The auto-updated technology brief
ModelExpress by NVIDIA addresses the high cost of moving large model checkpoints, enabling faster distribution of model weights across clusters for tasks like cold starts, autoscaling, and rolling updates.
NVIDIA OptiX is a GPU ray tracing engine framework. The NVIDIA OptiX Toolkit (OTK) offers debugging tools to diagnose issues like invalid API arguments, black frames, or GPU-side bugs in ray tracing applications.
NVIDIA DeepStream 9.1 enables developers to build video analytics applications that track objects across multiple camera views using 3D tracking, overcoming limitations of single-camera 2D tracking.
NVIDIA's GB300 NVL72 achieved a world record in pre-training the DeepSeek-V3 671B model using mixture of experts (MoE), reaching 1,648 TFLOPs per GPU and advancing large-scale AI training efficiency.
Healthcare robotics faces unique challenges due to limited data and experimentation opportunities. NVIDIA explores GPU-native medical physics simulation to advance robotic surgery development.
Agentic AI transforms AI factory infrastructure by triggering multiple model calls and data operations per request, requiring fast data movement and protection. NVIDIA BlueField enables scaling through extreme co-design.
NVIDIA's Nemotron 3 Ultra model excels in accuracy and efficiency for agentic register transfer level (RTL) coding, accelerating chip design by leveraging AI agents and LLMs to improve code generation and verification.
NVIDIA introduces the Vera CPU featuring Olympus cores optimized for maximum single-thread performance to enhance agentic AI workloads, improving per-agent responsiveness and AI factory throughput.
Increasing AI workloads are driving explosive compute demand, pushing the semiconductor industry to meet unprecedented performance targets while addressing thermal and power challenges through system-level engineering breakthroughs.
This tutorial explains how to deploy a validated AI coding assistant in regulated or source-sensitive environments using NVIDIA NeMo Guardrails, addressing challenges like network source restrictions, supply-chain risks, and audit trails.
The NVIDIA Nemotron Model Reasoning Challenge engaged over 5,000 Kaggle participants to explore techniques for improving AI reasoning accuracy starting from the same open model and evaluation constraints.
NVIDIA explores co-designing AI model attention mechanisms optimized for GPUs to improve inference performance in agentic and long-context workloads where attention dominates computational cost.