By Enterprises are investing in AI while carrying unoptimized cloud infrastructure underneath it. The result is that every dollar of AI spend lands on a foundation that is already leaking, and visibility tools and spot automation solutions do not change that equation. The organizations that will absorb AI costs without destroying margins are the ones that have already solved the application layer.
In this interview on TFiR, Peter Maloney, CFO and COO at Azul, breaks down why cloud cost optimization is the prerequisite to AI adoption, why most FinOps tooling stops short of where real savings live, and how Java runtime tuning delivers the most pervasive and material cost reduction available to enterprise teams today.
Guest: Peter Maloney, CFO and COO at Azul
Show: TFiR
Here is what every FinOps practitioner, platform engineer, and enterprise architect needs to know.
Technical Deep Dive
Q: Is there a hidden cost to AI adoption that enterprises are not accounting for?
Peter Maloney, CFO and COO at Azul, confirms that the hidden cost is real and structural. When cloud spend is already unoptimized, AI workloads do not arrive in a clean environment. They compound existing inefficiencies, driving up token costs, compute spend, and operational complexity simultaneously. The enterprises most exposed are those that have not yet established clear, outcome-defined cloud spending before layering in AI investment.
“AI can just compound the problem.” — Peter Maloney, CFO and COO, Azul
Q: What is the relationship between cloud cost optimization and AI adoption readiness?
Maloney frames cloud cost optimization as a direct enabler of AI adoption, not a separate workstream. Companies that have their cloud spend under control free up budget to invest in AI faster, and their employees experience smoother AI tool adoption because the underlying infrastructure is not fighting against them. The two are linked: cloud discipline creates the financial and operational headroom that AI requires.
“Companies that are focused on their cloud cost optimization will have an easier time adopting AI and will free up dollars to be able to more quickly invest in it.” — Peter Maloney, CFO and COO, Azul
Q: Will AI cost optimization follow the same industry pattern as cloud cost optimization?
Maloney believes it will. Just as an entire industry emerged to help enterprises manage and reduce cloud costs, he expects an equivalent market to develop around AI cost optimization. The pattern is consistent: unchecked spending creates pain, pain creates market demand, and tooling and services follow. The lesson from the cloud era is that organizations should not wait for that market to mature before acting on cost discipline.
“Just like there’s a whole industry around helping companies optimize their cloud costs, I think the same thing is going to happen with AI.” — Peter Maloney, CFO and COO, Azul
Q: Are enterprises applying quick fixes or pursuing long-term architectural solutions to rising cloud costs?
Maloney separates the response into two distinct stages. The first stage is the consumption and commercial bucket, which includes automation tools from cloud vendors and spot solutions that provide visibility and basic automation. The second, more impactful stage is going to the application layer, where tuning Java runtimes and modernizing applications produces returns that are both pervasive and material. Most enterprises stop at stage one and leave the largest cost reduction opportunity untouched.
“What you really need is to go to the application layer. Companies that focus on tuning their Java runtimes and modernizing their applications are the ones that will get the biggest return.” — Peter Maloney, CFO and COO, Azul
Q: Why are visibility and automation tools not enough for meaningful FinOps results?
Maloney distinguishes between tools that surface information and tools that take action. Phase one FinOps tooling, covering visibility, governance, and accountability, is valuable but passive. It tells you where money is going without changing how efficiently the workload runs. Application-layer optimization is where action actually happens, because it changes the compute efficiency of the software itself rather than simply reporting on its consumption.
“A lot of the tools in phase one of FinOps are focused on visibility, governance, and accountability. Where you can really take action is by making your application layer run more efficiently.” — Peter Maloney, CFO and COO, Azul
Q: How does Java runtime tuning deliver FinOps results at scale?
Maloney describes Java runtime tuning and application modernization as pervasive and material, two characteristics that distinguish it from point solutions. Because Java underpins a large share of enterprise application workloads, optimizing the runtime layer affects cost across the entire application portfolio rather than in isolated instances. Azul’s customer base demonstrates this pattern, with organizations using Azul’s solutions to achieve this level of FinOps impact directly through application-layer efficiency.
“That is pervasive and material. And that is actually what Azul is very, very good at.” — Peter Maloney, CFO and COO, Azul
Resources and Documentation
- Azul, Java runtime solutions for cloud cost optimization and application performance at enterprise scale
***
👇 Click to Read Full Raw Transcript
Swapnil Bhartiya: I was looking at the report, I think almost 56% of CFO said that AI and automation is going to be their top financial priority. Yet ironically 43% admit that AI is making cloud cost harder to manage. There is a hidden cost of AI itself. Token cost can pile up very quickly. Is there hidden AI tax that organizations are walking into with their eyes open? While yes, automation, AI, it seems to be low hanging fruit, seems to be very exciting. But what are the things, what are the lessons we can learn from the cloud era? That even here, just the way unchecked cloudways code not only was killing roi, we can make the same mistakes with AI as well.
Peter Maloney: No, I agree. I think once again, right, making sure that you have your cloud layer and your cloud spend under control and running efficiently with accurate and clearly defined outcomes is important. AI can just sort of compound the problem. Right. I think what we’re going to find is the companies that are focused on their cloud cost optimization will have an easier time of adopting and getting adoption from their employees, their AI tools. You’ll also see that they’ll free up dollars to be able to more quickly invest in AI. And so I think they’re completely linked. I actually think just like there’s a whole industry around trying to help companies optimize their cloud costs, I think the same thing is going to happen with AI. Right. And so you have to start with your cloud costs, make sure you understand them, make sure that you’re spending for real outcomes and that will help you to be able to implement and get traction with AI.
Swapnil Bhartiya: I think the only difference between AI and other is that AI can actually help you also to figure out how to tame. I mean we use AI to solve the problem that AI creates here as well. So that will happen. If you look at, in general, when you look at how enterprises are responding to rising cloud cost, at the same time they also want to invest in AI. But we also hear of course that cost also climb up. Are we looking at something called quick fix band aids or we should look at long term solutions which could be re architecture which is more optimized for AI, more for automation, more for efficiency, performance, leaner, smaller teams also which is also ready because AI world is changing so fast, new technologies are coming so fast, openclaw or mcp so there is more predictability, being able to move faster and also being able to use all those latest technologies.
Peter Maloney: Yeah, you know there’s a lot there. And I’d say as I said earlier, I think there’s a couple of stages of getting ready and preparing yourself and doing better at managing your cloud costs and optimizing. One is I put it in sort of the consumption and commercial bucket and some automation tools, right? There’s a lot of automation tools out there. The cloud vendors provide them. There are spot solutions that are very good that are used, but that really doesn’t give you anything more than just visibility and automation. What you really need is go to the application layer. Companies that are focusing on tuning their Java runtimes and modernizing their applications and their Java environment are the ones that are going to get the biggest return. And by doing that, it’s more pervasive and you’re actually investing in something that takes action. Right? If you think about a lot of the tools in sort of phase one of FinOps that you’re using, it’s for visibility, governance, accountability, where you can really take action is something that is going to help you to make your application layer run more efficiently. And that is pervasive and material. And that is actually what Azul is very, very good at. We have many customers using our solutions to achieve that level of finops. Excellent.
