Nvidia brings RTX video upscaling to older footage

By Splitfeed EditorialLast updated Oct 12, 2026, 20:00 UTC

On October 12, 2026, Nvidia announced that owners of RTX GPUs can now use the company's video upscaling technology to improve the visual quality of older footage. The rollout has two distinct fronts: a consumer-facing side, covered by outlets explaining how everyday RTX users can apply the feature to their existing video libraries, and a developer-facing side, in which Nvidia pushed RTX video upscaling support directly into ComfyUI, the widely used open-source image and video generation pipeline. Nvidia is framing this moment under the marketing label the RTX Spark era, suggesting the company sees it as a platform-level milestone rather than a minor feature update.

The integration with ComfyUI is what gives this announcement a broader significance. ComfyUI sits at the center of a large community of independent developers and researchers who build custom workflows on top of open-source tools, so embedding RTX upscaling there extends Nvidia's reach well beyond the typical retail consumer. It also continues a pattern of Nvidia positioning its hardware as the preferred substrate for AI-driven media tools, both proprietary and open-source. What remains genuinely unclear from available information is which specific RTX models are required, how the upscaling performs against competing tools, and whether any costs attach to the feature for developers building commercial workflows on top of ComfyUI.

None of the three camps have published reactions to this story yet. The Pro-AI camp would typically welcome the announcement as a practical demonstration of AI improving everyday media experiences with minimal friction for end users. The Anti-AI camp would likely raise questions about computational resource consumption, the environmental cost of running upscaling workloads at scale, or concerns about AI-altered footage being mistaken for original high-resolution material. The Middle Ground camp would probably focus on the specifics that remain unconfirmed, such as performance benchmarks and hardware requirements, and argue that the value of the feature depends heavily on details Nvidia has not yet disclosed.

The clearest next signals to watch are any independent benchmark comparisons between RTX upscaling and existing tools like Topaz Video AI, formal documentation from Nvidia clarifying minimum GPU requirements, and community feedback from ComfyUI developers on how smoothly the integration functions in real workflows.