CAST AI

When a DevOps engineer looks at their company's cloud bill, they often find a bleak landscape: oversized Kubernetes clusters, instances running at thirty or forty percent of actual utilization, resources allocated for peaks that will never materialize. According to the Kubernetes Cost Benchmark Report published by CAST AI in 2025 — based on analysis of over 2,100 organizations across AWS, Google Cloud, and Azure — applications use on average only 10% of CPU and 23% of memory actually allocated to them. The gap between paid resources and used resources is a structural waste worth billions of dollars a year, which no manual cost review can systematically resolve. It is on this fracture that CAST AI built its business. Founded in 2019 in Estonia by Yuri Frayman, Laurent Gil, Leon Kuperman, and Domantas Lubys — with operational headquarters subsequently moved to Miami — the startup developed an automatic Kubernetes workload optimization platform that promises cloud cost reductions between 50 and 80%, without requiring changes to application code and without manual intervention from engineering teams. The mechanism is elegant in its simplicity: the platform continuously analyzes actual resource utilization, selects the most cost-effective cloud instances for each workload type, dynamically scales capacity based on actual demand, and systematically leverages spot and preemptible instances — the discounted capacity made available by cloud providers when they have unused headroom — to cut the bill without compromising performance. All automatically, transparently, and reversibly. Commercial traction is significant. CAST AI's customer list includes Banking Circle, Yotpo, NielsenIQ, and Akamai — names that do not entrust their infrastructure to a technology unless it has proven reliability in critical production environments. The pricing model, based on a percentage of savings actually generated, perfectly aligns the startup's incentives with those of its customers: CAST AI earns only when the customer saves. The financial trajectory tells a story of rapid growth. After a $20 million Series A in March 2023 and a $35 million Series B in November of the same year, in April 2025 the startup closed a $108 million round — bringing total funding raised beyond $163 million — at a valuation approaching one billion dollars, on the cusp of unicorn status. The round was raised with the objective of accelerating expansion in US and European markets and enhancing the platform's capabilities for AI workload optimization, which requires enormously more expensive GPU resources and where the optimization margin is even greater. Particularly noteworthy is the expansion of scope beyond pure cost optimization: CAST AI has integrated security features into the platform — container vulnerability scanning, access management, policy enforcement — transforming itself from a FinOps tool into a holistic Kubernetes cluster management platform. This broadened positioning increases product stickiness: a company that entrusts CAST AI with not only the costs but also the security of its clusters is unlikely to switch vendors. At a historic moment when corporate cloud spending continues to grow — driven by AI workloads and the proliferation of microservices — CAST AI finds itself riding two converging structural trends: the need to contain infrastructure costs and the growing complexity of cloud-native systems. A combination that, in the FinOps and Kubernetes optimization market, is already worth billions.