The Savings Zesty Can’t Unlock for ML/Biotech/Variable Workloads

Streamlined visibility, cost allocation, reports & budgets for your Snowflake spending

The Savings Zesty Can’t Unlock for ML/Biotech/Variable Workloads

Over two-thirds of Vertexis’s AWS compute spend remained On-Demand, even after leveraging Zesty Reserved Instances (RIs)—leaving thousands of dollars on the table annually.

Zesty had helped cover the predictable baseline usage. But Vertexis’s high-performance research workloads spiked unpredictably, sometimes for just a few hours or several days before being turned off. And when those spikes hit, traditional approaches to commitment management couldn’t cover them with discounts. And when Vertexis asked for help optimizing further, they got little guidance for handling expiring commitments or improving savings.

Recognizing these challenges, Vertexis turned to nOps, who identified a previously overlooked savings opportunity by targeting these high-cost, short-term bursts. By strategically deploying hourly Convertible Reserved Instances (CRIs)—a technique uniquely optimized by nOps through hourly analysis and automated management—Vertexis was able to capture savings that traditional RI management tools missed.

Even in the first year, Vertexis saw their net savings tick up from 13.9% to 14.3%. But the real win comes in year two: $24,000 in additional projected savings, with no operational changes and no risk. If usage drops, nOps “squishes” the commitments. If usage changes, nOps adapts in real time — with a 100% utilization guarantee and automatic refund for any unused commitment.

Critically, nOps offered flexible arrangements without penalties or fees for breaking contracts, removing the risk associated with typical RI commitments. This clear, low-risk financial optimization ensures that Vertexis can focus on accelerating scientific breakthroughs rather than infrastructure costs.

*Note: the company name has been changed for privacy reasons

Demo