Rafay Systems Accelerates Enterprise AI with NVIDIA BlueField‑3 DPUs and NVIDIA RTX PRO Servers
SUNNYVALE, Calif.–(BUSINESS WIRE)–Rafay Systems today announced integrations with NVIDIA BlueField‑3 DPUs and NVIDIA RTX PRO Servers featuring NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. Together with Rafay’s platform, enterprises and NVIDIA Cloud Partners (NCPs)get a practical, business‑ready path to stand up and operate AI services with stronger security, predictable performance, and faster time‑to‑value—aligned with the NVIDIA AI Enterprise software platform, including NVIDIA NIM microservices.
With this latest set of integrations, Rafay customers can now utilize NVIDIA RTX PRO Servers as part of their AI Factories, and deploy workloads such as F5’s BIG‑IP Next for Kubernetes offering on NVIDIA BlueField DPUs to deliver networking, security, traffic management, and load balancing—reducing the operational complexity of AI deployments. This RTX PRO 6000 integration builds on Rafay’s work with the NVIDIA DOCA Platform Framework (DPF), which streamlines the deployment, orchestration, and scaling of BlueField‑accelerated infrastructure.
Haseeb Budhani, CEO of Rafay Systems, said:
Customers want AI outcomes quickly—without compromising security or control,
“By pairing Rafay with NVIDIA BlueField‑3 and RTX PRO Servers, enterprises can launch and scale AI services on their terms.”
Rafay Integrates BlueField‑3 & DOCA Platform Framework
With Rafay’s DPF integration, teams can stand up the full DPU hardware and software stack in a single step, abstract infrastructure complexity with reusable templates, maintain GitOps‑driven consistency from staging to production, and apply automated lifecycle and policy governance across every DPU node—then scale DPU‑enabled workloads across clouds and edges with minimal effort. In addition, DPF and the DOCA software framework bring prebuilt microservices for software‑defined hardware- accelerated networking, distributed routing, real‑time threat detection, and service orchestration, offloading data‑plane work to the DPU and improving overall GPU utilization.
Host‑trusted security with NVIDIA RTX PRO Servers
Rafay’s integration takes advantage of the RTX PRO 6000 Server Edition’s data‑center security capabilities—including secure boot with a hardware root of trust, confidential compute support, and MIG‑based GPU partitioning—to enforce host‑trusted execution and strong workload isolation. Policies and templates in Rafay help apply these controls consistently across fleets, ensuring trusted GPU firmware at boot and segmenting resources per tenant and workload.
The Outcomes For Enterprises and NCPs
- Security & isolation by design (BlueField‑3 + DOCA). BlueField‑3 accelerates networking, data access, and AI runtime security for secure multi‑tenancy and zero‑trust isolation, with broad adoption across production AI infrastructure and validation on RTX PRO Servers.
- Universal GPU platform (RTX PRO Servers). Configure up to eight RTX PRO 6000 Server Edition GPUs per node—each with 96GB GDDR7 and MIG (up to 4× 24GB)—to power generative/agentic AI, visual computing, and scientific workloads.
- Business‑ready consumption model. Rafay enables self‑service access, enterprise controls (SSO/RBAC, quotas, audit/chargeback), and environment management for NVIDIA AI Enterprise and NIM to speed the journey from pilot to production.
Looking Ahead
Rafay is continuing development to support the new NVIDIA BlueField‑4 DPU upon availability. BlueField‑4 is designed to deliver 6× the compute of BlueField‑3, 800 Gb/s networking, and the BlueField Airgapped Resource Controller (ARC) for fabric‑level security and manageability—while maintaining full backward compatibility.
READ the latest news shaping the Nvidia market at Newsvidia
Rafay Systems Accelerates Enterprise AI with NVIDIA BlueField‑3 DPUs and NVIDIA RTX PRO Servers, source
