GreenOps & FinOps – WisdomAI: Why zombie data is the hidden cost of cloud

Data waste, at source

Most organisations have three or four copies of every meaningful dataset. An original system of record. A derivative copy created through ETL for analytics or reporting. One or more test or experimental versions. A production copy feeding dashboards, models or downstream applications. Each copy consumes storage, compute and operational effort.

Data without a time value keeps piling up… so the direct ongoing effect on data storage, datacenter capacity requirements and the logical (negative) impact that has on any strategy designed to achieve sustainable GreenOps should be fairly obvious. Much of this data has limited ‘time value’ and so those information streams should be strategic targets for GreenOps wellbeing.

A dataset may be critical during a product launch, a forecasting cycle or an AI experiment. Once that window closes, its value drops. Yet the data persists. Storage feels cheap. Compute feels elastic. So the data stays – rarely accessed, rarely validated, almost never deleted… that’s not good for global emissions and (insert need to save species of rainforest animal of your choice) and so on. 

This isn’t a failure of engineering discipline. Engineers focus on the immediate problem: moving data, transforming it, connecting systems. Nobody rewards garbage collection. In the cloud, creating resources is easy. There are few incentives to clean them up.

Elasticity without limits creates waste

I saw something different in quota-driven environments. At Google, teams got explicit quotas for storage and compute. Those quotas could be large, but they existed. Scarcity forced prioritisation. If a dataset or pipeline no longer mattered, it had to justify its continued existence. This produced healthier systems with fewer forgotten assets.

Public cloud environments prioritise elasticity but rarely impose limits. Resources expand automatically. Nothing pushes back when usage drifts or multiplies. You end up with idle storage, unused pipelines and zombie compute, racking up cost and carbon impact.

So now we’re at a point where manual accountability no longer scales.

WisdomAI CEO Mazumdar: GreenOps can now stop being an abstract idea and become operational.

The instinctive response is to demand more accountability from engineers. That doesn’t work anymore. In the age of AI, accountability moves in the opposite direction. Teams experiment, prototype and connect data to new models as fast as possible. Temporary pipelines and datasets proliferate. Manual processes can’t keep up.

What you need is automation that tracks liveness. Systems that identify datasets not accessed in months, pipelines that no longer produce outputs, compute services that receive no traffic. These signals should trigger action: archiving, tiering to cold storage or elimination.

Consumption platforms enter FinOps

This issue goes beyond infrastructure services. Consumption-based platforms like Snowflake and Databricks make it worse. Zombie data and zombie operations don’t just sit idle. They generate cost every time a forgotten query runs or a scheduled job executes. These platforms belong in the FinOps and GreenOps conversation, not siloed off as a separate concern.

Cost and carbon are now inseparable, yes, this is a truism we need to say out loud. Every unnecessary copy of data, every idle workload, every forgotten pipeline carries both a financial and environmental cost. You can’t optimise one without addressing the other.

Our mission now is to move from aspiration to operational reality. The path forward isn’t austerity. It’s visibility, liveness tracking and automated discipline built into data systems. When you understand which data is alive, which is dormant and which is genuinely needed, you reduce waste, lower environmental impact and create a healthier foundation for analytics and AI.

That’s where GreenOps stops being an abstract idea and becomes operational.