Chainguard Assemble 2026 in New York City features a conversation with Dustin Kirkland of Chainguard, senior vice president of engineering. Kirkland discusses product innovation and security strategy and focuses on software factory models, artificial intelligence hereafter AI and secure software supply chains. They draw on experience securing cloud-native deployments and explain Chainguard's hardened images, Chainguard Libraries, operating system packages hereafter OS packages and the evolving Chainguard Factory.
Rebecca Knight of theCUBE Research and Paul Nashawaty of theCUBE Research host the conversation and explore AI-driven development, developer tooling and supply chain security. Key takeaways highlight the industry shift from experimentation to implementation in 2026 and emphasize automation, governance and developer-friendly security. Kirkland explains how Chainguard's expanded catalog, OS packages and agentic capabilities help teams adopt AI safely while maintaining developer productivity. Hosts and analysts also emphasize hardened artifacts, enhanced software bill of materials hereafter SBOM and self-service models, and the need to meet emerging regulations such as the European Union Cyber Resilience Act hereafter EU CRA.
This discussion provides practical guidance for product leaders, security professionals and engineering teams seeking to secure cloud-native development, implement software factory approaches and adopt AI responsibly across the software supply chain.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Chainguard Assemble 2026. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for Chainguard Assemble 2026
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for Chainguard Assemble 2026.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Chainguard Assemble 2026. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to Chainguard Assemble 2026
Please sign in with LinkedIn to continue to Chainguard Assemble 2026. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Dustin Kirkland, Chainguard
In this session Paul Nashawaty of theCUBE Research joins Rebecca Knight of SiliconANGLE Media to unpack the main messages from the Chainguard Assemble 2026 keynote. The conversation addresses Chainguard's customer growth, developer-first security, automated software factories, container hardening, Software Bill of Materials improvements and perspectives shared on the main stage by Chainguard and artificial intelligence industry leaders.
Key takeaways include: Nashawaty asserts that secure software is built at the build stage rather than patched later; they position automated software factories as a competitive advantage; they identify open source as the primary software supply chain risk. Chainguard's product updates include Chainguard factories and Chainguard OS, along with enhanced catalogs and hardened charts, which increase automation and support a self-service developer model. Nashawaty also emphasizes that rising AI-generated code requires continued human accountability and trust and that organizations must integrate security early and scale developer-first practices to manage risk.
Practice Lead and Principal AnalysttheCUBE Research
HOST
Rebecca Knight
HostSiliconANGLE Media
HOST
In this interview from Chainguard Assemble in New York City, Dustin Kirkland, senior vice president of engineering at Chainguard, joins theCUBE's Rebecca Knight and theCUBE Research's Paul Nashawaty to discuss how the company is industrializing software security for the agentic AI era. Kirkland unpacks today's wave of product announcements, including Chainguard OS packages, millions of hardened library artifacts spanning Python, Java and JavaScript, and a new self-service Catalog Starter tier designed to meet developers where they are. With the customer base ...Read more
exploreKeep Exploring
How is adoption of your software-factory capabilities progressing, and how many of your customers are mature enough to take advantage of them?add
How do Chainguard OS packages work, who gets access to them, and how do they relate to the Chainguard Catalog and Chainguard's hardened container images?add
What does the Chainguard Catalog Starter announcement mean, and does it introduce or emphasize a self-service model for developers?add
How can customers use Chainguard’s tools and automation (for example the Driftless Agentic Framework and the GitHub app) to detect, request, and replace third‑party software with Chainguard images, and how are changes proposed and validated?add
>> Hello everyone, and welcome to theCUBE's coverage of Chainguard Assemble here in New York City, the Big Apple. I'm your host, Rebecca Knight, sitting alongside Paul Nashawaty, Principal Analyst at SiliconANGLE Media. We are joined by Dustin Kirkland, SVP of Engineering here at Chainguard. Welcome. Welcome, Dustin.
Dustin Kirkland
>> Yeah, thanks, Rebecca. Thanks, Paul. Appreciate it. Glad to be here.
Rebecca Knight
>> Yeah, no-
Dustin Kirkland
>> Thanks for having me....
Rebecca Knight
>> it's a beautiful day. And we're talking tech. We're talking software factory.
Dustin Kirkland
>> Yes.
Rebecca Knight
>> Which is the way that Chainguard describes itself. And it's a really interesting analogy. Obviously, we had Dan Lawrence up on the main stage, brought up a real power tool, but this manufacturing analogy that the idea that software production is going through these same changes that physical manufacturing did too.
Dustin Kirkland
>> Sure.
Rebecca Knight
>> How does that framing click for the executive leadership team at Chainguard?
Dustin Kirkland
>> Such a powerful visual, right?
Rebecca Knight
>> Yeah. Yeah.
Dustin Kirkland
>> Dan had on stage a piece of wood, and he asked a volunteer to come up and use an old school hand saw to cut it. It took a few minutes and we all had a few laughs. And then, big reveal, pulls out the power tool and just, zip, right across. And the point that he was making, which I think is very apropos to software developers, is we are entering the era of power tools for software development. And here we're talking, of course, AI, agentic development, and we got a lot to talk about with respect to that.
Paul Nashawaty
>> No, it's definitely ... It's exciting times. The market is shifting, there's things happening. I like to say that 2025 was the year of experimentation, '26 is there year of implementation.
Dustin Kirkland
>> Yeah.
Paul Nashawaty
>> I agree with you. I agree what Dan was talking about in the keynote. I agree with the software factory approach. Automation is key. Definitely an area that I think a lot of companies aspire to get to. They want to get there. But I would ask from the perspective of maturity, because there's a lot of ... Obviously, there's a competitive advantage. You have to put this in the automation, you have to make sure things are working right. We see AI creating a lot of production code. But from a software factory perspective, how is the adoption going? We see your customer base growing substantially. That's great. Year over year, went from 150 customers to just under 500.
Dustin Kirkland
>> Almost 500 now.
Paul Nashawaty
>> Yeah. That's amazing. So congrats on that. But I think that from a maturity perspective on the software factory piece, they want to get there. How many of your customers are fully mature to take advantage of that?
Dustin Kirkland
>> Yeah, we see the full range in the spectrum, but with the product announcements that we made today, we're really helping many of those customers dramatically accelerate those timelines. Chainguard's product started with our hardened container images, which really helped a lot of those customers professionalize their usage of Google Kubernetes engine and Amazon's Elastic Kubernetes Service, AKS, EKS, GKE, or their own Kubernetes, by publishing secure hardened images that allowed those customers to meet their compliance requirements, address regulatory issues, and just eliminate software vulnerabilities. Last year we took it from there to a Chainguard Libraries announcement. And here today, we announced millions of Chainguard-built library artifacts across the Python, Java and JavaScript ecosystem, which helps those developers who aren't operating at the containers and Kubernetes level, but instead they're writing software and PIP installing Python packages or Maven adding Java packages or NPM installing node packages, but to do so safely from a secure repository. Now, that's just the starting point, Paul. From there, we get into, well, now I want to adopt these AI tools and technologies. How do I do that safely? How do I help my developers do that safely? If I'm a security engineer, how do I ensure that we're not giving away the keys to the kingdom and inviting malicious actors and threats into the system? And so we're now talking about Chainguard hardened agents. We're talking about extending the Chainguard factory, the same things we use AI to build, making that available to our customers. Yeah.
Paul Nashawaty
>> Yeah. No, that's a lot there to take in. I think that I'm going to use the keynote's words saying you're making security invisible to the developer, which is incredibly powerful and important, especially when you start looking at what does the term developer actually mean. You're looking at citizen developers or professional developers, right?
Dustin Kirkland
>> Sure.
Paul Nashawaty
>> And when we start using these applications, the thing that comes into my mind, and the market has been echoing this a lot, is governance, compliance and regulation. When you think of things like the new EU CRA that's going into effect, by September 2026, reporting has to be in place. And by December 2027, all applications have to comply, otherwise you have steep penalties.
Dustin Kirkland
>> Indeed.
Paul Nashawaty
>> This is an area that I think the hardened images really comes across as a security factor that know I can trust something that you can deliver. Would you agree with that or is that the area you're going after?
Dustin Kirkland
>> Yeah, I think it starts with your opening volley, which is, man, the definition of a developer has dramatically expanded. Once upon a time, a developer was a person who could sit down and write code. Code of some kind. And you can go all the way up from machine code and assembly code to low-level C and C++ to Java, to high-level interpreted languages. But still, the classical developer was someone with or without a computer science degree who would open up an editor of some kind, type some code out, and then try to get it to run. And great developers were able to do that quickly. They were able to do it efficiently. They were able to do it in ways that prevented problems later. But now we're seeing where developer means someone who can describe the state of a system that they want to exist in a prompt to an AI, and the AI does the code writing aspects of it. But that developer, I would still say that an important part of the developer definition is they have to be able to judge whether or not that's good and that's right.
Paul Nashawaty
>> Absolutely.
Dustin Kirkland
>> And so, those are the pieces that many of the announcements we made today is how Chainguard is helping that broadened definition of a developer succeed in a space where the tools are so much more powerful than they've ever been before.
Rebecca Knight
>> And to that point, you're also really changing the skillset of what a good developer does and needs to do in their day-to-day.
Dustin Kirkland
>> Yeah, insightful. Very insightful. What a developer actually has to do, describe the state of a system that they want to exist, churn, churn, churn, evaluate the results and know whether those results are good or bad, and then go back into the prompt. A huge piece of this has been the reconciliation. That's what we call that loop that happens inside of that churn, churn, churn piece, that continuously looks at the output of the system and puts it back to the AI and the agent to redevelop it and fix it again. There's still humans in the loop, and I think there's going to be humans in almost every loop for a very long time, but the amount of work that's able to be done inside of that agentic process, it's incredible, Rebecca.
Rebecca Knight
>> Yeah. So let's talk a little bit about some of the new product services and features that have been announced today. Want to start with Chainguard OS packages.
Dustin Kirkland
>> Yes.
Rebecca Knight
>> You already mentioned them a little bit, but talk a little bit-
Dustin Kirkland
>> Yeah, I love this one. It's foundational. It's atomic. It's in some ways more empowering to our customers who really want to build using their own tools, but start from the most secure base possible. We've opened that up. So any customer of ours who is already a Chainguard Catalog customer, meaning they had purchased access to the entire image catalog, automatically has access to Chainguard OS packages. Customers who want access to Chainguard OS packages simply become a Catalog customer. So they go hand-in-hand. Now, our Catalog customers get access to over 2,000, and growing, hardened container images that we've built, we've procured, we continue to test and quality assure and remediate the CVEs in. But we've had customers who've wanted to build their new and interesting images in new and interesting combinations of our packages. And so now we've made that available. Customers can take Chainguard APK packages, still covered by our SLAs to remediate CVs and security vulnerabilities in those packages, and then they can assemble them using whatever tool they want, not necessarily getting the hardened container image from Chainguard. Or I guess, in addition to getting the hardened container images from Chainguard, they can assemble their own images to however they like using our packages. So that's our Chainguard OS packages. It's the most fundamental atomic unit, high-quality operating system packages provided, fresh, free of CVEs and quality assured.
Paul Nashawaty
>> And it's well-adopted. The numbers you were showing at the keynote, the customer adoption was pretty substantial from year-over-year.
Dustin Kirkland
>> Yeah, absolutely.
Paul Nashawaty
>> So I think the Catalog growth is there as well. You mentioned, speaking of the Catalog growth, the innovation provided by the hardened charts and also enhanced SBOMs. Those were new features. But one of the things that was interesting to me was the self-service model.
Dustin Kirkland
>> Oh, yeah.
Paul Nashawaty
>> Again, I kind of lean back to the previous comment of the developer, they're doing their thing, everything's being shifted left, or shifted anywhere at this point. What does that mean, Chainguard Catalog starter as a new announcement and that also has a focus with the self-service model? Is that a-
Dustin Kirkland
>> Yeah, so look, I'll put on my developer hat. I usually know what I want and I know where to get it from, and we've essentially adapted the Chainguard enterprise sales model to meet many developers where they are. They just want access. So there's some number of free images that any developer can get access to now at this point. What we've found is that often, those individual developers bring Chainguard into an organization. They already fall in love with it and then it spreads, and spreads quite virally, in a good way that solves security issues for other teams. So we've expanded that tremendously, I think, through the developer program. Self-service just means developers can go directly to Chainguard, get the images they want, start building immediately without having a trial period and a sales executive involved.
Rebecca Knight
>> So this philosophy that you're describing where developers own the recipes but Chainguard owns the ingredients, is this clean mental model. But what do you do when those worlds collide, when a developer wants to use a package that doesn't meet your security standards, for example? What's the ethos there?
Dustin Kirkland
>> Yeah. Certainly, we have processes by which we add packages to the catalog. We add dozens of packages to the catalog every single day, a couple of images every day or every other day. That's roughly the pace. We have an intake process by which any customer can request new unpackaged software to become packaged, new images to become created by us. And we run it through a series of quality gates to ensure that the software is well maintained, that it's free of malware or can be made free of malware and CVEs. And then we put it into basically the automation where our factory will continuously rebuild and rebuild that software, to detect new releases and build that. Occasionally we encounter something that doesn't meet our standards. We'll usually inform the customer why. It's not a thoughtless decision. It's very thoughtful. "Here's reasons why we're not going to take that software on." And it's rare, but when we encounter that, we're usually a trusted advisor to our customer that, "Hey, this isn't good software that we're going to broadly depend on." We'll often also advise them on alternatives to that. Now, if they want to continue building that for other reasons, they certainly can, and they can add that to their images using the mechanisms that I just described. But we'll take a position occasionally on something that we don't believe belongs in our enterprise catalog. It's very exceptional though.
Rebecca Knight
>> Okay. Okay. And it also would, I would imagine builds credibility and trust too, with your customers-
Dustin Kirkland
>> Sure. Yeah....
Rebecca Knight
>> as that advisory role.
Dustin Kirkland
>> Especially when we can offer, "Here's a well-maintained alternative. Here's a fork of that dead project that someone else is maintaining." Occasionally we take on the maintenance of that. We have this process called Emeritas to build on Emeritus and then open source software at the end, where Chainguard is taking on the maintenance of some extremely well-used, high-quality software that was simply abandoned by the maintainers.
Paul Nashawaty
>> Yeah. So let's double-click on that. I like the trust point that you were talking about, Rebecca, but I also want to get into open source a little bit more. Obviously, the ecosystem of using open source technologies and the technology stacks is important to a lot of developers. It's actually critical to most developers in how they develop their applications. The Chainguard approach of making sure that that works within their environment, you put some, I'll call it belt and suspenders in place to make sure that things actually work the way they're designed. Can you talk a little bit about that?
Dustin Kirkland
>> Yeah, the belt and suspenders is definitely, we go heavy on the quality assurance. I think that's very important. Customers come to us saying, "Hey, we love this idea of a rolling release or a distroless approach, but how do we know that it's going to work?" And so I'll usually explain to them the entire process by which we detect upstream releases, we download that source code, we check the check sums, ensure that those match. We apply the build rules. We build, configure, make, make install, or whatever the build process is. We'll run any upstream tests that are available for that project. We've also enhanced every single package, every one of our 7,000-plus packages. 100% of them have Chainguard-developed proprietary tests that we run on top of that, and ensure that every subsequent release can meet those quality gates. If anything fails, then it goes into a queue where humans then review. And then, we then calculate build dependencies and run time dependencies of those. We'll retest there. Then we assimilate those into images, run integration tests across those images, and then we'll put them into helm charts and federate multiple services together and run yet another set of tests. So that's your belt and suspenders, Paul. It's multiple levels of quality assurance gates. If anything fails, and I mean anything fails, any one of the dozens and dozens of checks, it ends up in a queue where a human analyzes what went wrong, figures out how to solve that, and then retrains our models so that we never encounter that problem again.
Paul Nashawaty
>> Built for scalability.
Dustin Kirkland
>> Totally. With a flywheel where this factory continues to get better every time it runs.
Rebecca Knight
>> You are architecting for a future that is still taking shape. How do you think about that when you're engineering systems today where AI is autonomously writing code and generating and maintaining the code? How do you think about that and how do you work with customers to help them approach?
Dustin Kirkland
>> You're more right than you realize, Rebecca. About this time last year, me and my engineering team celebrated the release of our Chainguard Factory. And we didn't call it 1.0 at the time, we just called it the Factory. We're done. We had this entire initiative called Finish the Factory, and we finished it. And then guess what? We had to rebuild it all over again. And now I think we realize ... Dan today was talking about the Chainguard Factory 2.0. We created the 2.0. We had to go back and call the previous one the 1.0. And it turns out 2.0 isn't the finishing point either. This thing is going to continue evolving. So yes, we're absolutely shaping the future. I would say in, I don't know, the humblest way possible, that we certainly were proud of what we've done and we feel like we're ahead of the curve, but we're also ready, willing, and able to share all of those advances with our customers as much as possible. And to that end, we've extended much of the Chainguard Factory itself to our customers. They can leverage much of the AI and agentic approaches, our Driftless Agentic Framework, the reconcilers that we use to just make that software better over time. Our customers can take advantage of that directly. They can leverage an app that we built for GitHub. They can install that GitHub app into their repository, and then that GitHub app can detect all the places where those customers and developers are using software, software that could come from Chainguard that does it at this point. So it'll look at Docker files and determine what is the source of those images, and say, "Hey, there's actually a Chainguard image here." It could also say, "There isn't a Chainguard image, but there could be. Would you like to open up a request for an image here?"
It'll detect places where developers are installing third-party software off the internet. This is the PIP install, the Maven add, and the NPM install that's just coming from somewhere off the internet. It will determine that, "Hey, these artifacts could come from Chainguard." And by the way, we're not just going to complain at you and tell you all of these things. We're actually going to send you a PR, a pull request, that says, "Here are the changes that you would hypothetically need to make your Docker file, so your build processes to your code to take advantage of the Chainguard approach." And then it will write tests even, to determine whether the previous version of this matches the successive version once it's rebuilt with Chainguard, and then ensure that those match. It's still ultimately up to those customers, their humans, their developers, their architects to review the results and approve that pull requests or make modifications to that pull request. So that's the future that, it is under development, Rebecca, but man, it is here as well.
Rebecca Knight
>> It is. The future is bright. Well, Dustin Kirkland, thank you so much for coming on the show. Great conversation.
Dustin Kirkland
>> Thank you so much. Yes, absolutely.
Rebecca Knight
>> I'm Rebecca Knight for Paul Nashawaty. Stay tuned for more of theCUBE's coverage of Chainguard Assemble. You're watching theCUBE, the leader in enterprise tech news and analysis.
