CAST AI – a leading Kubernetes cost optimization platform – recently announced that it has closed a $35 million Series B round from Vintage Investment Partners and existing investors Creandum and Uncorrelated Ventures. The new funding round follows a $20 million investment round led by Creandum in March, bringing the company’s total funding to $73 million and underscoring its strong growth and increasing demand for its platform.
Running applications within the cloud is complex and expensive. And the costs will continue to rise as legacy applications are modernized and migrated to the cloud. New use cases emerge, such as companies that have started running and training AI models that demand specialized GPUs and require billions of computations per second and petabytes of data. For enterprises training and running their large language models, the resulting cloud bill can cost as much as $700,000 a day, as is the case for OpenAI’s infrastructure cost on Azure.
Even though some solutions offer basic capabilities that merely monitor clusters and make recommendations, CAST AI’s platform utilizes advanced Machine Learning Algorithms and Heuristics to automatically optimize clusters – saving customers 50% or more on their cloud spend, improving performance and reliability and boosting DevOps and engineering productivity.
At the KubeCon Chicago 2023 event, CAST AI unveiled two pivotal new platform capabilities – Automated Workload Rightsizing and PrecisionPack. These new features are not just standalone capabilities but significant drivers in a robust platform. When integrated with CAST AI’s Cost-Aware Autoscaler, Spot Instance Management, and the complete Cost Reporting Suite, it forms a powerful combination that drives cost-efficiency and performance optimization to the next level. The Automated Workload Rightsizing ensures that Kubernetes workloads are not just efficiently sized but are managed in a cost-aware manner, embodying a holistic approach to cost optimization and resource utilization.
Accurately forecasting the resources required for Kubernetes workloads is complex and has traditionally been a stumbling block for organizations that require a lot of human intervention. And the propensity to under or over-provision resources leads to operational risks and resource wastage. Workload Rightsizing overcomes the issue by automating the scaling of workload requests in near real-time, ensuring optimal performance while being cost-effective.
Resource allocations are regenerated in near real-time and offer granular control and configurability per workload. The flexibility extends to specifying additional overhead for CPU and RAM, adjusting percentile values and setting thresholds for applying this automatically. CAST AI’s product roadmap includes the introduction of seasonality models to better predict resource needs across varying time cycles, which increases response time and availability.
Along with Workload Rightsizing, CAST AI also introduced PrecisionPack – a next-generation Kubernetes scheduling approach that eradicates randomness in pod placement. And this employs a sophisticated bin-packing algorithm to ensure strategic pod positioning onto the designated set of nodes, maximizing resource utilization while increasing efficiency and predictability across Kubernetes clusters. The workload movement is reduced, which improves both uptime and reliability of workloads while maintaining a perfect blueprint for cluster cost optimization.
“AWS, Google and Azure bills can significantly impact gross margins for businesses, leading to excessive financial stress. CAST AI has proven that this scenario can be entirely avoided by consistently slashing customers’ cloud spend. What is unique about CAST AI is that it has developed a robust platform that goes beyond monitoring and recommendations; it automatically optimizes customers’ cloud resources, supercharging their savings.”
— Vintage Investment Partners’ Barrel Kfir
“Every single person at CAST AI is relentlessly focused on helping customers slash their cloud spend by automating tasks that are best performed by machine learning systems. That’s why customer growth continues to accelerate and we’ve recently welcomed marquee customers like Akamai and Yotpo. The new funding will further bolster customer savings and productivity as we expand our platform’s capabilities and automate even more aspects of Kubernetes.”
— CAST AI Co-Founder and CEO Yuri Frayman
“Our vision transcends individual features. We are on a mission to deliver a fully automated Kubernetes experience. The introductions of Workload Rightsizing and PrecisionPack is a significant stride towards that horizon. It’s not just about rightsizing; it’s about creating a synergistic ecosystem where every component – from cost reporting to instance management – works in unison to deliver unprecedented value.”
“The journey has just begun. The recent capital raise has supercharged our pace of innovation. As we continue to unfold platform capabilities that encapsulate our vision of a fully automated Kubernetes reality, the value proposition for our customers becomes significantly amplified.”
— Laurent Gil, co-founder and chief product officer of CAST AI