Operations | Monitoring | ITSM | DevOps | Cloud

For most of the last decade, the direction of cloud strategy was clear: standardize, consolidate, and reduce sprawl. Engineering teams worked to pick a primary cloud, reduce vendor dependencies, and simplify their stacks. FinOps teams unwound years of fragmentation. Platform teams built guardrails to make sure it didn’t happen again. Then AI arrived, and it’s a fundamentally different class of workload. AI demands specialized hardware and, increasingly, diverging providers.

AI is the new frontier for FinOps maturity. It introduces fresh spend patterns and new opportunities for value. As GPUs, inference, and retraining reshape costs, FinOps maturity grows through visibility, forecasting, and shared mindset about how these workloads drive business impact. In this 2025 post, I gave my guidelines for implementing AI tagging to give business context and clarity to vague AI invoices. Now, I’m sharing the next level up: how to drive FinOps in AI with AI.

When Anthony Leung talks about FinOps, he’s speaking from operating at real scale — not theory. As VP of Engineering Platforms and Security Research at Forcepoint, he led a transformation that cut cloud spend in half while improving availability, and built a culture where engineers own their economics.

Tags have long been the backbone of cloud cost visibility and governance. They help teams understand who owns what, where spend comes from, and how infrastructure maps back to the value the business delivers. However, AI workloads have altered that model, and exposed the limitations of traditional AI tags in the process. In fact, many of the most expensive AI operations don’t run on taggable cloud resources at all.

You may have heard that keeping an existing customer is five times cheaper than acquiring a new one. But that isn’t always true. “Hidden costs” often accompany customer retention, loyalty, and increasing “share of customer”. Could you be spending more on customer retention than on winning new customers? This quick guide will walk you through the meaning of Customer Retention Cost (CRC), why it’s important to calculate it, and how to calculate it.

The inevitable happened. A while back, Gartner projected that in 2026, 30–50% of all new SaaS product features would use LLM inference. That meant OpenAI-style costs would become a standard part of SaaS COGS. Today, OpenAI has become one of the most operationally significant line items for SaaS companies. But for many teams, this creates an uncomfortable gap. Engineering sees OpenAI as a fast path to innovation.

In 2025, Oracle shocked the market. Its cloud growth was so aggressive that Oracle’s stock surged, briefly making founder Larry Ellison the world’s richest person. That didn’t happen by accident. Oracle closed fiscal 2025 with $57.4 billion in revenue, mainly driven by cloud services. Oracle Cloud Infrastructure (OCI) grew roughly 50% year over year, driven by enterprise databases, AI workloads, and network-intensive applications migrating from more expensive platforms.

Right now, 48% of organizations say they’re being asked to measure or report on AI-related costs. The problem is that they’re still figuring out how to do it. That was a very telling stat from a recent CloudZero webinar on AI and profitability, and speaks loudly to the reality that many organizations are still struggling to get a grasp on AI spend which our data shows to be rising sharply as a part of total spend in recent months.

Think of a stateful application like a conversation with a barista who remembers your order every time you walk in. They know what you had yesterday, how you like it prepared, and what you’ll probably want next. That memory makes the experience smoother, but it also means that if that barista isn’t around, your experience can break down entirely. A stateless application, on the other hand, is similar to ordering from a self-service kiosk.