Cast AI, the leading Kubernetes automation platform, today published its third annual Kubernetes Cost Benchmark Report, which highlights ongoing inefficiencies in cloud resource utilization despite the growing maturity of the cloud-native ecosystem. The report also features an analysis of GPU availability and pricing.
Many DevOps teams still manually manage infrastructure for Kubernetes applications, leading to significant overprovisioning and cloud waste. According to the report, average CPU utilization across Kubernetes clusters remains low at just 10%, down from 13% last year. Memory utilization is 23% – a modest 3% increase from the previous year.
Cast AI’s research identifies a clear opportunity for cost savings through the use of Spot Instances. For example, clusters that partially leverage Spot Instances achieve an average compute cost reduction of 59%, while those running exclusively on Spot Instances see an even greater cost reduction of 77%.
This year’s Kubernetes Cost Benchmark Report also features an analysis of GPU availability and pricing. According to the report, GPU availability varies by cloud provider. Cast AI analyzed different regions and Availability Zones to find where specific GPU chips are most available and compared the cost of running workloads on some of the hardest-to-get GPUs. The key finding is that companies that are able to move workloads to more cost-effective locations can cut costs substantially:
- From 2x to 7x savings compared to the average Spot Instance price worldwide
- From 3x to 10x savings compared to the average On-Demand Instance price
“Our report spotlights that overprovisioning is still a key issue for teams manually managing Kubernetes clusters in public clouds. While not unexpected, it is still frustrating to see because it is completely avoidable using automation,” said Laurent Gil, president and co-founder of Cast AI. “When we onboard new customers, they have two ah-ha moments. The first is when they enable automation and see immediate cost savings via workload rightsizing, bin-packing, and instance-type selection. The second is when they realize automation isn’t just saving them money, it’s freeing them up to uplevel their creative thinking and spend more time solving mission-critical business problems.”
About the 2025 Kubernetes Cost Benchmark Report
To view other insights from the report, please visit Kubernetes Cost Benchmark Report. The report is based on Cast AI’s analysis of 2,100+ organizations across Amazon Web Services (AWS), Google Cloud Platform (GCP), and Microsoft Azure (Azure) between January 1 and December 31, 2024. This analysis excludes clusters with fewer than 50 CPUs and focuses on data collected before these organizations used Cast AI’s automation.
