The auto-updated technology brief
NVIDIA introduces AI factories powered by accelerated computing to enhance training, fine-tuning, and deployment of autonomous AI agents, featuring new in-silicon security with DOCA.
AI factories require new data-center infrastructure to handle power-dense AI workloads and deliver predictable performance. This article discusses designing battery energy storage systems tailored for AI factory needs.
NVIDIA's Blackwell GPU sets a new STAC-AI benchmark record for large language model inference in financial trading, enabling advanced analysis of unstructured data to improve trading insights.
Step 3.7 Flash from StepFun delivers enterprise-scale multimodal AI capabilities on NVIDIA GPUs, enabling real-time perception, search, and reasoning across images, documents, video, and language.
NVIDIA introduces Dynamo Snapshot to address the cold-start delay in GPU inference workloads on Kubernetes, enabling faster startup and improved scalability during traffic spikes.
NVIDIA Cosmos 3 is a frontier foundation model enabling physical AI systems like robots and autonomous vehicles to understand, predict, and act in the real world.
NVIDIA introduces CUDA Tile programming to enable developers to create highly optimized GPU kernels within large existing C++ GPU codebases using tile-based programming techniques.
Training speech AI models to accurately recognize clinical terminology is challenging. NVIDIA's Nemotron Speech and agent skills help accelerate evaluation of clinical ASR models, improving recognition of complex drug names and medical terms.
AWS Transform – continuous modernization (preview) automatically scans code repositories to detect, prioritize, and remediate technical debt at scale, enabling continuous autonomous modernization and integration with security workflows.
NVIDIA JetPack 7.2 enhances real-world AI agent deployment on Jetson devices with optimized memory and performance, supporting one-command deployment of NemoClaw for privacy and security.
NVIDIA Quantum InfiniBand introduces intent-based security profiles in Unified Fabric Manager enabling one-click multi-tenant fabric security, supporting General, Bare Metal Cloud, and Secured Bare Metal Cloud profiles to reduce deployment time from hours or days to minutes.
NVIDIA introduces AI agents and FLARE Auto-FL to accelerate federated learning research by automating experiment design and evaluation, addressing challenges in metric improvement assessment.