Topaz Labs has announced its Precision Update, which introduces new AI models for image and video enhancement, including Gaia 2, Starlight Precise 2.5, and Wonder 2. The update features optimized support for AMD GPUs using Topaz NeuroStream technology to facilitate local processing.
This update signals a strategic shift toward specialized, local-first processing for high-compute tasks like GenAI video upscaling. By utilizing NeuroStream to optimize VRAM, Topaz is lowering the hardware barrier for professional-grade enhancement, directly challenging cloud-heavy competitors. For the streaming ecosystem, this local capability allows production houses to enhance AI-generated content or archival footage with greater security and zero per-gigapixel cloud costs. Watch for increased adoption of AMD hardware in creative workstations as specialized software-hardware partnerships improve local ROI for high-resolution 4K pipelines.
Topaz Labs first introduced NeuroStream in early March 2026 as a proprietary VRAM optimization layer that reduces VRAM usage by up to 95 percent, enabling large AI models to run on consumer-grade GPUs without sacrificing output quality. The technology was initially optimized for NVIDIA GeForce RTX and RTX PRO GPUs, but the March 31 Precision Update extends that capability to AMD Radeon hardware through a direct engineering collaboration. AMD Fellow Hisham Chowdhury confirmed that AMD Variable Graphics Memory further expands compatibility, allowing users to run models beyond those currently leveraging NeuroStream, signaling that the partnership is designed to scale as new models ship.
The AMD collaboration builds on an earlier joint effort that brought Starlight, the first diffusion-based video enhancement model, to AMD Radeon GPUs with up to 3.3x performance improvement and significant memory reduction versus the base model when running on AMD platforms. AMD's GPUOpen technical write-up details how the optimization pipeline used AMD Optimizer toolchains alongside advanced GPU pipelining and buffer-sharing techniques to eliminate redundant memory copies and reduce CPU synchronization overhead across the full video processing workflow. That same engineering foundation now underpins Wonder 2 (Local) and Starlight Precise 2.5 on AMD hardware.
On the product side, Topaz Video 1.6.0 brought Starlight Precise 2.5 support to both AMD and Mac platforms, specializing in realistic human face and skin textures while reducing flickering on smaller textures. The model outputs up to 4K and is available via cloud export for all subscription tiers, though local rendering on AMD and Mac requires meeting specific hardware thresholds. This release cadence, moving from NVIDIA-only local inference to cross-platform AMD and Apple Silicon support within a single quarter, positions Topaz as one of the few vendors offering diffusion-based video enhancement entirely on consumer hardware without cloud dependency.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source