Operations | Monitoring | ITSM | DevOps | Cloud

Running multiple LLMs without aggregation can feel like managing five different clouds with no dashboard. Sure, you can make it work, but you won’t like the bill. And most SaaS teams didn’t start with a multi-LLM strategy. It just happened. You added one model for reasoning, another for summarization, or maybe a fine-tuned version for customer support. Fast-forward six months, and your AI stack looks like a tangle of APIs. And each charges tokens on its own terms.

KubeCon Atlanta 2025 felt different this year — and CloudZero had a full team on the ground to capture it. Engineers, product leaders, sales reps, and CTO Erik Peterson spent three days embedded across the show floor. Their vantage points were complementary: the outbound conversations, the inbound questions, the demos, the technical deep-dives, and the quieter moments between sessions. Five perspectives stood out.

Traditionally, companies had to use separate tools for ETL, data storage, and analytics. Often, this resulted in slow, complex, and expensive data workflows A good example is PwC’s Deals, Insights & Analytics (DIA) team, which faced similar challenges. According to Microsoft, bespoke solutions often took months to build and were difficult to merge, slowing projects and driving up costs. That changed when PwC adopted Azure Synapse Analytics. The result?

If you’ve ever opened your Google Cloud bill and felt confused, you’re not alone. Costs in GCP can spike up fast. One project here, a few APIs there, until the total looks nothing like what you expected. That’s because Google Cloud pricing is built for flexibility. You pay for what you use, across dozens of services, each with its own rules, discounts, and data charges. It’s robust, but also easy to lose sight of what’s driving your spend. The good news?

Welcome to CloudZero’s inaugural Cloud Economics Pulse! This is our monthly snapshot of how cloud spend is evolving across providers, services, and emerging AI workloads. Each month, we’ll surface key trends, highlight where the money’s moving in cloud spend, and offer practical insight to help FinOps and cloud cost teams stay on top of things. October — and preceding months — shows a turbulent stretch for cloud economics.

For every CFO who’s ever asked, “Can we reduce engineering costs by 15%?” there’s a CTO quietly thinking, “But at what cost?” That tug-of-war plays out inside most SaaS companies. Finance wants to tighten budgets to protect margins, while engineering and product teams push for the resources to build, test, and innovate faster. The truth is, the teams that spend most efficiently — not necessarily the least — drive the highest returns.

Kubernetes is now the backbone of modern software infrastructure, helping teams deploy, scale, and manage applications efficiently across clouds. But when it comes to understanding costs, Kubernetes remains opaque. Teams often can’t answer basic questions like: How do you solve the gap between engineering usage and financial visibility? CloudZero’s new Kubernetes capabilities are built to address this challenge.

Data has exploded — and so have the challenges that come with it. Every click, transaction, and sensor ping generates mountains of data that traditional databases can’t handle. That’s why more than 94% of organizations now rely on cloud platforms, according to CloudZero’s 2025 cloud report. The goal isn’t just to store data, but rather, to make sense of it fast. And this is exactly where tools such as Google BigQuery step in.

As companies race to ship AI features into production, FinOps teams everywhere are scrambling to keep up with the variability of their engineers’ new favorite toys. The rapid pace of AI experimentation is creating a new wave of unpredictable usage patterns and costs.

Last month was certainly eventful. (If you use AWS or Azure, you probably already know what we’re talking about). But in case you missed it, here are the must-see headlines from October 2025.