Milpitas, CA, Aug. 12, 2025 — ProphetStor Data Services, Inc. today announced that the United States Patent and Trademark Office has granted its patent for a breakthrough in predictive, self-driving autoscaling that enables organizations to accelerate AI adoption, boost workload throughput, and deliver consistent performance across Kubernetes, VMware, and cloud environments.

Previously published as US 2023/0267006 A1, the newly granted patent covers a forward-looking scaling method that plans resource decisions before demand changes occur. This approach allows IT and AI operations teams to maintain stable service levels while controlling infrastructure costs and maximizing resource utilization.

Solving the Limits of Reactive Autoscaling

Traditional autoscaling reacts after demand changes, often leading to overprovisioning, underperformance, or sudden cost spikes. ProphetStor’s patented method takes a proactive, predictive approach, using short-term workload forecasts to plan capacity several steps ahead.

The system evaluates both the operational cost of running resources and the overhead cost of changing them, then selects the most efficient, stable scaling path. The result is a clear execution plan for pods, GPUs, VMs, or containers that integrates seamlessly with Kubernetes Horizontal Pod Autoscaler (HPA), Vertical Pod Autoscaler (VPA), VMware vSphere, and public cloud autoscaling groups.

“Enterprises want infrastructure that doesn’t just react to the past but anticipates what’s ahead,” said Eric Chen, CEO of ProphetStor. “This patent enables a self-driving feature with forward-looking optimization, scanning the road ahead, predicting changes, and adjusting workloads in real time so AI applications run faster and more reliably.”

Why This Matters for AI and Cloud Workloads

• Accelerated AI adoption through faster deployment of inference and training clusters
• Higher workload throughput by keeping compute resources optimally allocated
• Predictable performance and costs for easier planning and budgeting
• Seamless integration with existing Kubernetes, VMware, and cloud operations pipelines

Part of a Larger Autonomous Infrastructure Vision

This patent is also an important building block of ADDC.ai, ProphetStor’s vision for the AI-powered data center. In this model, the Federator.ai DataCenter OS serves as the intelligent control layer, virtualizing multi-layer platforms and devices, providing comprehensive visibility into operations, and applying AI-driven optimization to improve energy efficiency, cooling, and workload throughput.

Through the AboveCloud Platform, data centers can also participate in global compute trading, dynamically sharing and monetizing resources to improve availability and profitability.