Operations | Monitoring | ITSM | DevOps | Cloud

Welcome to December’s edition of CloudZero’s Cloud Economics Pulse — your monthly read on how cloud spend is shifting across providers, services, and AI workloads. No surprises here — November continued the quiet reshaping trend we’ve seen all year. Compute softened, data layers grew, and AI/ML hit its highest share yet. AWS extended its lead, Azure and GCP nudged upward, and the emerging “AI layer” of providers continued to take shape.

A few months back, a backend team at a fast-growing SaaS company shipped what seemed like a harmless feature. Just a simple request validation layer. No new service. No major dependencies. No architectural shock. Yet two months later, their cloud costs had climbed 38% without any significant increase in traffic, storage, or compute load. What they’d missed was that the validation layer triggered a fan-out pattern.

New Relic provides full-stack observability and monitoring. It provides almost every type of system monitoring on a single platform. This includes monitoring tools for infrastructure, application performance monitoring (APM), synthetics, user, log, mobile, network, and Kubernetes components. DevOps, security, and business professionals use these capabilities to detect anomalies, analyze root causes, and fix software performance issues.

AI adoption is exploding, but margins aren’t. In fact, an MIT analysis reports that 95% of organizations have yet to see measurable ROI from GenAI. This gap becomes obvious as soon as teams push a model into production and usage begins to scale. For most workloads, the pressure comes after training. Every message, call, query, completion, or retrieval triggers compute behind the scenes. That real-time execution is what AI inference is all about.

Machine learning pipelines are getting heavier by the day. From model training to large-scale inference and data preprocessing, compute demands are scaling faster than teams can manage. Kubernetes clusters groan under unpredictable job spikes. Static infrastructure wastes money when workloads slow down. The result? Organizations are perpetually chasing flexibility, automation, and cost efficiency. AWS has quietly built a solution to establish that balance.

Ask a SaaS team how they measure cloud efficiency, and you’ll hear familiar things. Total cloud spend. Average cost per customer. Maybe a breakdown of spend by service. All useful, but rather wobbly. Now ask, “What does it cost you to serve one more customer?” That’s when the room goes quiet. And that’s often where cloud economics gets really wobbly. Because that number, your marginal cost, is what actually determines your margins. Not your total cloud bill.

The AI landscape today feels a lot like the early days of the cloud: exciting, fast-moving, and completely fragmented. Every week, engineering teams are experimenting with dozens of large language models (LLMs) from providers like OpenAI, Anthropic, Google, Mistral, Meta, and beyond. They’re tweaking prompts, testing model performance, swapping context windows, and even running multiple models in parallel to figure out which one works best for each unique use case.

Happy Holidays, and happy almost new year from us at CloudZero! November is always a busy time! In case you missed the headlines, here’s what happened last month in relation to cloud and AI costs: Get the details below!

One year ago, at AWS re:Invent, we launched CloudZero Advisor, a free, standalone AI assistant that enables anyone to ask questions about cloud spend in plain language. It was the first experiment of its kind in FinOps, a chance to see what people really wanted to know when cost data finally became conversational. Over the past year, Advisor has become a learning engine.

FinOps thrives on clarity, and clarity is built on metrics. Metrics give engineering and finance a shared language to understand costs, evaluate trade-offs, and guide innovation. The most impactful metrics go beyond “how much are we spending?” and help us answer: When we measure these things, we stretch beyond tracking progress to fueling it.