By Nayan Lad,
Santa Clara, California, April 15, 2025 — Exostellar, a leader in AI-Driven Autonomous Compute Orchestration, today announced that DataRobot has selected Exostellar’s Infrastructure Optimizer following a successful proof of value (POV) that demonstrated the potential for over 40% reduction in EC2 costs through intelligent spot instance utilization for stateful workloads.
DataRobot delivers an Enterprise AI Suite that easily integrates into core business processes so customers can develop, deliver, and govern AI agents and applications at scale.
During the multi-week POV, Exostellar’s AI-powered optimization platform achieved 94%+ spot instance utilization while maintaining workload stability, validating its ability to handle stateful services with no disruptions. These results confirmed the platform’s potential to drive significant cloud savings without requiring architectural changes or operational compromises.
“Based on the outcomes of the POV, we saw clear potential to reduce our EC2 spend using Exostellar’s infrastructure optimizer,” said Amit Purohit, SVP of Engineering at DataRobot. “This aligns with our long-term strategy to optimize cloud efficiency at scale.”
“We’re excited to welcome DataRobot to the Exostellar platform,” said Tony Shakib, CEO of Exostellar. “This decision reinforces the value of our AI-driven infrastructure automation in helping modern AI companies scale with confidence and efficiency.”
Exostellar’s Infrastructure Optimizer uses live migration and real-time policy engines to dynamically shift workloads across Spot, On-Demand, and Reserved instances—maximizing savings and maintaining high reliability. The product requires no changes to existing applications.
Coming Soon: Workload Optimizer — Predictive Rightsizing Without Downtime
In addition to Infrastructure Optimizer, Exostellar is preparing to launch Workload Optimizer, a new solution that dynamically resizes Kubernetes and HPC workloads during runtime—without restarts or downtime.
By combining predictive scaling, live migration, and real-time CPU/memory monitoring, Workload Optimizer prevents overprovisioning and out-of-memory failures. Early trials have shown 50–80% cost savings and 2–3x improvements in utilization and throughput for long-running, memory-intensive jobs.
With both Infrastructure and Workload Optimizers, Exostellar is setting a new standard for intelligent, cost-aware cloud operations.
Schedule demo: https://exostellar.io/getting-started/