• Home
• Technology

F5 (NASDAQ: FFIV), a global leader in application delivery and API security, has announced the expansion of its ongoing collaboration with NVIDIA to enhance the performance and efficiency of infrastructure powering real-world AI model deployments.

The expanded integration combines F5 BIG-IP Next for Kubernetes with NVIDIA BlueField-3 data processing units (DPUs), creating an intelligent infrastructure layer driven by real-time operational data insights. This approach is designed to significantly improve AI inference performance, optimize GPU utilization, reduce latency, and enable secure, scalable multi-tenant AI environments.

Within AI systems, “tokens” serve as the fundamental unit of output—whether words, signals, or data fragments—generated and processed during inference. The volume and speed of token generation are critical factors influencing user experience, infrastructure efficiency, and overall return on accelerated computing resources.

By leveraging this integration, organizations can better manage AI workloads at scale, ensuring faster processing times and more efficient resource allocation, while maintaining high levels of security and operational reliability.