I am currently working on an AI-based video analytics solution and using the latest RTX 5080 and RTX 5090 GPUs for real-time multi-stream inference workloads. We are building our pipeline on the NVIDIA DeepStream SDK, but we have encountered compatibility issues since the current release does not appear to support the RTX 50 series (Blackwell architecture) GPUs.
Could you please advise on the expected release date or roadmap for DeepStream SDK with official support for RTX 50 series cards? Knowing this timeline would help us plan our deployments and decide whether to continue with workarounds (such as Triton Inference Server) or wait for a stable DeepStream update.