Jul 3, 2026 · 19 min · 11 segments
What does the future of FinOps and AI spend optimization look like heading into 2027? In this interview recorded live at FinOps X in San Diego, we sit down with Zach Johnson to break down the biggest…
Zach JohnsonGuestVictorHostAnd well, I'm going to start with the first thing.
The first day, you know, we had the announcement of the Tokenomics Foundation and everything about tokens, tokens, tokens, tokens.
So, you know, what are your thoughts? How do you envision this? How do you think? What do you think is going to happen?
Yeah, I think the industry is still struggling to define what people define as a universally agreed upon KPR for metric.
And I think as all of the providers are moving away from named seats to usage-based or token-based consumption, people's budgets are growing out of control.
And once they're growing out of control, when finance starts to ask, what's the value I'm getting for all of this money? And help me understand it in a metric that we can all agree upon.
Yeah, that's true.
I think when things get out of control on the budget way, then finance starts to ask questions.
And when finance starts to ask questions, it gets difficult to get answers.
And then the token gets redacted and then the things going on, right?
And we want to kind of figure out how can we not slow down all the velocity that's coming out of engineering from using these tools, but how can we not blow through our AI budget
like Uber did in a three-month
timeframe for 12 months? And then the forecasting, you know, finance really struggles to forecast growth that is growing in an exponential hockey curve, you know, growth pattern, you know, and there's just not a lot of financial modeling out there that can support growth of that
volume.
Yeah, you need to support it with exponential growth and that's not easy for most of the sectors, right? It's very difficult to grow in exponential unless you're like a new startup.
Yeah, 100%.
And I wanted to ask you about that because from your perspective and from what you've been learning here in the conference, what are the main challenges apart from token matching and token consumption that you are seeing in the conference from people?
I think people are struggling to define what really is a token and what's a credit.
And is there a universally agreed upon sort of like a per hour per CPU? That's the industry can get behind and everybody, no matter what provider you go to.
And I think everybody's trying to understand what's a token cost in one provider versus another.
And well, I'm going to start with the first thing.
The first day, you know, we had the announcement of the Tokenomics Foundation and everything about tokens, tokens, tokens, tokens.
So, you know, what are your thoughts? How do you envision this? How do you think? What do you think is going to happen?
Yeah, I think the industry is still struggling to define what people define as a universally agreed upon KPR for metric.
And I think as all of the providers are moving away from named seats to usage-based or token-based consumption, people's budgets are growing out of control.
And once they're growing out of control, when finance starts to ask, what's the value I'm getting for all of this money? And help me understand it in a metric that we can all agree upon.
Yeah, that's true.
I think when things get out of control on the budget way, then finance starts to ask questions.
And when finance starts to ask questions, it gets difficult to get answers.
And then the token gets redacted and then the things going on, right?
And we want to kind of figure out how can we not slow down all the velocity that's coming out of engineering from using these tools, but how can we not blow through our AI budget
like Uber did in a three-month
timeframe for 12 months? And then the forecasting, you know, finance really struggles to forecast growth that is growing in an exponential hockey curve, you know, growth pattern, you know, and there's just not a lot of financial modeling out there that can support growth of that
volume.
Yeah, you need to support it with exponential growth and that's not easy for most of the sectors, right? It's very difficult to grow in exponential unless you're like a new startup.
Yeah, 100%.
And I wanted to ask you about that because from your perspective and from what you've been learning here in the conference, what are the main challenges apart from token matching and token consumption that you are seeing in the conference from people?
I think people are struggling to define what really is a token and what's a credit.
And is there a universally agreed upon sort of like a per hour per CPU? That's the industry can get behind and everybody, no matter what provider you go to.
And I think everybody's trying to understand what's a token cost in one provider versus another.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
