID configurations and computational tasks for various AI applications. The system provides high performance and scalability at a competitive price, addressing energy efficiency challenges. With plans for expansion and a growing customer base, PEAK:AIO aims to refine its product-market fit and explore opportunities in financial services and deep analytics.

explore Keep Exploring

What was Mark Klarzynski's role in helping NVIDIA channel with their storage purchases for AI workloads? add

What are some of the challenges faced by customers in terms of power constraints when transitioning to AI technology, and what innovative solutions are being developed to address these challenges? add

bolt Powered by CUBE AI

Roger Cummings, PEAK:AIO

search

Dave Vellante

>> Hi, everybody. Welcome back to Media Week. This is NYSE Wired and theCUBE's ongoing coverage of AI innovators and cyber leaders. And we're super excited to have Roger Cummings here as the CEO of PEAK:AIO, a company you may not have heard of, but doing some really interesting things in software-defined storage and AI. Roger, welcome to theCUBE's NYSE Studio.

Roger Cummings

>> Yeah, thank you. Thank you for having me.

Dave Vellante

>> Great to have you.

Roger Cummings

>> I appreciate it very much.

Dave Vellante

>> Yeah, you bet. Okay, so let's get into it. So you guys started the company in 2021. Right now you're self-funded.

Roger Cummings

>> Correct.

Dave Vellante

>> Tell us about why was the company formed, you came in, what, about eight months ago, and what are you guys all about?

Roger Cummings

>> Yeah, actually it's a great story. Mark Klarzynski is the founder of our company and has been doing this, we both have the gray hair, we've been doing this a long time.

Dave Vellante

>> I dye my hair gray too, Roger.

Roger Cummings

>> So Mark really spent time and was brought in by the NVIDIA channel to help them really uncover why they weren't buying storage with the NVIDIA purchases. So what Mark really found out is this AI workload was different, this AI workload didn't value the enterprise features and functions that were out there. And really it was just looking for fast high-performance storage that could address their solution at a cost-effective manner. So he spent a good solid two years of working with these companies and tried to take the legacy infrastructure and boil it down to something that was consumable at a very small footprint and it just didn't make sense economically, or from a complexity perspective. So he spent the first two years really defining that and providing a product that was very successful. And now we've taken that product, we've made it software-defined so it could sit on any commodity hardware out there and provide customers the performance, price, power, savings and economics to really be successful in AI.

Dave Vellante

>> So it's not HPC, it's really designed for enterprise AI, is that right?

Roger Cummings

>> It's designed for AI from the research person, from the computational biologist, all the way to organizations running larger training modules. So it can scale.

Dave Vellante

>> They would configure or leverage your system in different ways depending upon the scale. What are the dimensions that we should be thinking about there?

Roger Cummings

>> Yeah, a big part about it that we've learned is AI is a lot of times a different customer or doesn't have the support that they need. So a big part of our solution is simplicity. So setting up the RAID configurations, doing all these things, computational biologists or even doctors or some of these folks-

Dave Vellante

>> Guys in lab coats aren't good at that stuff....

Roger Cummings

>> aren't good at that at all. So even at the enterprise level, we make that AI infrastructure faster, a higher performance, dense associated with, and really make it simple for them to use across any AI workload.

Dave Vellante

>> Interesting. Okay. And so where are you at with the buzz phrase, "But it's important, a product market fit"?

Roger Cummings

>> Product market fit, great. So we have some really good use case success stories in healthcare, and healthcare being at the edge. Because we can scale down to a single node and not only run the training at that single node and be dense thanks to our partners at Solidigm as well, you have incredible density at that single node. So we can provide all the training that you need as well as inference at that edge. So we're seeing healthcare applications and radiology and imagery and video. We're seeing mapping chemistry in life sciences. We're seeing government and tactical edge scenarios, financial services, doing deep analytics, and energy doing the three and four dimensional kind of data collection at the edge as well. So we're really excited, the ability to do so much at the edge.

Dave Vellante

>> Okay, so you were brought in to really bring the operational excellence, right, raise some money and to scale the company?

Roger Cummings

>> Correct.

Dave Vellante

>> Before we get off the sort of basic problems that you're solving and the trends, what are the options that are out there today? I could use traditional enterprise storage, you've got emerging players, they got VAST and Hammerspace and stuff like that. So where do you guys fit? Where are you adding value and differentiation relative to the legacy storage players and some of these new startups?

Roger Cummings

>> So if you think about the legacy players as well as maybe VAST and VECA, really it's the complexity, being able to provide the same performance in a single node and then scale out from there. I recently saw a number, 80% of AI workloads are failing. Well, why? It's because you bought this GPU, you now see the data bottleneck, and all of a sudden you see I have to spend as much or more than it did for the GPU just to have the storage infrastructure to feed that GPU. And there's no need for that. We provide a great solution. And the big differentiation is being able to scale down at a price point and a performance that exceeds anybody in the marketplace. Now you bring up Hammerspace. Hammerspace is a partner of ours. We're a great target for Hammerspace.

Dave Vellante

>> Yeah, it makes sense. Right. Okay. One of the biggest problem, if not the biggest problem in AI, the blocker that everybody's talking about and will really come to fruition as we scale this whole wave up is energy and power consumption. When you were at Super Compute, I was as well, I was struck by how many liquid cooling players there are in the market and guys doing hoses. We had Omni Services on theCUBE, "What do you guys do?" They make hoses to cool GPUs. It's like, wow, this is bringing together all kinds of new innovation. So talk about the energy problem and where you guys fit.

Roger Cummings

>> Yeah. And you bring up a great point because when we first started this company, it's like, well, AI is just more storage, take the legacy, move it over here. Well, as a result, we've seen different types of rack systems so we know it's a different type of workload. So we have to address power, and our customers are begging and really are handcuffed from growing with the power constraints that are out there. So we have some innovation where we're able to look at those workloads and really determine the best path to energy efficiency. And we're looking forward, we're still testing it out, but we're looking forward to saving customers anywhere from 50 to 70% in their storage savings or in their environmental savings. We're really excited about that.

Dave Vellante

>> Okay. And so of course the GPU gets most of the blame, but storage still sucks up tremendous quite a bit.

Roger Cummings

>> Tremendous amount, tremendous amount.

Dave Vellante

>> The spinning discs were really culprit. And of course flash is, but flash is growing.

Roger Cummings

>> You're bringing in our gray hair again.

Dave Vellante

>> Yeah, yeah, there you go. Okay. So you mentioned Hammerspace, what's your partner ecosystem and strategy.

Roger Cummings

>> So Hammerspace has an incredible scale out system, and we have a number of customers that rave about it. For us, we fit as a target for Hammerspace that allows them to have the performance that they need at the price point. And that's key is having a cost-effective solution to complement that existing file system that's out there.

Dave Vellante

>> So you're not at the point, I presume, that you're scaling go-to-market, or are you scaling go-to-market?

Roger Cummings

>> Yeah. I came on board to take the great work. We've got some incredible customers in the UK that we're bringing to the United States. So we have 30 plus customers that were taking their use cases, as I mentioned before, here to the United States, and we're scaling that success. And we have wonderful customers for prospects to talk to and really understand their use cases because they're identical to here in the United States, maybe a little larger, but here in the United States as well.

Dave Vellante

>> So 30 plus, these are revenue customers?

Roger Cummings

>> Oh, very much.

Dave Vellante

>> Okay. I was asking about product market fit before, but do you have product market fit? Do you feel like you're there?

Roger Cummings

>> Yeah, I'm a big believer product market fit can always be tweaked and learn more and more.

Dave Vellante

>> Sure, it's a you're never done kind of thing.

Roger Cummings

>> Never done. And that's something that, as a company, we're very close to the customer and trying to understand more and more of what the challenges they face. So we feel fairly comfortable where we are with product market fit.

Dave Vellante

>> Well, so as a CEO of a blank sheet of paper startup, how do you even determine product market fit? Market says, "Well, it's when your servers are blowing up and you can't meet the demand." Okay, but maybe there's a more scientific way to evaluate that.

Roger Cummings

>> Well, there's a lot of metrics that we throw behind this and its utilization, how they're using our solution and what they're using our solution for. And the biggest point that we drill down on from market fit is what was it before and what was it after? And the significance that we can make in the before and after from performance, price, power, and density.

Dave Vellante

>> So it's really around adoption, how they're using it, the use cases, how sticky they are, how much land and expand you do. The reason I ask is because a lot of startups, they try to scale go-to-market before they have product market fit.

Roger Cummings

>> Exactly.

Dave Vellante

>> And as you well know, Roger, in the early phases of go-to-market, your sales motion is you're consultants, you're technical consultants, you go deep, you got to build trust with the customer. It's not a transactional business, it's really you literally are an extension, you're a consulting arm essentially.

Roger Cummings

>> Exactly right. And I had the pleasure of working with David Scott back in the day, and David was the king of metrics.

Dave Vellante

>> David Scott of 3PAR fame.

Roger Cummings

>> Exactly right.

Dave Vellante

>> He made HP at the time 3PAR, and he was an amazing executive.

Roger Cummings

>> Yeah. So David Scott has really taught me early in my career that, "Listen, reduce friction. Whatever you're doing, get close as you can to the customer and reduce that friction. Understand what he wants out of that technology and work as hard as you can to make it a seamless path to the success of that customer."

Dave Vellante

>> And that's what 3PAR... When did you work with David? Was it at HPE?

Roger Cummings

>> Actually, David Scott was on our board at AppIQ.

Dave Vellante

>> Oh, right, okay. And didn't get acquired by HP?

Roger Cummings

>> AppIQ got acquired by HP.

Dave Vellante

>> And then 3PAR got acquired by HP.

Roger Cummings

>> Yeah, I believe 3PAR was acquired a little bit before us, but yes, I believe so.

Dave Vellante

>> Either way.

Roger Cummings

>> We've been doing this too long.

Dave Vellante

>> David Scott, great human being-

Roger Cummings

>> Yes, very much so....

Dave Vellante

>> and an amazing leader. So what's happening here at the New York Stock Exchange? You got invited by Brian, I presume.

Roger Cummings

>> So I was invited by Brian. We have a number of opportunities that we're talking to, learning from, as I mentioned. And we're excited to be here. We're talking a few a little bit later today. And financial services and the deep analytics they want to do across various verticals, product and information verticals they have is really exciting for us. I think we have a great opportunity to provide some value.

Dave Vellante

>> Amazing. Well, Brian, good luck with the raise, the scale. Would love to have you back and track the progress. You're based in the Midwest in Chicago.

Roger Cummings

>> Correct?

Dave Vellante

>> We have studios, one in the West Coast in Palo Alto, two in the East Coast, one here in New York and one outside of Boston. So I'd love to have you back and see how you're doing.

Roger Cummings

>> Appreciate your time very much.

Dave Vellante

>> All right, you're very welcome. All right, thank you.

Roger Cummings

>> Thank you.

Dave Vellante

>> Now, keep it right there everybody. This is Dave Vellante for John Furrier, NYSE Wired and theCUBE community's Media Week covering AI innovators and cyber. Keep it right there, right back after the short break.