This week on fxpodcast, Foundry Creative Director Juan Salazar talks about the latest advancements and strategic changes within Nuke 17. The main focus of our conversation will be the integration of artificial intelligence through tools like CopyCat and BigCat, allowing artists to automate tedious tasks and manage large-scale data training across sequences.
This interview also highlights a major overhaul of the 3D system, including the introduction of Gaussian splats and improved USD support for better performance and creative flexibility. Salazar explains that these updates are aimed at streamlining compositing workflows and allowing artists to more efficiently handle complex reflections and projections within a single environment.
We also discuss technical improvements in this release of Nuke 17, including Aces 2 integration, faster deep rendering, and an improved annotation system for better team collaboration. Listen to this week’s fxpodcast as Salazar explains Foundry’s commitment to providing sophisticated tools that empower individual artists while maintaining fast iteration cycles.
See also our previous article on Nuke 17’s Gauss Splats.
NUKE 17.0 Beta with Gauss Splats
BigCat: Scaling machine learning beyond shots
Nuke 17 expands Foundry’s machine learning ecosystem with the introduction of BigCat, an evolution of the CopyCat toolset designed for large training datasets. While CopyCat was originally designed for individual shots or small groups of shots, BigCat enables training of entire sequences or entire shows, responding to feedback from studios seeking greater control and scalability.
Interestingly, CopyCat adoption has been quietly increasing across the industry, with many studios using CopyCat internally in their productions, even if its usage is not widely discussed publicly. BigCat is building on its momentum by enabling more technology users to manage large datasets and training workflows directly within Nuke. Artists can start with CopyCat and move to BigCat depending on the size of the project, but training on a broader dataset from the beginning may reduce the risk of overfitting to a particular shot.
The discussion also touched on the current sensitivities in the industry regarding AI in general. Many of the uses for machine learning at Nuke involve practical tasks such as cleanup, edge refinement, transfer of looks between renderers, and lens distortion workflows rather than generative content creation, but the studio often remains cautious about publicly discussing AI’s involvement in production.
GripTape integration and artist amplification
Foundry previously announced the completion of its acquisition of Griptape, a company focused on enterprise-grade AI orchestration, a move that signals an important next step in the evolution of AI in professional VFX and animation pipelines. Listen to that interview here
Foundry acquires Griptape – exclusive interview with fxpodcast
Going forward, Foundry’s acquisition of AI orchestration company GripTape is expected to significantly expand its machine learning workflows. Our long-term vision is to encapsulate complex AI processes into artist-friendly tools within Nuke, reducing technical barriers while increasing automation capabilities.
For compositors, this represents a change in productivity. Rather than reducing the role of the artist, automation is intended to eliminate repetitive tasks and allow individuals to focus more on creative decision-making and storytelling across sequences. Allowing a single artist to accomplish more in the pipeline could partially reverse decades of increased specialization in the sector.
