• VP Land
  • Posts
  • Project DIGITS, RTX 50, AI Ray Tracing: NVIDIA's CES Updates

Project DIGITS, RTX 50, AI Ray Tracing: NVIDIA's CES Updates

NVIDIA unveiled its next generation of AI and graphics technology, headlined by dramatic improvements in graphics processing, AI capabilities, and a surprise desktop AI supercomputer.

The announcements signal a major shift in how media professionals will create and process content in the near future.

Behind the Scenes

GeForce RTX 50 Series

The new Blackwell architecture brings unprecedented power to content creation and rendering. The RTX 5090 delivers twice the performance of the previous 4090, while the RTX 5070 matches the 4090's performance in video games at just $549.

Its 12GB of VRAM will limit it in AI workloads and professional applications such as Unreal Engine.

The entire lineup features:

  • 92 billion transistors

  • 4 petaflops of AI processing (3x previous generation)

  • 380 ray tracing teraflops

  • 125 shader teraflops with concurrent integer processing

  • G7 memory delivering 1.8 terabytes per second bandwidth

Neural Rendering Revolution

NVIDIA's new approach to ray tracing uses AI to dramatically reduce computational demands. The system:

  • Renders only 2 million essential pixels while AI predicts the remaining 33 million

  • Generates three additional frames for every computed frame

  • Introduces neural texture compression and neural material shading

  • Enables real-time ray tracing of every pixel, previously thought impossible

Project DIGITS

This compact AI supercomputer represents a significant shift in how creative professionals can access AI processing power.

Key features include:

  • Built on the new GB10 chip developed with MediaTek

  • Runs the entire NVIDIA AI software stack

  • Works as a cloud platform that sits on your desk

  • Available around May 2025

  • Can be paired for additional processing power

Limitations include:

  • Limited VRAM for training

  • VRAM Speed

  • Not enterprise ready - no management solution / wrong form factor

  • Unified Memory - OS and GPU share memory - Bad for 3d workloads and precaching

Final Take

For media professionals, these announcements represent a fundamental shift in content creation workflows. The combination of AI-assisted rendering, dramatically improved graphics processing, and desktop AI supercomputing means faster rendering times, more realistic real-time previews, and new creative possibilities previously limited by processing power.

The ability to generate and process content using AI at your desk, rather than relying on cloud services, could reshape how studios approach both creative development and final rendering.

As these technologies roll out through 2025, they'll likely become essential tools for staying competitive in the rapidly evolving media production landscape.

Reply

or to participate.