Cloud Costs Are Getting Smarter. Are We?
Here’s a conversation I keep hearing about: An engineering leader walks into a meeting with finance. They've just launched a new AI-powered feature, usage is spiking, and cloud bills are doing the same. The finance team wants to know how those costs relate to revenue. The engineering team shrugs. The conversation goes nowhere.
This isn’t a one-off. It’s a pattern—and one I’ve seen across companies big and small. As cloud ecosystems become more dynamic and intelligent, our approach to cost management needs to evolve just as fast. That starts with shifting the conversation from "How much are we spending?" to "What are we spending on—and what are we getting in return?"
That’s where COGS comes in.
COGS Is the Future of Cloud Cost Management
Let’s level-set: Cost of Goods Sold (COGS) is the direct cost of delivering your product or service to customers. In a SaaS company, cloud infrastructure is very often one of those direct costs. If your product lives in the cloud, and customers can’t use it without compute, storage, APIs, or inference calls to your LLM—then that spend belongs in COGS.
But here’s the thing: most companies are still treating cloud like generic OpEx. And that’s a mistake. Why? Because cloud is no longer a shared utility that can’t be allocated—it’s a critical part of product infrastructure. If we don’t align cloud spend with the value it delivers, we’re flying blind.
Complexity Is the New Normal
We’ve seen firsthand how fast architectures are changing. Companies are moving from monoliths to microservices, from long-lived VMs to event-driven functions. With that shift comes fragmentation: more teams, more services, more vendors, more noise.
Add to that the rise of serverless, ML workloads, and agentic AI—and suddenly you’re dealing with infrastructure that scales dynamically, across multiple clouds, with pricing models that shift by the minute. Traditional cost tracking simply can’t keep up.
We’re entering an era where compute costs aren’t just an engineering concern—they’re a business driver. If we want to stay ahead, we need a way to assign those costs clearly and reliably to the services—and the customers—that generate revenue.
That’s where COGS becomes more than a finance metric. It becomes a lens.
The Real Innovation? Accountability
Let’s talk about the shiny stuff for a minute. There’s a lot of excitement right now around AI ops, predictive scaling, and automated budget enforcement. And those are all great tools—but they don’t mean much without a foundation of cost accountability.
And cost accountability shouldn’t start with engineers being told to keep the cost low, optimize and shrink it. It should start with a focused understanding of what belongs in COGS and how much revenue it enables.
Here’s how we see the landscape changing:
1. Cloud architectures are getting more granular.
Teams are spinning up services on demand. Infrastructure is short-lived and usage patterns are bursty. The old model of fixed cost centers doesn’t work. We need flexible attribution models that match the speed of deployment.
2. Service models are getting more abstract.
You’re not managing infrastructure anymore—you’re consuming it. From platform-as-a-service to usage-based SaaS, costs are hidden behind APIs and bundled pricing. That makes visibility harder, but also more important.
3. The AI wave is changing cost profiles.
Inference, training, model hosting—these aren’t side projects anymore. They’re customer-facing features. Which means their costs are no longer optional or experimental. They’re COGS.
What Needs to Happen Next
To keep up with all this change, we need to rethink our cloud cost playbook. And that starts with three shifts:
Shift 1: From OpEx to COGS
Don’t just track what you spend—track what it’s for. Tag and allocate costs to the specific services, features, and teams responsible. If a resource is directly involved in delivering your product, it belongs in COGS.
Shift 2: From Cost Centers to Business Context
Move beyond vague labels like “infra team” or “shared services.” Start mapping spend to business outcomes—revenue-generating workloads, critical customer features, SLAs. If it impacts the user experience, it needs a name and a number.
Shift 3: From Manual Reports to Real-Time Insight
Finance shouldn’t need to wait until the end of the month to understand cloud margins. Engineers shouldn’t be surprised by cost anomalies. We need cost observability built into workflows—not bolted on after the fact.
What We're Watching at Yotascale
We're keeping a close eye on a few important developments that will shape the next 12–18 months:
- FOCUS standardization. The FinOps Foundation’s push toward standardized cost and usage formats will make cross-vendor reporting more consistent—and COGS classification more defensible.
- Cloud-native telemetry improvements. Hyperscalers are starting to offer more granular cost data, which will improve automation and accuracy in cost attribution.
- AI-native product architectures. As more teams embed LLMs and agentic AI into their products, we expect a wave of innovation around how to track and allocate inference costs to the right place in the P&L.
Why This Matters for the Whole Business
When companies start classifying cloud infrastructure as COGS, a core shift takes place. Finance gets better visibility into gross margins. Engineering gets clearer goals around cost-efficiency. Product teams gain a more accurate view of unit economics.
Most importantly, it aligns the business around a shared understanding of value. We stop arguing over who owns the AWS bill—and start asking smarter questions:
- How does this new feature impact margins?
- Are our infrastructure investments scaling with customer growth?
- What’s the real ROI of that AI integration?
Final Thought: COGS Isn’t Just Accounting—It’s Strategy
This isn’t just about better accounting. It’s about building a more connected, responsive organization—one that can move fast without losing financial clarity.
In the future, cloud cost management won’t be a back-office function. It’ll be part of how we build, price, and grow great products. COGS gives us the structure to make that happen.
Let’s stop reacting to cloud bills and start designing around them.
Let’s treat cloud spend not as overhead—but as investment.
Because in a world where your product is digital, understanding your COGS might just be your biggest competitive advantage.
