FinOps X 2026 | Shawn Alpay & Matt Cowsert, FinOps Foundation

Clips
More from FinOps X 2026

Shawn Alpay

Director of Data Engineering

FinOps Foundation

Matt Cowsert

Principal Product Manager

FinOps Foundation

The next frontier for FOCUS: AI token economics

AI token economics is creating a data normalization crisis across the technology stack as enterprises scramble to apply consistent financial controls to graphics processing unit clusters, AI factories and token-based consumption that existing cloud billing frameworks were never designed to handle.As the FinOps Foundation Project, a Series of LF Projects LLC, community gathers in San Diego for FinOps X 2026, attention is turning to FOCUS — the FinOps Open Cost and Usage Specification — and whether the open standard can extend its multicloud billing common language to cover a new generation of AI spending

play_circle_outline FOCUS 1.4 Release & 1.5 Roadmap: Granular Tokenomics, Multi-Persona Support, and Opt-In Fine-Grained Telemetry

play_circle_outline Use both native provider exports and FOCUS: optimization vs unified reporting

play_circle_outline Hybrid and on‑prem AI adoption expands FOCUS scope beyond public cloud

play_circle_outline Community-driven development: open GitHub, daily working group sessions, provider advocacy

Info
Transcript

Shawn Alpay & Matt Cowsert, FinOps Foundation

Shawn Alpay

Director of Data Engineering FinOps Foundation

Matt Cowsert

Principal Product Manager FinOps Foundation

In this interview from FinOps X 2026, Shawn Alpay, director of data engineering at FinOps Foundation, joins Matt Cowsert, principal product manager at FinOps Foundation, to talk with theCUBE's John Furrier and theCUBE Research's Paul Nashawaty about the launch of FOCUS 1.4 and how the open billing specification is maturing into a unified language for technology value. Alpay and Cowsert explain how FOCUS normalizes billing data across public cloud, SaaS and AI providers — giving engineering, finance and procurement teams a shared vocabulary for expressing cost... Read more

explore Keep Exploring

What changes and additions were made in FOCUS 1.4 compared to earlier releases, and what is being considered or planned for the upcoming 1.5 release? add

Should one use FOCUS or the native export for cost reporting and optimization? add

How has the shift of AI into on‑premises and hybrid environments changed the scope or need for a common language for technology value? add

What is FOCUS? add

bolt Powered by CUBE AI

Shawn Alpay & Matt Cowsert, FinOps Foundation

search

John Furrier

>> Welcome back to theCUBE's live coverage here at FinOps X 2026 in San Diego. I'm John Furrier, host of theCUBE with Paul Nashawaty with theCUBE Research, breaking down all the action, getting the data, we're analyzing it. And FinOps is about cost, and if you don't get control or understand your data, that's a huge problem. It's a big part of this community's work. They have the FOCUS project. FinOps, open cost usage specification. The leaders are here. Shawn Alpay, director of data engineering, FinOps Foundation. Thanks for coming on. Appreciate it. Matt Cowsert, principal product manager for the FinOps Foundation. Thanks for joining us.

>> Thank you very much. Yeah.

John Furrier

>> You guys are doing great work. Obviously cloud going strong, getting good work done. AI hits the scene. More cost analysis. The data's out there. There's more data lakes than I've ever seen before. You got neo clouds. You got AI GPU clouds. Everything's happening so fast. So take us through what's changed in FOCUS, which is an open source initiative around billing cycles and all the data. And there's more data than ever before.

>> Yeah. I think you've set it up well in terms of the challenge. The data is not normalized across providers, across technology categories, public cloud, AI, SaaS, et cetera. Being able to tell that story with the same names of the columns, with the same definitions of the allowed values, what's allowed to be put in those files, in those columns. Being able to have that story told the same way across all providers, it's incredibly valuable. And then to be able to land that into your data lake, your data store, your data warehouse, just shorten the time it takes to get to market for your organization's analytical needs, for your accounting needs, your financial needs, whatever. That was all possible before, but we're just trying to really help people do that faster and more easily. And yes, of course, we started with public cloud support, moved into SaaS. Of course then talking about AI and the Tokenomics Foundation and all of those things that are coming along very dynamically. Yeah, it's certainly a big challenge. And I don't know if you have any thoughts on that.

John Furrier

>> Yeah. I think the biggest thing that we've observed is we have new technology categories that we're introducing. We have new providers coming on board expressing interest every day. But what's nice is because we've set the foundation already within FOCUS, a lot of what we're observing is that these patterns, these billing behaviors, these pricing patterns, they already fit into what we've defined within FOCUS and for the things that are different, we're then spending the time and effort within the working group to define what that looks like.

Paul Nashawaty

>> So I'll tell you, from my perspective, I cover application monetization and application development, right? It's a big growing area. Monetization, as you know, is a treadmill. Everyone's always running. You never really stop. But when you look at it, our research shows that 20% of application portability is critical to the organizations. 67% of respondents indicate that it's very important to them. This is incredibly important because when you think of something like FOCUS and what you're doing, what you're harmonizing across these different platforms, you're harmonizing the data, you're allowing for these different access points to come in, whether it's heritage systems or new systems that would come in, you're bridging the gap between old and new. One thing I recognize year over year with FOCUS from 1.3 and now in 1.4, you've done a lot. So why don't we touch on that and expand on that. What have you done in 1.4 that some of the deltas from the past?

>> Yeah. So in 1.3, we launched our first second data set for FOCUS and we're now expanding into additional data sets that are complimentary to the costume usage that we originally started with. And so what you're seeing is we're both bringing in new technologies as categories, but also bringing in new personas into FOCUS as well. So we started with the FinOps practitioner, we started with the data engineers and then we moved into leadership, but then we're introducing finance and then we're introducing the procurement. And so I feel like we're able to tell that story. We talked before around FOCUS is a specification, but FOCUS is also a language that you're then able to leverage across your organization. And I think with our last two releases, we've done a really solid job of making sure that those additional personas in your organization actually know the value that they can get out of FOCUS as well.

John Furrier

>> So as 1.4s out, now 1.5, an odd number, we all know odd numbers are better than even on releases, what are you guys considering? Only kidding, by the way. That's an OPC joke. 1.5, what are you guys looking at? What's on the horizon? What's getting your interest? What are you focused on?

>> Yeah. So it's incredibly important, I think, to think about, when it comes to token counts and everything AI related, you're asking questions around token economics. It's not just the tokens that you burn through. It's also what's the value, what's the outcome that came from that? So in order to tell that story, you need not only the token count, but you also need that information per various entities. Yes, per service, per resource, but when it comes to AI and tokens in that realm, you're starting to talk about per user. We're starting to talk about per session, per request, per operation. And the farther down this ladder you go, the larger the cardinality of the data and the harder it is to gather that information in order to tell that story. We were just chatting with a member of the steering committee, they're making a good point, it could end up being that a non-trivial percentage of your cost is then having to be attributed to just getting your data and putting it through a pipeline. It's a non-zero cost just to tell the story. So we have to be really thoughtful about how we deliver that information, potentially allowing you, as a consumer, to choose whether to consume data at that grain or not. You want to opt in or out to using per session data, per operation data, per user data. So we're having those conversations actively in the working group and we'll be certainly here at the booth over the next couple of days.

John Furrier

>> I mean, the engineering side, we said data engineering problem. There's more data coming in, you got to slice it. There's policy, there's service level. It feels a lot like an OS kind of problem.

>> This is really where, I mean, maybe you want to talk about this a bit, but like this is where billing and observability data really starts to blur.

John Furrier

>> Yeah. So there's some foundational things that we want to get right when it comes to understanding your AI consumption. So your model identity, your model family, like there's individual dimensions that currently sit within like your SKU price details. And so like how do you break that down in a very structured way? So it's not just what your consumption is, but it's around like what outcome are you trying to drive with it and it's going to vary depending on the model that you're using. And so I think right now because like the AI consumption is so new, it's kind of like where we were at with cloud 10 or 15 years ago where you don't know what you don't know yet. And so there's some uncertainty that adding the additional granularity is going to help believe that. But for us to do that, like whether we're talking about doing like a net new data set for observability or we're tying into something that exists today like the open telemetry project, like we're still trying to figure out exactly what that looks like. But at the end of the day, there's going to be very specific use cases where understanding that session ID, that particular request is going to be really meaningful.

Paul Nashawaty

>> There's a lot of conversion going on in these different pipelines, right? We see the CI/CD pipeline converging with the telemetry pipeline, converging with the security pipeline, and it's all in an effort to accelerate delivery. So one of the things when we're talking and when you were describing, I like the audience to understand that there's different levels of entry points here for FOCUS. It's not like you have to be a full on expert in order to do this. You can come in at different stages and your applications can come in at different stages as well. So maybe it would be good to kind of double click down on that and talk about what does this mean if somebody's just getting started versus somebody that might be fully mature?

>> I think it's really challenging to build a specification that solves for the data savviness across the entire spectrum, right?

Paul Nashawaty

>> Yes.

>> You have to be able to solve for a specification of somebody who's going to download a CSV and dump it into Excel all the way to a person who has a whole data engineering team and is operating at scale of petabytes and everybody in between. A lot of the influencers in our working group are the people who experience these challenges at the higher scale every day, but we definitely have other people in the working group who are just trying to like take tens of megabytes or hundreds of megabytes or gigabytes and they're solving for different kinds of challenges because they're so limited in their support in their organization. So it's not just sides, it's also understanding the data that's available there. Some people are coming at this from like a data engineering background, some are coming from an accounting background, finance background, a non-tech background. So it's a challenge for sure to be able to tell that story across those dimensions, but it is necessary to build as big a tent as possible, to bring all those people along and then tell that story, to encourage all those folks to advocate to their providers, "Please generate data in the FOCUS format so that we can grow the list of data generators from the 15 we have today to the hundreds if not thousands that we see coming in the very near term."

John Furrier

>> That's a great observation about the multidisciplinary participants. So how does that change people's decision? I'm sure the enterprises I took, they look a lot like the FinOps Foundation because there's a lot of constructive conversations in projects. It's happening in the enterprise. How should people think about preparing the data when they try to solve the token economics problem or frame it? I mean, just to zoom out like, okay, what are we talking about here?

>> Yeah. I think it's really a challenge to know how much data you need to answer those questions. So we were just talking about this in the working group today in our meeting, that there's going to be these smaller and smaller concentric circles of people who care about more and more niche use cases. We might all care about telling story per user, but you might not care about telling the story per request and I might. And if I ask that question, what is the cost or usage per request, what am I solving? What is the actual thing I'm going after? And you may not need that data all the time. And so having the agency to decide whether you want to consume that information or not is really at the heart, I think, of the need there.

Paul Nashawaty

>> Yeah. I think the other thing to keep in mind as well is like when I first joined the foundation, I got this question a lot of like, should I use FOCUS or should I use the native export? And the reality is that you need both of them. The native export, it's very critical from an optimization standpoint if I'm an engineer, but if I'm trying to report my cost across my five providers, FOCUS is a much more palatable way of going about doing that and that way you don't need to reexplain the way that you're like delivering value through your technology usage and choices. And that's again, I think like when we're talking about like what's the value of FOCUS, it's one language for your executives, one language for your finance team, one language for procurement. And then if there's a hyper specialized optimization need for your engineers, then it may make sense to dive deeper into the native format.

Paul Nashawaty

>> I like that. I like where you're going with that. And again, that kind of addresses the crawl, walk, run approach, and you can enter it at any point. But I do know with our research, we see that 67% of organizations are hiring generalists over specialists because the two main barriers they run into is complexity and skill gap issues. And whether they're pulling the native format and trying to compile data together in very siloed approaches or using a unified view or a harmonized view like FOCUS, it seems like the generalists will be able to use the FOCUS approach much more easily than trying to compile different things together. Is that fair?

John Furrier

>> Yeah, I think that's a great way of thinking about it. So we've talked a lot about FOCUS as a specification, but the language of FOCUS permeates through job descriptions and it permeates through training and it's really like using FOCUS as the lexicon for expressing value, whether you're choosing to use a particular version of the SPAC or whether your favorite provider supports FOCUS yet or not, like that's kind of secondary to being able to communicate in a way that's understood across your entire organization.

John Furrier

>> FOCUS began as cloud billing. We talked about AI earlier. AI is exploding in the data center and the on premises in the enterprise because it's hybrid cloud. Okay, great. We like that. But now the users are using AI so it's not going to the cloud and the data's in the enterprise. So you got the users adopting, that's causing a lot of engagement. So a lot more on prem services are being consumed. Smaller AI factories, different size clusters, smaller, faster, cheaper gear It's bringing it into the enterprise. How has that changed the scope or the need for a common language?

>> The slogan of FOCUS is the unifying language of technology value, and that slogan has been true since inception, but it's a question of whether we're fulfilling on that promise. And so as we talked about earlier, we started with traction amongst public cloud. We always have said that any technology category, any technology provider can generate focused data. You simply have to add more columns that we call custom columns.

Paul Nashawaty

>> Yeah.

>> Now, okay, that's good enough, but if we really want to look people in the eye and say that we support AI or data center, we have to go through a scrubbing exercise, an audit exercise to make sure that we have everything that the many subject matter experts that we talk to in these spaces believe that they need in order to answer those questions. And so we do recognize to some extent, because we're coming from public cloud roots, that there could be a perception of square peg, round hole, but I do believe that FOCUS in its current format can be extended and is being extended to answer the questions across all those technology categories.

John Furrier

>> And the value is being unlocked on the enterprise. Big time. All right, final point. Put a plug in for FOCUS. How do people contribute? What's going on? What are you guys looking for? What are you optimizing for?

John Furrier

>> Yeah. So what is FOCUS? FOCUS's entire job is like how are we normalizing the expression of billing data across whatever technology it is that you're using? That's our goal. So if that's something of interest to you, we would love for you to join the working group. We literally have a session every single day. So if you want to hear me talk more, that's where you go. But like we have synchronous engagement, we have an open repository on GitHub. So we build in public that's an extremely important part of our operating model. And again, it's not Shawn's idea or Matt's idea, it's the community's specification and we adapt and evolve to meet the needs of practitioners and then we go lobby on behalf of the practitioners to get the right set of data generators to support FOCUS.

John Furrier

>> And so someone's goal might be to turn on their dataset, for instance.

Paul Nashawaty

>> Exactly.

>> 100%.

>> If it's not super interesting to you to spend every week interacting with us to build a specification and you simply want to enjoy the fruits of our labor, then please we beg of you to go advocate to your providers to either generate FOCUS data or to advocate for generating a newer version of FOCUS data. Because we are announcing FOCUS 1.4 here at the conference, if your provider's on 1.2 and there's functionality in 1.3 or 1.4, definitely advocate to your provider to use those newer version.

John Furrier

>> Versions. Shawn and Matt, we love what you do. Thanks for what you do.

>> Thank you very much.

John Furrier

>> Providers watching out there, put the hooks into them, get that focused data.

John Furrier

>> Build the agents.

John Furrier

>> We need that token data, especially if you're upgrading, moving from one model to the other. You got to burn tokens just to kind of make sure everything works. Guys, thanks so much for coming on theCUBE.

>> Thank you very much. Appreciate it.

John Furrier

>> I'm John Furrier with Paul Nashawaty. FinOps is becoming the most important story because to understand the value, you've got to understand the data, you understand the metrics. Being open, having one common language is an approach people are liking, certainly here to show as FOCUS 1.4 hits. And of course we're looking to see what happens with 1.5 and an odd number, which I'm superstitious. I like odd numbers. Guys, thanks for coming on.

>> Thank you so much.

John Furrier

>> Thanks for watching.

>> All right.