NVIDIA Blackwell Powers Real-Time AI for Entertainment Workflows

AI has been shaping the media and entertainment industry for decades, from early recommendation engines to AI-driven editing and visual effects automation. Real-time AI - which lets companies actively drive content creation, personalize viewing experiences and rapidly deliver data insights - marks the next wave of that transformation.

With the NVIDIA RTX PRO Blackwell GPU series, announced yesterday at the NVIDIA GTC global AI conference, media companies can now harness real-time AI for media workflows with unprecedented speed, efficiency and creative potential.

NVIDIA Blackwell serves as the foundation of NVIDIA Media2, an initiative that enables real-time AI by bringing together NVIDIA technologies - including NVIDIA NIM microservices, NVIDIA AI Blueprints, accelerated computing platforms and generative AI software - to transform all aspects of production workflows and experiences, starting with content creation, streaming and live media.

Powering Intelligent Content Creation Accelerated computing enables AI-driven workflows to process massive datasets in real time, unlocking faster rendering, simulation and content generation.

NVIDIA RTX PRO Blackwell GPUs series include new features that enable unprecedented graphics and AI performance. The NVIDIA Streaming Multiprocessor offers up to 1.5x faster throughput over the NVIDIA Ada generation, and new neural shaders that integrate AI inside of programmable shaders for advanced content creation.

Fourth-generation RT Cores deliver up to 2x the performance of the previous generation, enabling the creation of massive photoreal and physically accurate animated scenes. Fifth-generation Tensor Cores deliver up to 4,000 AI trillion operations per second and add support for FP4 precision. And up to 96GB of GDDR7 memory boosts GPU bandwidth and capacity, allowing applications to run faster and work with larger, more complex datasets for massive 3D and AI projects, large-scale virtual-reality environments and more.

Elio Disney/Pixar One of the most exciting aspects of new technology is how it empowers our artists with tools to enhance their creative workflows, said Steve May, chief technology officer of Pixar Animation Studios. With Pixar's next-generation renderer, RenderMan XPU - optimized for the NVIDIA Blackwell platform - 99% of Pixar shots can now fit within the 96GB of memory on the NVIDIA RTX PRO 6000 Blackwell GPUs. This breakthrough will fundamentally improve the way we make movies.

Lucasfilm Ltd. Our artists were frequently maxing out our 48GB cards with ILM StageCraft environments and having to battle performance issues on set for 6K and 8K real-time renders, said Stephen Hill, principal rendering engineer at Lucasfilm. The new NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition GPU lifts these limitations - we're seeing upwards of a 2.5x performance increase over our current production GPUs, and with 96GB of VRAM we now have twice as much memory to play with.

In addition, neural rendering with NVIDIA RTX Kit brings cinematic-quality ray tracing and AI-enhanced graphics to real-time engines, elevating visual fidelity in film, TV and interactive media. Including neural texture compression, neural shaders, RTX Global Illumination and Mega Geometry, RTX Kit is a suite of neural rendering technologies that enhance graphics for games, animation, virtual production scenes and immersive experiences.

Fueling the Future of Streaming and Data Analytics

Data analytics is transforming raw audience insights into actionable intelligence faster than ever. NVIDIA accelerated computing and AI-powered frameworks enable studios to analyze viewer behavior, predict engagement patterns and optimize content in real time, driving hyper-personalized experiences and smarter creative decisions.

With the new GPUs, users can achieve real-time ingestion and data transformation with GPU-accelerated data loading and cleansing at scale.

The NVIDIA technologies accelerating streaming and data analytics include a suite of NVIDIA CUDA-X data processing libraries that enable immediate insights from continuous data streams and reduce latency, such as:

NVIDIA cuML: Enables GPU-accelerated training and inference for recommendation models using scikit-learn algorithms, providing real-time personalization capabilities and up-to-date relevant content recommendations that boost viewer engagement while reducing churn.

NVIDIA cuDF: Offers pandas DataFrame operations on GPUs, enabling faster and more efficient NVIDIA-accelerated extract, transform and load operations and analytics. cuDF helps optimize content delivery by analyzing user data to predict demand and adjust content distribution in real time, improving overall user experiences.

Along with cuML and cuDF, accelerated data science libraries provide seamless integration with the open-source Dask library for multi-GPU or multi-node clusters. NVIDIA RTX Blackwell PRO GPUs' large GPU memory can further assist with handling massive datasets and spikes in usage without sacrificing performance.

And, the video search and summarization blueprint integrates vision language models and large language models and provides cloud-native building blocks to build video analytics, search and summarization applications.

Breathing Life Into Live Media With NVIDIA RTX PRO Blackwell GPUs, broadcasters can achieve higher performance than ever in high-resolution video processing, real-time augmented reality and AI-driven content production and video analytics.

New features include:

Ninth-Generation NVIDIA NVENC: Adds support for 4:2:2 encoding, accelerating video encoding speed and improving quality for broadcast and live media applications while reducing costs of storing uncompressed video.

Sixth-Generation NVIDIA NVDEC: Provides up to double H.264 decoding throughput and offers support for 4:2:2 H.264

LINK:	https://blogs.nvidia.com/blog/media2-rtx-pro-blackwell/...
	See more stories from nvidia

More from Nvidia