Tencent backs ComfyUI 3D
Tencent Cloud partnered with ComfyUI to accelerate open‑source 3D AI workflows, aiming to bring production‑ready 3D pipelines to millions of users and boost China’s creative‑AI stack (prnewswire.com). The deal signals a push to embed 3D generative tools into cloud stacks—another front in the race to commoditize multimodal content creation (prnewswire.com).
Tencent Cloud and ComfyUI formalized the partnership in a press release dated March 30, 2026, declaring Tencent HY 3D Global will be exposed as native Partner Nodes inside ComfyUI. (prnewswire.com) ComfyUI’s own blog shows Hunyuan (Hunyuan 3D 3.0) support landed in ComfyUI earlier on Feb. 13, 2026, with explicit text-to-3D, image-to-3D and sketch-to-3D generation paths implemented as Partner Nodes. (blog.comfy.org) The announcement cites ComfyUI’s scale — more than 105,000 GitHub stars and roughly 60,000 custom nodes — as the immediate distribution surface for Tencent’s 3D APIs. (prnewswire.com) ComfyUI shipped a Dynamic VRAM feature this month for Nvidia GPUs on Windows and Linux to reduce local RAM demands when running large models, a change that lowers hardware barriers for 3D model inference. (comfyui.org) Comfy’s March changelog (v0.18.2 and related releases) also records Partner Node and Tencent TextToModel/ImageToModel API fixes intended to stabilize cloud‑to‑local integration. (docs.comfy.org) Comfy’s fundraising post says the project raised $17 million to scale Comfy Cloud and related services, and third‑party listings show Comfy Cloud offerings that include 200+ pre‑installed models for browser/cloud use. (blog.comfy.org) (aicreators.tools) Tencent’s statement quotes Fred Sun, General Manager of Tencent Cloud Europe, tying the integration to Tencent’s Hunyuan/HY 3D Global APIs and listing those APIs as available through Tencent Cloud’s platform. (prnewswire.com)
