Breaking

NVIDIA Shows AI Agent Orchestrating Rhino, ComfyUI and Blender

June 11, 2026 at 19:14 EDT

AI Agents
Media Generation
Infra & Chips

NVIDIA has demonstrated a design workflow in which an AI agent handles everything from a concept sketch to a photorealistic render in a single flow, showcasing "agentic AI" that links specialist tools such as Rhino, ComfyUI and Blender. The company frames this as a response to an era in which workflows are defined by the connections between tools rather than the tools themselves.

Computex 2026 · NVIDIA RTX Spark

One Workflow, Many Tools: AI Agents Stitch the Design Pipeline Together

NVIDIA demos autonomous agents that run locally on RTX Spark — chaining Rhino, ComfyUI and Blender from concept sketch to photorealistic render in a single flow. The value, it argues, lives "in the connections between tools, not the tools themselves."

~2×

inference speedup for top agentic models via multi-token prediction

128GB

unified memory on the DGX Spark hardware example

1 PFLOP

up to, of AI performance (DGX Spark)

~$4.7K

price example for the DGX Spark unit

The Agent Pipeline

A Hermes Agent orchestrates a multi-step process as one flow — running locally on RTX Spark inside a secure OpenShell sandbox.

Concept

sketches & mood boards

→

Rhino

site & structure 3D

→

ComfyUI

image generation

→

Blender

photoreal render

Agentic Inference Throughput

Multi-token prediction in llama.cpp / vLLM roughly doubles performance for top agentic models.

1×

baseline

~2×

with MTP

What works

Removes tool-switching friction — one flow, concept to render
Local execution: privacy, low latency, less cloud dependence
Designers focus on intent and feedback

Still maturing

Aesthetic judgment & design quality still depend on humans
UX aimed at beginners — generation-heavy, little component reuse
Accuracy like flawless mesh generation is a work in progress

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.