VAST Forward 2026 | John Mao, VAST

Clips
More from VAST Forward 2026

John Mao

Global Head of Business Development

VAST Data

play_circle_outline VAST's AI Operating System: Hosting Hardware and Apps Through a Strategic Compute–Storage–Networking Alliance

play_circle_outline CrowdStrike Falcon–VAST Native Integration: Protecting Identity, Data Access, and Preventing Exfiltration by Autonomous AI Agents

play_circle_outline Integration options: CrowdStrike agents on VAST or invoking remote services; business model TBD

play_circle_outline TwelveLabs Partnership Brings Frontier Computer Vision On-Prem for Regulated Industries, Fast Video Understanding, Secure Visual Search

Info
Transcript

John Mao, VAST

John Mao

Global Head of Business Development VAST Data

In this segment from VAST Forward 2026, John Mao, vice president of business development and alliances at VAST Data, joins theCUBE’s Rebecca Knight and Dave Vellante to discuss VAST’s evolution into a comprehensive AI operating system. Mao explains how VAST is building a strategic alliance ecosystem that bridges the gap between hardware and applications to serve as the foundational layer for the modern AI lifecycle. A major highlight of the discussion is the native integration with CrowdStrike, which brings best-in-class cybersecurity directly into the VAST p... Read more

explore Keep Exploring

What does it mean for VAST to become the AI operating system, and what is your team's role in achieving that? add

What does the announced integration with CrowdStrike's Falcon platform mean for the AI operating system (VAST AI OS) and how will it affect the future approach to securing AI applications? add

How will CrowdStrike's technology be integrated with VAST's platform — for example, will CrowdStrike agents run within VAST and how will their services (threat intelligence, detection/response, data-exfiltration protections) be used? add

What does TwelveLabs' technology do for video understanding, and why are they partnering with VAST AI OS? add

bolt Powered by CUBE AI

John Mao, VAST

search

Rebecca Knight

>> Good afternoon everyone, and welcome back to theCUBE's live coverage of VAST Forward 2026 here in Salt Lake City. I'm your host, Rebecca Knight, alongside Dave Vellante, my co-host and analyst. I'd like to welcome back to theCUBE, John Mao, he is the Vice President Business Development and Alliances at VAST. Thanks so much for coming on today.

John Mao

>> Thanks for having us. Thanks for being here.

Rebecca Knight

>> Yeah, no, we're so excited to be at your first ever customer conference.

John Mao

>> It's amazing. It's amazing.

Rebecca Knight

>> It is. It is. I'm sure for you who's been almost here since the beginning. I mean, yeah.

John Mao

>> Yeah, I've been here for six years now. I mean, I looked in the crowd today during the keynote and it was kind of hard to take in. We are such a small company when we started and now it's amazing.

Dave Vellante

>> You're need a bigger boat for next year.

John Mao

>> We are going to need a bigger boat. We're excited. We're excited.

Rebecca Knight

>> Well, let's actually talk about that for a second, because you sit at the center of VAST, for lack of a better word, VAST alliance ecosystem. At this year's forward, several major partnerships were announced, which we're going to talk about all of them, but I'd like to just hear you talk in broad brushstrokes about what this moment says about VAST and how it's thinking about its ecosystem versus platform.

John Mao

>> Yeah, I mean, we've been talking about VAST becoming the AI operating system, and I think it's a very great analogy. Because if you think about an operating system, it has really two main purposes. One, which is to serve as the layer that runs on hardware, right? That's obviously very important. And then two, it's to serve and host applications at the end of the day. So from my team's perspective, that's kind of our charter. That's our charter is to go figure out how to build strategic partnerships with hardware vendors to make sure we're ... And that's everything from compute to storage technologies to networking technologies, to obviously the accelerated compute technologies. And then similarly, my team is actually divided like this. We have another half of my team that's focused on applications, whether applications at the solution level. Now, in the AI world, it's more about models as another type of application as well. Yeah, it's very exciting. I am very fortunate and lucky because I get to sit right in the middle of all of the innovation and the technology that's happening with our partners. So, yeah, it's fun.

Dave Vellante

>> So let's get into some of these announcements. So CrowdStrike, obviously, a gold standard for cybersecurity. You're building CrowdStrike natively into your platform. So when I think of cloud native, AI native, we throw these words around, what did you guys announce and what does it mean to be sort of native inside of your stack?

John Mao

>> Yeah, I mean, I think it's two-sided. So we have some native integrations that fit now into CrowdStrike's Falcon platform, and we have large customers already doing that and using that today. But I think the more interesting part of the announcement is we're bringing their technologies into the AI operating system as well. To run natively means being a part of the pipelines in which we're hosting now, the AI life cycle, that's going to run more and more on top of the VAST AI OS. So they have a great set of technologies around data access. We do a lot of the data layer from a protection perspective. Those are very complimentary, but they also do a lot around things like identity access. If you think about the future, this is where my brain hurts sometimes thinking about this, but we're moving from this world of humans having a security policy, a profile. You can access this, you can do that to autonomous AI agents who have now authority to not only read and access maybe enterprise's data, but eventually take action. And that requires a whole new mindset and approach to securing AI applications in the future that a lot of old approaches and old solutions may not been a really good fit for. So, yeah, we're super excited. They bring best in class technology. We're looking at lots of different ways to integrate up and down the stack with us.

Dave Vellante

>> So what does that mean? You bring the CrowdStrike agent into VAST?

John Mao

>> There's CrowdStrike agents that can come into our platform. We can also integrate them into our pipeline by invoking specific services that maybe even hosted remotely. So there's lots of different paths in which customers could eventually consume this, but it is about inserting different capabilities. We're not naive to think we can do everything. We're not a security expert. We're not doing all of the things that those companies do in terms of tracking, bad actors and things that are out there from a software perspective, so we need to lean on them.

Dave Vellante

>> Then their threat intelligence is amazing. But even, John, within the CrowdStrike Falcon platform, I mean, there's a lot. So I presume you're ... It's early in the relationship.

John Mao

>> It's early, yeah.

Dave Vellante

>> So you're trying to pick your spots.

John Mao

>> Yes.

Dave Vellante

>> Is it detection and response? Is it stop the breach as George Kurtz? I mean, where do you start? Because that's a VAST.

John Mao

>> I think there's two logical places.

Dave Vellante

>> Logical places?

John Mao

>> Yeah, there's two logical places to start. One is around data access and then one is around more on the data exfiltration side. So that's a big common problem. We've seen it on, it's published everywhere. When you roll out AI, one of the big concerns is always around data leakage. What goes out and goes through those models, they have sets of technologies that can focus on that. How do we insert those into these pipelines, into these applications that will eventually run on VAST?

Rebecca Knight

>> So is this partnership in your mind a signal that AI infrastructure and cybersecurity are no longer separate conversations? I mean, if they ever were really because-

John Mao

>> No, I think this is ... When I talk to enterprise customers, many of them ... A lot of people have kind of said, "Oh, enterprises are slow. They're kind of laggards moving in AI from an adoption perspective, but they have serious concerns." This is one of them. You can't go to production as an enterprise and have an accident when things like this leak out or people access the wrong information or maybe-

Rebecca Knight

>> Or the agents accessing the wrong-

John Mao

>> That's right. Or maybe even basic things like, obviously, supply chain risk. At the hardware level, you're familiar, but at the software level it's the same. And now, it's within models. It's also the same. So it's just a bigger surface area for the enterprise to have to wrestle with on top of all the new technologies. So I think they will start to fuse and become one ... Not one and the same, but they have to be thought of together as a single system.

Dave Vellante

>> What's the business model? Is it an OEM deal? Is it a licensing deal? Is it capacity?

John Mao

>> We haven't announced that yet, but more and more details of that will come out over the next several months.

Dave Vellante

>> Let's talk about TwelveLabs. I think Jeff introduced this on stage.

John Mao

>> He did. Yeah.

Dave Vellante

>> And he talked about they're actually a frontier model for video. So explain to the audience. There aren't a lot of frontier models out there.

John Mao

>> There aren't.

Dave Vellante

>> So throw that word around. So what makes them a frontier model and the video specialty? Explain.

John Mao

>> Yeah, so frontier model of companies, obviously, the big names like an OpenAI.

Dave Vellante

>> AI, Anthropic.

John Mao

>> That's right.

Dave Vellante

>> Groq.

John Mao

>> So they built a model from scratch. They built their own, they trained their own models, and they iterate on those models and they improve on those models, and they build subsequent applications based off of those models.

Dave Vellante

>> Versus, say, DeepSeek. That's not a frontier model, but it's a derivative, right?

John Mao

>> It could be derivative. Yeah. So there are a lot of derivatives, tons of derivatives now because there are open source foundational frontier models as well. Obviously, Llama is a big one out of Meta and so forth. TwelveLabs is a special company. They can focus on computer vision. And this is really exciting to me, because we now have the technology in AI that can actually look at video and understand it. They can tell you exactly what's happening scene by scene as if a human was watching. But at 1000x fast-forward speed. So what humans would normally take linear time to watch, AI can watch it in nonlinear time and understand everything and remember everything and connect to everything. That's wild to me. They can also provide capabilities to do visual search. So everyone's familiar with a visual search of you take a picture and you can find something. They can do similar things now in video by looking at a specific, maybe it's specific scene, and then look for that same occurrence across a massive amount of data. So they have two models that they built. Their foundational models, like you mentioned Dave, and they've been doing really well. They've been really focused in the early days in media and entertainment. Natural, right? Makes sense. They've been doing that predominantly in the public cloud. And what they realize is that there is a much bigger market outside of media entertainment for their models, for their, and so they've been actually going into a place that we've been also focused on, which is in other markets that are also computer vision oriented, but need to stay on-prem. Or maybe it's a regulated industry or there's a lot of security concerns, or maybe for some public government entities they have special needs. And so the partnership is about them now bringing a lot of their technologies onto the VAST AI OS to help enable some of these types of workloads.

Dave Vellante

>> I mean, the investments required to build a frontier model are massive, particularly because you need a lot of compute. So my obvious question is what's their kind of protection against the mainstream, the big four LLM vendors? They have some unique IP, they've got business traction. What's their kind of moat?

John Mao

>> And I am not the expert, so I don't want to put the wrong words out there.

Dave Vellante

>> Yeah. I shouldn't asked you to speak for TwelveLabs.

John Mao

>> But they do have unique-

Dave Vellante

>> But they're not widely known and they are in the community.

John Mao

>> Like I said, they're very specialized in computer vision. I've heard this directly from our customers who have smarter people than me. The CTOs of a global bank, for example, told me recently that he's looked at a whole bunch of different computer vision models and theirs is the best. So that's my only data point I can give you.

Rebecca Knight

>> My question for context is why is this computer vision and video intelligence becoming more of a priority for organizations right now?

John Mao

>> Okay, yeah, that's great.

Rebecca Knight

>> Because you had said this is used in predominantly in media and entertainment. Yeah, I get that, but what about for the global financial institutions?

John Mao

>> So in this case, a lot of the global banks are starting to for, I don't know exactly the catalyst for this, but I've seen this pattern now where a lot of the banks are looking to leverage video to help do things like prevent money laundering or prevent fraud. You can obviously prevent fraud on the digital world, but there's still this kind of vector of the branch locations, like money laundering happens with the guy walking in and doing something at a teller or at an ATM machine. That's the video source. So that's one use case that we've actually heard and we're actually working on together, but there are other applications of this. So I've talked with some retailers, everybody has video-

Rebecca Knight

>> Shoplifting issues. Yeah.

John Mao

>> It's not just shoplifting. So prevention is one, but they're actually looking at this from a customer analytics perspective. How many customers walk through each of the aisles throughout this time to that time? How long did they stop at the end cap of the aisle to look at the advertising display? Because that data is so valuable to them, not only on product placement, but also back to advertisers to kind of upsell and upcharge those premium locations for product placement. And there's just so many more smart manufacturing, being able to look at quality of things coming off the assembly line. Inside it, it could be also around personnel safety. There's actually a wide number of use cases, most of which is already being recorded by CCTV systems, video surveillance systems, which no one's really watching other than security guard. But now, you can use technology to extract more insight and value out of a massively untapped asset for most organizations.

Dave Vellante

>> I could see some government surveillance.

Rebecca Knight

>> Yeah, me too, I could obviously say.

Dave Vellante

>> The wheelhouse there.

John Mao

>> I can't comment on that. No.

Dave Vellante

>> All this Anthropic stuff that's going on, they just relax some of their conditions I heard, but okay. And now, I got to ask you about Cosmos. When I first heard you guys introduce Cosmos, I was like, "Okay, these guys are getting into our business, content creation." But it's become much more than that, right? To explain what Cosmos is, what it started out as, how it's evolved and where you are today.

John Mao

>> So Cosmos actually, we launched last year, I think October. So two Octobers ago and its original incarnation was really focused around enabling a developer community. Because as we're moving up the stack, we need builders, we need people that will build the next generation of AI applications on top of our platform, and so that's where we started. But the vision for Cosmos even back then was to use that as a way to build the broader community and ecosystem for VAST. So that becomes kind of our umbrella, if you will, for everything. So the announcements we actually made this week are about extending that to kind of realize our full vision. We are now rolling out partner tracks within the Cosmos. It's not just about developers, but it's also about our channel partners, our distribution partners, our OEM partners, our cloud partners, our services partners, our technology alliance partners, all under that same umbrella. Because at the end of the day, like I said earlier, even in the security context, we are not naïve to think that we can do everything. And Renen was very clear, we want to facilitate a broader conversation across a wider community because that's the only way that a lot of our vision will come to fruition.

Dave Vellante

>> And when you go into Cosmos, whatever that community is, I forget what you call it, but it's kind of cool. It's a good interface. There's a lot of content coming in, a lot of Q&A going on.

John Mao

>> There'll be more. Yeah.

Dave Vellante

>> It's a good forum.

John Mao

>> Yeah. We're actually doubling down also on the developer community. So this probably didn't make the press release headlines, but we've been publishing a lot of open source blueprints to help kind of jumpstart the developer community. So it's about kind of bringing everybody together. I mean, this event is also a great kind of example of that. You have customers here, partners here, channel partners here, service partners here, cloud partner, everybody is here. And we actually encouraged them to talk to each other.

Dave Vellante

>> To me anyway, John, and maybe I just missed it when you announced Cosmos, that wasn't clear to me originally, but it's very much now sort of geared toward the developer, because you've got a voice that can help them as opposed to that's our job to be the sort of independent journalist. So that I think is a great fit for you guys. And when I look-

John Mao

>> We're not coming for you. AI might be, but we're not coming for you.

Dave Vellante

>> Honestly, we're not worried.

John Mao

>> It's all good.

Dave Vellante

>> You talk about VAST, you know inside and out, but there you go.

Rebecca Knight

>> I'm curious about how you're going to measure its success if it is about deepening collaboration and accelerating adoption. What are your hopes for this partnership?

John Mao

>> So my team's responsible mostly for the technology partner track, and I have a peer that's responsible for the others. On the developer side, it's also very interesting just to look at the engagement. So we have an entire DevRel team now in our company focused on engaging with developers. On the technology partner track, it's a very simple mission I give my team. We want to work with everybody that shares the same vision. So I got asked a question yesterday, you guys are ... VAST is so closely tied with NVIDIA, for example. We love NVIDIA. We partner with them heavily, but at the end of the day, there will be other hardware platforms we will also have to support over time. Similarly, from an ecosystem perspective, on the software side, we launched with about 60 partners this week. I want to see that number become 600 and then eventually 6,000, right? Because that shows a strengthening community and an ecosystem. We all focused on driving value and solutions to the customer, and so that's my hope, my aspiration. We're going to work hard to get there.

Dave Vellante

>> Well, congrats on getting here and having your first customer event. It's good to see you.

John Mao

>> Glad you were able to make it from Boston.

Dave Vellante

>> No problem.

Rebecca Knight

>> Man, this is Blizzard. Anyway.

John Mao

>> Yeah.

Rebecca Knight

>> John. A pleasure having you on the show.

John Mao

>> Thanks for having me.

Rebecca Knight

>> I'm Rebecca Knight. For Dave Vellante, stay tuned for more of theCUBE's live coverage of Vast forward 2026. You're watching theCUBE, the leader in enterprise tech news and analysis.