KubeCon + CloudNativeCon NA 2025 | Kelsey Hightower & Bobby Allen, Google & Eric Hanselman, S&P Global Market Intelligence

Clips
More from KubeCon + CloudNativeCon NA 2025

Eric Hanselman

Chief Analyst

S&P Global Market Intelligence

Kelsey Hightower

Distinguished Engineer

Kubernetes Thought Pioneer

Bobby Allen

Cloud Therapist

Google Cloud

play_circle_outline KubeCon event celebrating 10 years of Google Kubernetes Engine (GKE) and its impact.

play_circle_outline From Inception to Impact: Kelsey's Role in GKE's Journey and Its Influence on the Cloud Landscape

play_circle_outline Bobby discusses GKE’s versatile capabilities supporting various workloads, including AI.

play_circle_outline The Future of Kubernetes: Seamless Integration and Its Evolution as the Invisible Backbone of Technology

Info
Transcript

Kelsey Hightower & Bobby Allen, Google & Eric Hanselman, S&P Global Market Intelligence

Eric Hanselman

Chief Analyst S&P Global Market Intelligence

Kelsey Hightower

Distinguished Engineer Kubernetes Thought Pioneer

Bobby Allen

Cloud Therapist Google Cloud

In this KubeCon + CloudNativeCon North America conversation from Atlanta, theCUBE’s Savannah Peterson sits down with Bobby Allen, Kelsey Hightower and Eric Hanselman to celebrate 10 years of Google Kubernetes Engine (GKE) and its evolution from open-source project to production platform. They reflect on how GKE helped normalize cloud around a common Kubernetes API, gave Google Cloud new enterprise relevance and became a foundation for modern AI workloads by abstracting away much of the infrastructure complexity.

The discussion also spotlights GKE’s n... Read more

explore Keep Exploring

What is the significance of the 10th anniversary of the Google Kubernetes Engine, and how does it feel for Kelsey Hightower to be part of the celebration? add

What impact did GKE have on Google Cloud Platform's position in the enterprise market? add

What are the capabilities of Kubernetes in supporting different types of workloads, particularly in relation to AI? add

What is the future outlook for Kubernetes and how are tools like GKE Autopilot facilitating its use? add

bolt Powered by CUBE AI

Kelsey Hightower & Bobby Allen, Google & Eric Hanselman, S&P Global Market Intelligence

search

Savannah Peterson

>> Good afternoon, nerd fam and welcome back to surprisingly arctic Atlanta, Georgia. We're here on day one of three days of coverage at KubeCon, but we're starting off a very exclusive and special series with our friends here at Google Cloud. My name's Savannah Peterson, bringing you all the fun with an absolutely impressive lineup next to me, to my left. We've got Bobby, Kelsey, and Eric. Thank you gentlemen so much for taking the time.

Bobby Allen

>> Thank you.

Savannah Peterson

>> So this is a really cool segment. We are here low-key to party, to celebrate the 10 years of the Google Kubernetes Engine. Very big day to day. First of all, congratulations particularly to you two. This is a big deal. Kelsey, it would be awkward if I didn't open this up with you. You've been a part of this entire movement, I'd like to call it, beyond the project and the tooling since inception. What is it like for you to be in the room today, look around, get completely bombarded by all of the wonderful open-source fans of yours?

Kelsey Hightower

>> Well, number one, we took a project to a product and made it 10 years.

Savannah Peterson

>> That's like 100 years in normal lifetime.

Kelsey Hightower

>> 100%. And someone else there when they were struggling to figure out what to name this thing and we just put the K in the middle and kept it there as a placeholder and it stayed there all this time. And the fact that people are still using it for the future of workloads is tremendous.

Savannah Peterson

>> I love that. Wait, so just take us to the naming room. What was that conversation like? Was it heated? Were you just trying to get something done?

Kelsey Hightower

>> No, the name Kubernetes is a project name.

Savannah Peterson

>> Of course.

Kelsey Hightower

>> So you have things like BigQuery, you have all these other things. And so you have this open-source project, which is not a product, so what do you name the commercial service? You got to hurry up, and so GKE, Google's Kubernetes Engine, and we'll get a better name later. But it stuck. I think it's super clear that it's one of those things that we do all the hard work to make this thing usable in production. And that was the vision 10 years ago, and it's still true today.

Savannah Peterson

>> It is. And so exciting. For those who might not realize, not every project becomes a product or a platform or a full-on technology solution. It's a big deal for all the starting projects here in the wings. I hope they're looking up to this and taking it. Eric, I want to ask you, you've been following the journey, you're very familiar with this space. Did you anticipate that GKE was going to be that just power player that it is now?

Eric Hanselman

>> If you look at an effort to try and revolutionize what are the foundational elements of infrastructure? That's a big ask. You think about being able to move from what was a perfectly functional virtualized environment we've been living with for ages and to look at what's necessary to actually make that shift happen, to be able to go move... Look at all the attempts that people have made to do this over the past in all these different disciplines, being able to make that work. Now, granted, you had the advantage of being able to take the work production environments, to be able to prove it out and harden it, but hey, 10 years, it stood the test of time.

Bobby Allen

>> It has.

Savannah Peterson

>> Well, and now, as we were just talking about before we went live, it's not just momentum right now. We're right now. With AI, we are totally getting catalyzed. What's it like for you on the inside hitting this incredible milestone?

Bobby Allen

>> It's amazing. It's humbling to have a product, really a platform that's at the center of so many different things historically and going forward. And to see it accelerate, it's just amazing to be a part of that, to play a really small part in it. So here's my food analogy for this.

Savannah Peterson

>> You know I was waiting for it.

Bobby Allen

>> So my daughter is like the air fryer queen, right? Or air fryer

Savannah Peterson

>> Oh, I'm going to have to ask her for some tips.

Bobby Allen

>> And so she exclusively makes chicken nuggets in the air fryer, they're crispy and all that kind of stuff. I'm a tater tot guy myself, by the way, for your listeners. Tater tots are the superior potato compliment, especially the little crispy ones. But what changed my life, Savannah, was being at grandma's house one day, at my mother's house, she put a croissant in the air fryer. Oh my gosh. It was warm, it was crispy, it was just right.

Savannah Peterson

>> Were you in Paris all of a sudden that moment?

Bobby Allen

>> That's what it felt like.

Savannah Peterson

>> I love this for us.

Bobby Allen

>> But I feel like AI is like the croissant in the air fryer. Now that we've had a taste of that, it's like, ooh, this is so good. And so that's allowing us to take it to the next level because we took something and combined it with something else, and now that combination is magical.

Savannah Peterson

>> Well, now I'm hungry. And also, brilliant analogy. I love that because it was something you didn't know you could have as good essentially is what you're saying. When we think about so many different things that are powered by GKE, I'm curious... Actually, I'd love to get all three of your opinions on this. Kelsey, what's your favorite solution you've seen come to life as the result of the work that you and the team have done?

Kelsey Hightower

>> Honestly, when I joined GCP 2016-ish, it was hard for Google Cloud to get a seat at the table, the enterprise table. GKE comes out and it's the talk of the town. So much so that it was the first time that I saw GCP inspire and influence all the other cloud providers. And so when we think about a solution, what GKE does, maybe people that haven't looked under the hood, it takes what was traditionally infrastructure as a service and makes it one thing. So before then, you're doing DevOps, platform engineering, you're writing a bunch of scripts and tooling. And I often tell people, "Kubernetes is what your system administrators would've built if they had enough time and they knew how to do it." And so when GKE comes out, it was the solution to GCP's place in the marketplace and it advanced the art of cloud to the next level. So another big part about GKE that's important is that we almost normalized the cloud around this one API abstraction that no one thought was possible, because before that, everything was proprietary. We finally got the cloud providers to agree on something. So GKE was a solution to get the cloud providers to agree on something. And we did it.

Savannah Peterson

>> So not only did it rocket your relevancy, but it also helped create a bit of a standardization across the board so that we could all innovate at the speed we want to without wasting a bunch of money and reinventing the wheel.

Kelsey Hightower

>> Google went all in and it paid off.

Savannah Peterson

>> I love to hear it. Well, you can certainly feel it and see it in this room. This is my ninth KubeCon. I've been a part of the community for a long time and it's really powerful. You can tell that right now, to your point, we are having, I'm going to start saying this all the time, our magic croissant moment, which I am super, super into. Eric, I'm curious, with your non-biased outside perspective, what's impressed you the most or excited you the most as an analyst?

Eric Hanselman

>> I was trying to come up with the croissant comparison here, but I don't know if I can go so quite far as Bobby on this.

Savannah Peterson

>> I have faith in you, Eric.

Eric Hanselman

>> I may have to work on that one. The realization of run anywhere, how long had we been talking about the ability to be able to take capabilities and have some reasonable level of assurance that they can get to where they need to go? Especially now with AI. One of the things that we see in a lot of our end-user study data is that as we get into AI, there's this hangover of cloud dependence that organizations look forward and they want to be able to get to open containers, give them the opportunity to have an interchange, to be able to have that run anywhere to deliver on that run anywhere promise. And to do it in ways that give them operational efficiency, give them all of the integration and capabilities to be able to build in the automation orchestration to really make this stuff work the way they have hoped that infrastructure has run for a long time.

Savannah Peterson

>> So it's like the full bakery? Is that what you're trying to tell me?

Eric Hanselman

>> It's the full display case of everything.

Savannah Peterson

>> Which is really exciting. So Bobby, I know there's some announcements today. 10 years later, where is GKE right now?

Bobby Allen

>> GKE is very broad and very deep in terms of what it can support, and I think what's exciting is I think folks have been surprised. AI was another class of workload, but the number of different things that Kubernetes can support, I think it's been amazing and I want to give the audience a little bit of Greek to expand on this.

Savannah Peterson

>> Please do.

Bobby Allen

>> So English is very word-poor as we know. We use a lot of adjectives, but other languages like Greek will have different words for different things. Greek has two different words for new, there's nèos and there's kenos. Think of nèos like a new pair of tennis shoes, like a new instantiation of existing recipe.

Savannah Peterson

>> Iteration.

Bobby Allen

>> Think of kenos like a jet pack. And so what I think is powerful is folks are seeing that Kubernetes can support the new stuff that's like a new iteration of something like tennis shoes, but we're also building jet packs now. So the fact that it's got broad enough shoulders to run things better, but also to run things that are different is what's exciting.

Savannah Peterson

>> That is really exciting. What are some of the different things people are going to be able to run now? I know there's a lot of announcements around inference, there's a lot of announcements around agentic.

Bobby Allen

>> Lots of announcements. And again, we want to come get the audience for those who may not be quite as technical as these gentlemen are up here. And so when I think about this, because there's going to be some other segments to go really, really deep into announcements, if I was going to categorize this, I got to go to Pixar, I want to go to the Incredibles. Because I think they're different types of features that relate to the different characters. There's some things that are feats of strength like Mr. Incredible. Being able to run things like 130,000 nodes in a single cluster. Sometimes you just need more horsepower. Other times, like Violet, the one who could be invisible and have the force field, we've got things like our agentic Sandbox in GKE because you need to sometimes protect your old workloads from the new stuff as the new agent. So operating and getting leveled up with tooling and all that, you need to have a little bit of separation. Sometimes you just need speed. You need things like pod snapshots that will let you resume things faster. Other times you want to just be cool and do things like autopilot to reduce the friction, you got to have a shout-out to Frozone and anytime you talk about anything cool, you got to bring him in. And then some things are just faster like Dash where you want to just do things faster than you did before. So there's a whole family of features that come together just like that family and they together can do things to defeat certain villains in the enterprise that we're facing as we try to take that new ground.

Savannah Peterson

>> I think you might be the person on the planet who makes this the most digestible of anyone. That was delightful. We've got our own little incredible team. Eric, I see you buzzing to say something right there.

Eric Hanselman

>> Oh, just all the points that you're making about, there's different constituencies that are trying to get towards AI capabilities. They're looking at some of the challenges in agentic and you've got those two ends of the spectrum. You've got those folks who are looking to get the templated guidance to be able to get in. And so if you look at a lot of those offerings, the ability to now actually have something that's an opinionated way to move forward. And then the other end of that spectrum are those applications where you got to scale, and you've got to be able to move the scale to be able to really deliver on the promise of agentic.

Savannah Peterson

>> And not just scale up but also scale out, which I think is one of the big linchpins of what's going on right now. We're talking about sovereignty, we're talking about compliance, regulation. There's so many different things. You got to have the right trusted ally who can actually interface and has thought about that before you even got there so you can keep running your workloads. Kelsey, I'm curious, if someone had asked you 10 years ago, where would GKE be in that moment, would you have predicted that we would be where we are today?

Kelsey Hightower

>> I did multiple times.

Savannah Peterson

>> Perfect.

Kelsey Hightower

>> I had analogies that Kubernetes was the Linux of the cloud. This is a 20-year technology we're working on. And when you say a 20-year technology, that means a ecosystem is required. This wasn't something that was going to rely on the fad. And I used to say things like, we're building a platform for platforms. Kubernetes is where you start but not end. And also everything, this is rare product that the commercial enterprise product makes the open source project better.

Savannah Peterson

>> Yes.

Kelsey Hightower

>> We always did things that way. So in the cloud, we had one unique advantage. You have thousands and thousands of customers that all want different things, whether it's NVIDIA for the scheduling stack around GPUs or someone that's just trying to run a web server better than they were doing before. All of those workloads are mixed and they run on the same platform and then it feeds back into the underlying ecosystem. That's very rare that you have a huge ecosystem underneath it. This is why I can't adapt because whenever a new ecosystem spawns up, they look around and say, "Hey, we might as well start there," versus recreating something. So I'm not as surprised. I am very grateful that the project is around, but that was somewhat by design.

Savannah Peterson

>> Yeah. And obviously a part of the vision from the gateway. That relationship between open source and enterprise I think is a really interesting one and I feel like you really laid the foundation for how to do that best. Eric, would you agree?

Eric Hanselman

>> Absolutely. If you think about what's necessary to really be able to give flexibility, to be able to give scale, to be able to create an environment that's going to be able to last through so many different iterations and capabilities-

Savannah Peterson

>> And hype cycles at this point....

Eric Hanselman

>> chasms and all the rest, take it back. But it's being able to put something together that's thought enough about what's really necessary to be able to evolve over time, to be able to adapt. Because even now as we enter all the craziness that's happening with the velocity acceleration of AI, to even be able to survive that in terms of where it gets to and what it's able to support.

Kelsey Hightower

>> I think there's another key point that's super critical. Google chose to compete on openness. There was very little fear that because a competitor was going to run Kubernetes that we would somehow be at a disadvantage. At that time, remember, Kubernetes sits on top of a cloud and we often say Kubernetes is no better than the cloud provider it runs on. And we had no fear of everyone having access to the similar bits because we always thought of GKE marrying the two. And so I think that decision is what allows us 10 years later to have such a broad ecosystem because you don't get the ecosystem without being open.

Savannah Peterson

>> I think that's a really good point. I'm so glad you brought that up. I know you were about to chime in, Bob.

Bobby Allen

>> I want to jump in just for a bit of transparency.

Savannah Peterson

>> Please.

Bobby Allen

>> I was not a believer. I was not. There are people that can find me on video where I was hating on Kubernetes. I was, because I saw how many...

Savannah Peterson

>> Spill that tea, Bobby.

Bobby Allen

>> As I was doing a lot of startup stuff and did a lot of enterprise stuff, worked at banks and all that, I saw them struggling with the tech and the complexity. And I was probably about five and a half, six years ago where I turned the corner because you couldn't argue with the data, you couldn't argue with the size of the community. You couldn't argue with the number of Google products that are powered by containers or Kubernetes directly or indirectly. And you couldn't argue with the fact that, I'm going to quote Gary Singh, one of the other guys, "We weren't selling it. People were buying it." And at some point you have to look at what the community is saying and say, "Look, Bobby, you got to get on this train." And so I have to admit, I was not there for 10 years like these gentlemen were. I'm a late later convert, so to speak.

Savannah Peterson

>> That's still a while ago. Just to be clear, five years ago, Kubernetes land is pretty solid

Bobby Allen

>> It is. And probably , but I just want to be honest with the community, I wasn't there. I was one of the skeptics. I remember debating with people like Damani Corbyn and Keith Townsend, "I'm not feeling it. I don't think it's there." And so now I've turned the corner because I've seen what the community's done. I've seen what people have built on it, and I've seen the openness that Google really gave it away to the community because the whole thing was, it's not about what we're building, it's about what we're empowering you to build. That's when I got on board.

Savannah Peterson

>> First of all, I love that you admitted that. We're not all right all the time, especially as analysts. You and I both know this gets a little... Maybe you're right all the time. I will be honest, I am not right all the time. But you brought up something really, it's interesting that you say it was five, six years ago. For me, it was five years ago. I distinctly remember when all of a sudden the light bulb went off and I realized, oh, this is going to be how this works in the future. And it was still lovingly clunky and hard and complex, and there were a lot of tools and features and projects that have come around to rally that up and GKE has obviously matured a lot. But yeah, there was a real shift, I remember, for me when I was like, "Oh, I am going to be all in on this Kubernetes thing and it'll be a part of my career." And I remember my mom being like, "Kuber-what? What do those words mean?"

Kelsey Hightower

>> There's a saying, "It's easy to predict the future when you're working on it." And a lot of us that have built various systems over the years, you saw it when you got in there. And then I think even when it got to cloud, there was still this thing of, are other platforms going to win? Are other projects going to win? Should we hedge our bets here? And we went all in and that meant everyone had to work on it. And then we made it be what it needed to be.

Savannah Peterson

>> I think that's a really good point. Wow, gentlemen, this has been awesome. I have one final question for each of you. No one gets off the hook on this one. Decade anniversary. Huge milestone. First of all, congratulations. That's a big deal. 10 years from now, what are we talking about on this stage? Kelsey, I'm starting with you because you said it was a 20-year technology to begin with.

Kelsey Hightower

>> So 10 years ago we were like, "Kubernetes will reach its peak when it disappears." And so we've been talking about GKE a lot here, but one of the lineages was GKE Autopilot, the part where we hide some of that complexity that Bobby was talking about. When people used to say Kubernetes was hard, they were talking about all that other stuff. And so Autopilot moves us to a world where that stuff is hidden from you. 10 years from now, you'll be using Kubernetes and not even know it. There are a lot of people that are using parts of GCP that have Kubernetes underpinnings and they don't even know it. So I think that just continues for the next decade.

Savannah Peterson

>> I feel like you could have swapped AI in for Kubernetes as well in that. It's all just going to be a part of where we are. It's going to come meet us where we are. I love that fiction. Eric, I'm going you next and then I'll keep Bobby in the hot seat.

Eric Hanselman

>> I'm going to continue the abstraction message because again, we abstract away, we do technology, we abstract away complexity to be able to enable. And to the extent to which we just simply make the provisioning of the computational environment the data we need that much more capable and visible, absolutely. Less visible, but still a foundation. And how much more can we get?

Savannah Peterson

>> Yeah. All right. That was nice of you just to concur there. Nice one. Nice one, Eric. All right, Bobby, what about you? Give us something fun and radical.

Bobby Allen

>> This a tough one. I do think Kubernetes is going to continue to go away, but I'll make a statement and some may disagree with this. I think Kubernetes' impact has mostly been in the digital world so far. I think we're going to be talking about physical impact.

Savannah Peterson

>> Oh, I like that one, Bobby.

Bobby Allen

>> Other mission-critical things, other things that our grandmothers and our cousins can touch that are going to be powered by, inspired by, running on Kubernetes. And I do think we're going to get to a point where we're not going to decide where it runs. The AI often that was trained on Kubernetes will decide the right place to run it, almost to channel the Jamiroquai song, Travelling Without Moving. You'll be running it at different runtime, which may be inspired by Kubernetes like a cloud run for example, but it'll just disappear. But I think it's going to be things we can touch.

Savannah Peterson

>> I like that. The physical world, physical AI, physical Kubernetes. Super fun. Gentlemen, thank you so much. This has been inspiring.

Bobby Allen

>> Thank you.

Savannah Peterson

>> I appreciate it. I hope you're all feeling as inspired as we are here in wonderful Atlanta, Georgia at KubeCon. My name's Savannah Peterson. You're watching theCUBE, the leading source for enterprise tech news.