AWS re:Invent 2025 | Julie Neumann, Honeycomb.io

Clips
More from AWS re:Invent 2025

Julie Neumann

CMO

Honeycomb.io

play_circle_outline Emergence of AI agents impacting the cloud computing landscape and workflows.

play_circle_outline The necessity of getting AI into production environments while ensuring resilience.

play_circle_outline Market consolidation in observability with major acquisitions indicating growing importance.

play_circle_outline Expanding Observability: The Crucial Role of AI and Security in Modern Application Development

play_circle_outline Emphasis on making observability tools accessible for all levels of engineering teams.

Info
Transcript

Julie Neumann, Honeycomb.io

Julie Neumann

CMO Honeycomb.io

In this interview from AWS re:Invent, Julie Neumann, chief marketing officer at Honeycomb.io, joins theCUBE’s John Furrier to discuss the rapid convergence of cloud-native architectures with the emerging AI-native world. Neumann compares the current pace of innovation to the industry's previous shift to the cloud but at "10x speed," emphasizing that while AI provides velocity, deep observability is essential for establishing trust. The conversation explores how organizations are moving beyond experimental "vibe coding" to deploying frontier agents in producti... Read more

explore Keep Exploring

What are the main developments and trends observed in the integration of cloud native and AI native technologies as discussed at AWS re:Invent? add

What does success look like in the AI era, particularly regarding accelerating things into production and the implications of mergers and acquisitions in the industry? add

What does success look like in the AI era as the worlds of observability and AI converge? add

What are the current trends and challenges in the observability space, particularly in relation to AI and larger foundation models? add

What principles does Honeycomb emphasize regarding telemetry data and its accessibility? add

bolt Powered by CUBE AI

Julie Neumann, Honeycomb.io

search

John Furrier

>> Welcome back to one of theCUBE's live coverage at AWS re:Invent, I'm John Furrier, host of theCUBE. We're here for three days. Our 13th re:Invent watching Amazon really run the table on cloud computing, get some competition. Now agents are hot, agents are the new cloud. Posted a story with my exclusive with Matt Garman this morning where we went into great detail around all the things that are going on in the stack around AI. Certainly a lot of it is around observability and also end-to-end workflows as these agents start doing the work and abstracting away those complexities, similar to what the cloud did with infrastructure and a lot of stuff going on. Julie Neumann is here, CMO at Honeycomb IO. Frequent stories on SiliconANGLE about you guys. Coverage at KubeCon. Thanks for coming on. Appreciate it.

Julie Neumann

>> Yeah, we're really excited to be here.

John Furrier

>> So a lot of conversations is that we've been seeing this cloud native world kind of connecting in with the AI native world. And for the most part, it's a lot of the same stuff that if you look back in the past decade, Kubernetes is now leveled up. It's kind of stable, very boring and in a good way. And that's setting the table for a lot of stuff we're hearing here at Amazon re:Invent, AWS re:Invent is that okay, end-to-end workflow, a lot of data, AI factory sovereignty, that means you got to figure that out. So a lot of instrumentation, telemetry, networking, a lot of stuff going on in AI right now that rhymes with observability concepts. We've seen this movie before. So that's a nice tailwind for you guys.

Julie Neumann

>> It's a super exciting time I think generally in tech right now. Just the pace of innovation, how quickly and everything is moving. It does remind me of the move to the cloud at 10X speed. It's crazy just to see how quickly we are moving. And absolutely, observability is critical in that space. The faster you're moving, you need to make sure that you're moving in the right direction, you're getting the right results. How can you trust the speed that AI is giving you? And that's where observability comes in.

John Furrier

>> We had our analyst kind of session where we break down the keynote analysis and Paul Nashawaty who covers AppDev were talking about how cloud abstracted away the infrastructure. Hello SaaS with SaaS applications, and then the cloud native DevOps community levels up. We get orchestration with Kubernetes, all these services, shift left for security. So all those things that were going on to build that out is replaying now. So I want to get your thoughts on what you guys are seeing with this wave because it's kind of the same game. It's not a pivot in any way. It's more of an extension to cloud native, and it's almost like that culture is, I won't say more advanced, but they've done a lot of hard work in cloud native. And the AI has been kind of like Wild West, vibe coding, a lot of experimentation, which is good. But now when you talk about end-to-end workflows and frontier agents, which is the big news here, you got to track stuff. State of the data, what are the agents doing? How do you measure them? How do you evaluate them? All these things are now on the table. What does that mean for you guys and the industry? Because to me, it seems like you got to have the data. Now agents can do a lot of work in that area. What's your position on that?

Julie Neumann

>> Yeah, I think that data really is the currency. I think when we think about AI and understanding data is incredibly valuable. Data is also coming at you now at a volume. I think that, again, somewhat unheard of, another massive acceleration and the amount of data you need to look at coming out of applications, coming out of the user experiences, coming out of your agentic workflows. So I think helping teams make sense of that data and understanding what is going on in software delivery life cycle, what's going on in production, what's happening before production. Can you actually get in front of some of these issues? There's just, I think a lot of people taking a step back now and looking at sort of what they've experimented with, what they've started to deploy, even things as simple as gen AI code, vibe coding, like what has been happening in your systems and getting a handle on that. And I think that is where observability really can make sense of a lot of that data now that's coming in and helping people see what's happening in production, what's happening with my agents, what's happening out in the world.

John Furrier

>> Complexity in the software life cycle is interesting because I was talking to Matt Garman in my exclusive with him and we're talking about how, okay, Copilots, chat before frontier models came out was kind of a toy. And then as it became more of a system, it was much more production already. So the scrutiny now is, you hear a lot of it in the keynote, production. So accelerating things into production. Now this is what the AppDev world's been doing for years. How do you get stuff into production? So I have to ask you, how do you see that move? Because when you're in production, you need to survive and thrive with the agents. So what does that look like to be successful in that model? Because seeing acquisitions like Chronosphere and Palo Alto, to me, that's a sign that you're going to start to see a lot of M&A for people who may not cross over. There's almost like there's a playbook. What's your view on the readiness? What does success look like in the AI era as these worlds come together?

Julie Neumann

>> I think when you think about the Chronosphere acquisition, I think that we see a lot of really strong validation in the market right now for how critical observability is to help you get ... Like we said, AI gets you speed, but you need trust. You need to understand what is happening in those systems. And I think where observability fits in that stack is really, really important, but not even so much as the tooling and the data. It is the practice of understanding what are agents doing in production? What are your users experiencing? How do you know that? Especially when you have now so many non-deterministic systems, how can you get a better handle on what is actually going to be happening in production and how you're going to be taking this really accelerated innovation and applying it in the correct direction and getting the outcomes that you want for your end users, for your teams, and really getting a handle on that.

John Furrier

>> What's interesting is I had last year at re:Invent here, I interviewed Lori Beer. She's the CIO of JPMorgan Chase. And I asked her, how many gen AI apps do you have in production? She said zero. And I'm like, "Well, we have a lot of machine learning because they do fraud detection." And I asked her, "So why?" She goes, "Well, it's just another app to us." And when you look at the software life cycle with say DevOps, and you're seeing agents announce, they announced that one of the agents was DevOps and they have a security agent and then obviously a Kiro for the software, autonomous software. So you're seeing Amazon already kind of bringing stuff to the table to kind of prime the pump, but from a software development standpoint, GenAI is just an app and generative is a runtime. So okay, if Nova Forge is the substrate and AgentCore is the runtime, that kind of sets the table for kind of a DevOps vibe, if you will. So what's your reaction there? You're smiling.

Julie Neumann

>> I think this is something that our founder Charity Majors has talked about quite a bit that like AI it's just another software problem. This is not some crazy new thing per se. Again, I think what we've said is the amount of data the speed, like that is quite different, but it is just another software development challenge. Just like anything else, you need good DevOps practices. You need to understand what's happening across your systems. And these systems are getting increasingly complex. It's not to say just because it is another software problem that it is an easy problem. I think that's what makes it so fun right now, but it is. And we see people like the JPMorgan Chases of the world learning how are you getting those things now from behind the scenes more into protection and treating it like a software problem.

John Furrier

>> And discipline.

Julie Neumann

>> Yes.

John Furrier

>> It's not about favoritism, the shiny new toys, but okay, we have a high resilience bar. Sorry, you have to clear that. I have to ask you this because one of the benefits of doing the queue for 16 years is you meet a lot of people over the years. And I ran into an engineer that I interviewed in 2012 and she was amazing. And so I ran into her GTC event in Washington DC and she's been so hardcore on DevOps with cloud native. I asked her what she's doing there and she's like, "Well, I really haven't figured it out yet." So you have this cultural DevOps crowd community that we've been covering with KubeCon since the beginning that have been just jamming so hard on these hard problems and then they kind of wake up and it's like, I won't say Neverland, but it's like, "Okay, AI is kind of crazy, fast and loose right now." But she was trying to orientate towards what's the problem. And so what's your thoughts on that? I don't say it's a culture cloud, it's more of, I think it's again, it's not a pivot. Cloud native is driving, look at Amazon's EKS container native strategy, that's clearly setting the table for this world. So there's the merging of those two communities, how do you see that happening culturally?

Julie Neumann

>> Yeah, I think that you're right that it's not a culture clash. I think it's an evolution, right? And I think that there is a lot that we, we see this in our own teams and we see it with a lot of our customers that there is a lot that you can learn from traditional DevOps practices, and I think there is also a lot of things that traditional DevOps people are learning on. There's some new ways to do things and it's kind of opening up the world a little bit in terms of how could you approach things differently, faster? Where do you think about how agents fit into your systems in different ways? I think that once we really start to see these things becoming truly integrated systems, I think that's where it's going to get really interesting on, you do need both sets of, you need these hardened practices that people have been really embracing for years now, and there's been a ton of success. The last time we had a major pivot in the tech industry going cloud. We develop those things in that sort of environment, and I think now it's an opportunity to update them and ask tough questions of them, learn some new tricks, but rely on the stuff that got us here as well.

John Furrier

>> Well, I was fascinated by the three major announcements that I thought jumped off the page to AI factories, validation. Well, something we've been talking about on theCUBE a lot, Amazon to acknowledge that essentially is good because it's now hybrid environment. Nova Forge, I think is killer and then frontier agents, which it gives it more scale and more intelligence. In order for that to happen, Amazon is also investing in lower cost price performance to get the power and the performance for tokens per watt going, right? Because it's going to be still a cloud game. So the AI factories is kind of an on prem cloud thing. If you're buying a big NVIDIA rack, if you're a neo-cloud, okay, you can buy a big factory and spend zillions of dollars and have NVIDIA finance that or whatever, but that's not the average enterprise.

Julie Neumann

>> Correct.

John Furrier

>> They don't have the CapEx. So what's your vision on how you see the market going from your perspective? Because the enterprise is going to want to have that data. The Nova fits for them. They can take that frontier model, make it their own, use their own data estate in there and create a custom model. Reddit is doing it. I talked to two other customers and they're seeing great results. So I think that's going to be a thing for sure. That means on prem, we'll still talk to the cloud.

Julie Neumann

>> Yeah.

John Furrier

>> And so a little bit of a data center outpost kind of feeling going on with Amazon. How do you see that piece going? Because I think this is going to, if this happens the way I think it's going to happen, the enterprise should open up and it's got a sovereign vibe because again, it's their data. They want to measure it, multi-tenant. What's your take on that whole on prem enterprise piece?

Julie Neumann

>> I think it's providing flexibility for enterprises, right? And I think especially with the cost associated with this, obviously that can easily run out of control. I think it is making sure that enterprises have the ability to pick and choose what's going to work best for them. And so I think that's some of the interesting things around the Amazon announcements today is there is a trend towards bring your own cloud and how are people doing more private cloud? How are they mixing private cloud on prem, off-prem workloads? And so I think that that ... Some of the things I think that were really interesting about the Amazon announcements today are really how are you empowering people to mix and match and actually get some of those efficiencies and figure out what's going to get them where they need to go. And especially when you have some of these really large at scale enterprises, I think they are going to want to take a lot more control over it. But things like the factory announcement shows that they also are still willing to figure out how they can help people connect the dots in different areas.

John Furrier

>> I think I asked about Honeycomb's business in your journey and you guys in the market, we talked about the Chronosphere acquisition three billion plus for that, a big number and they had pretty good revenues, but most people say they were kind of narrow in scope for what they had capabilities for. You mentioned it's getting bigger. If you go back maybe two years ago, there was everyone's unobservability. It seemed like everyone was doing observability and then there was speculation that that might consolidate, but interesting that you bring up the aperture of the TAM.

Julie Neumann

>> Right.

John Furrier

>> So talk about that because I think there's, I won't say reset, but almost like a new view on what the market looks like for observability. What is the new definition?

Julie Neumann

>> The new?

John Furrier

>> The new view, because it's just now it crosses over. It's almost horizontal. It's not like a category. It's not a product per se. It's a capability.

Julie Neumann

>> It is. I think it's a capability and I think it is a critical practice that really can go quite broad in who it's applicable to. So obviously with Chronosphere going to Palo Alto, a little bit of a different ICP alignment there and getting more into the CIO world, strong alignment with security. I think that we're seeing a lot of really interesting patterns around how critical observability is to AI and to more of the larger foundation models and sort of like what are they looking at. And so I think you have now observability just becoming so critical that more so than I think consolidation within the space. I think that it's other larger players saying, first of all, it's an expensive problem, right? And so there's a lot of work we do at Honeycomb to make sure that we are helping customers with their data, how that's flowing in and out, how they can be smarter about what their expenses and their costs are. But I think there is a lot of interesting crossovers between some of these, between security, between different DevOps tools and to looking at like how can you create more of that intelligence layer so people can be moving faster and that works in multiple areas. And so I think that there's a lot of movement happening right now in the observability space that'll get really interesting in the next couple of years.

John Furrier

>> It's interesting. If you go back, I remember when we first started talking about observability, it was like a segment, and then it's now industry scope. So it's much wider.

Julie Neumann

>> Yes.

John Furrier

>> And I think having the adjacency to the AI side is interesting because now all those same problems exist. You take software supply chain containers, cloud native services. It all looks the same over here. They just call it something different.

Julie Neumann

>> Right. Yeah.

John Furrier

>> Frameworks, MPC, A2A. So you have protocols, you've got frameworks. It's an app dev.

Julie Neumann

>> It is. It is.

John Furrier

>> It's a whole nother app dev open.

Julie Neumann

>> It is a very apt dev shaped problem, I think. And looking at just again, how critical it is to actually understand, you can hook those tools together. Think about the Claude Code environment on you can now move so fast as Charity likes to say, writing code was never, that was never the hard part. How you get code to do what you want it to in production, that's the hard part. And so when you are looking now as like app dev becoming a very AI problem, how are you making sure that ... And it is also like security has a lot of these problems and being able to see what's going on. And so I think that's the direction you sort of saw with the Chronosphere acquisition, but I think there's a lot of really interesting ways that we connect more into the AI space generally and observability why it's so critical and why you really do hear it in all of these AWS announcements as just if you're going to be innovating, if you're going to be embracing AI, make sure that you have observability alongside it.

John Furrier

>> I think what you said earlier about the software development life cycle is interesting because the code assistant and now the autonomous coding is the same vein. It's the linchpin to cloud native, one, because of the services. And two, the role of the developer for charities comment is like, "Okay, you're orchestrating, you're directing."

Julie Neumann

>> Correct.

John Furrier

>> You're the director, not the coder.

Julie Neumann

>> Yeah. Your responsibilities and engineering team now is outcomes, right? Every company is now a software company. That is how people are engaging with your business. That's how you're making money is how people are experiencing your applications. And so it really puts the customer experience very front and center for engineering teams. So you need to understand, you're no longer in the background. You're really close to the customer and how your company makes revenue. And so you really need to understand what those outcomes are, which is very empowering. It is great to see that evolution. I think that engineering teams have long been the innovation center of business. And so to really understand what's happening for them, it's a really great move forward.

John Furrier

>> What's interesting is that if you go back to the DevOps days, well, it's still there, but the software developers were actually real developers and they were shifting left for security and doing all these things. Now you have things like quick. You have now business people coding and the citizen developer was a concept. I'm not sure that's a word still, but like now you're going to have essentially like business people coding just by either, I won't say vibe coding, but vibe coding, just doing commands, write new software that does this. So the DevOps serve developers, now they'll be serving agents.

Julie Neumann

>> Yes.

John Furrier

>> So that seems to be the big wave. Agents are just fungible to the developer. Not fungible, more like the same service. They still got to run on infrastructure.

Julie Neumann

>> Right. And we actually, that is a principle that we talk about at Honeycomb is like, we want to make sure that our telemetry data, like our product, that it is built for engineers, it is also built for agents. You need to be able to talk, I think both ways. You need to consider that, and if you have great data, if you have powerful data, which is something we pride ourselves on, it's such a powerful input to agents and to AI development. And so you want to make sure that that's accessible.

John Furrier

>> Well, super exciting for the observability because it's no longer just a segment scope. It's more industry-wide. So it's a tailwind for sure and market TAM expansion. So exciting there. Put a plugin for Honeycomb, give some stats on the company, what you guys are working on, key value proposition, put a plugin.

Julie Neumann

>> Yeah, so I think we've had a really exciting year at Honeycomb and we have a big year plan next year. So Honeycomb created the concept of observability almost 10 years ago now. So next year is going to be the 10-year anniversary of Honeycomb and of observability, and it's amazing to see how much it's evolved yet to the discussion we've just had, how critical is more important than ever. So we've really been thinking about what is the next era of observability? What do the next 10 years of observability look like? And so this past year, we've been doing a lot on how are we setting the foundation for LLM observability and getting AI built in. There's so many great ways to apply AI to the observability process and being able to move faster with data, but also making sure that it is not just the top 1% of engineers at startups doing cool stuff get their hands on these products, that it is scalable up into the enterprise. And so we've been doing, had some really cool customers doing really great things this past year, being able to run alongside great innovators like a Netflix, but then also big financial institutions like a Vanguard and helping everybody come forward into like, first it was how do you get people into the cloud era and now it's how do you get people into the AI era and help them continue to evolve.

John Furrier

>> So you're responsible for Netflix's innovation about shutting down all my shared passwords?

Julie Neumann

>> Well, no, not that part.

John Furrier

>> Can you send me the new password or the QR codes? Sorry, you're busted. No, Netflix has always been a pioneer. They really were-

Julie Neumann

>> What they've been doing with live events now and that need to have fast feedback loops. And so many businesses right now, you need fast feedback loops in the age of AI. And I think that is where the depth, the breadth, the context that we have at Honeycomb with our data and the speed at which we can operate and really give those fast feedback loops to developers. I think we're seeing really cool stuff right now.

John Furrier

>> Velocities, to your point, is huge right now.

Julie Neumann

>> Massive. Yep.

John Furrier

>> Julie, thanks for coming on. Appreciate the update. Great to see you guys. Congratulations on the momentum and continued success. Thanks for coming on, share it on theCUBE, appreciate it.

Julie Neumann

>> Thanks for having us.

John Furrier

>> All right, this is theCUBE live coverage. I'm John Furrier here. More, three days of coverage. Day one is just kicking off. More. Stay with us for more after the short break.