NVIDIA 900-23834-0080-000 | IGX T5000 AI Module, Blackwell GPU, 128GB LPDDR5X, 14-Core

Q: What applications is the NVIDIA IGX T5000 Module designed for?
A: The IGX T5000 Module is purpose-built for industrial, medical, and robotics edge AI applications requiring real-time sensor processing, multi-sensor fusion, and generative AI inference. It excels in autonomous mobile robots, collaborative robots, industrial inspection systems, medical imaging, and mission-critical physical AI agents that require functional safety compliance.

Q: How does the IGX T5000 Module differ from NVIDIA Jetson modules?
A: While the IGX T5000 is form-factor and pin-compatible with Jetson T5000, it is specifically designed for industrial-grade and mission-critical deployments with enhanced functional safety features, including a Functional Safety Island (FSI), dedicated safety MCU cores, and compliance with ISO 26262 and IEC 61508 standards. It also offers 10-year lifecycle support for enterprise applications.

Q: What are the memory and compute capabilities of the IGX T5000?
A: The module features 128GB of 256-bit LPDDR5X memory with 273GB/s bandwidth and delivers up to 2,070 FP4 TFLOPS of AI compute performance. This represents 7.5x higher AI compute than previous-generation platforms while maintaining 3.5x better energy efficiency, making it suitable for running multiple generative AI models at the edge.

Q: Can the IGX T5000 Module be integrated into custom carrier boards?
A: Yes, the IGX T5000 is a System-on-Module designed for flexible integration with custom carrier boards. It provides extensive I/O options including 4x 25 GbE networking, up to 32 camera inputs via MIPI CSI-2, PCIe Gen5, USB, and other interfaces, allowing developers to create tailored solutions for specific industrial and medical edge computing requirements.

Q: What AI frameworks and models does the IGX T5000 support?
A: The IGX T5000 supports modern AI frameworks through NVIDIA AI Enterprise software, NVIDIA Isaac for robotics, NVIDIA Holoscan for sensor processing, and NVIDIA Metropolis for visual AI. The Blackwell architecture GPU includes Transformer Engine and Multi-Instance GPU (MIG) support, enabling efficient execution of large language models, vision-language models, and other generative AI workloads with multiple precision formats including FP4, FP8, and FP16.