ComfyUI has raised $30M at a $500M valuation, betting that creative professionals will pay for control over diffusion models rather than settle for what a prompt box hands back. Craft Ventures led the round, with Pace Capital, Chemistry, and TruArrow joining. The company says more than 4 million users now run its node-based workflow tool for image, video, and audio generation.
The new round values ComfyUI at roughly 26x the $19M Series A it closed in late 2024, when Chemistry Ventures, Cursor Capital, and Vercel founder Guillermo Rauch backed the company. That earlier financing came barely a year after ComfyUI emerged from open source in 2023, in the wake of the first wave of diffusion models from Midjourney and OpenAI's DALL-E.
Co-founder and CEO Yoland Yan frames the pitch in arithmetic. Prompt-based tools like Midjourney and ChatGPT, he told TechCrunch, get a creator 60% to 80% of the way to a usable output. The last 20% — adjusting a hand, a shadow, a piece of text — is where the work actually lives, and where prompt boxes break down.
Key facts
- 01ComfyUI raised $30M at a $500M valuation, led by Craft Ventures with Pace Capital, Chemistry, and TruArrow.
- 02The company raised a $19M Series A in late 2024 from Chemistry Ventures, Cursor Capital, and Vercel founder Guillermo Rauch.
- 03ComfyUI claims more than 4 million users across visual effects, animation, advertising, and industrial design.
- 04CEO Yoland Yan says prompt-based tools like Midjourney and ChatGPT get outputs 60% to 80% right, leaving the final 20% as a guessing game.
- 05ComfyUI started as an open-source project in 2023 and competes with Weavy, which Figma acquired last year.
Yan compares re-prompting to a slot machine. A small request for one tweak can return a completely different image, often overwriting the parts that were already correct. ComfyUI's interface lets creators wire individual steps of the generation pipeline together as nodes, so changes to one stage do not blow up the rest of the output.
“ComfyUI claims more than 4 million users and a $500M valuation built on a simple bet: prompt boxes get creators 60% to 80% of the way there, and the last 20% is where the work lives.”— Jaeden Schafer
"You cannot easily convey that message in the prompt box," Yan said of foundation models. The node graph is the workaround: explicit, visual, and reproducible, in a way that a chat turn is not.
The customer base reflects that pitch. ComfyUI says its users include technical artists in visual effects, animation, advertising, and industrial design, and that studio job boards now list "ComfyUI artist" or "ComfyUI engineer" as a discrete role. That is an unusual signal for a tool that is barely three years removed from its open-source origins.
The competitive map is thin but real. ComfyUI's closest analog, Weavy, was acquired by Figma last year, pulling one rival into a design suite aimed at a different buyer. That leaves ComfyUI as the dominant standalone option for the granular, professional end of generative media work, even as Midjourney, OpenAI, and a long list of video model startups push their own consumer-facing tools.
Yan's bet is that better foundation models do not eliminate his market — they expand it. "In the world where AI slop is going to be everywhere, the Comfy version of human-in-the-loop approach is going to win out most of the eyeballs in the end," he said. The argument: as raw generation gets cheaper, the premium shifts to outputs that look deliberate.
The risk runs the other way. If the next generation of video and image models closes more of that 20% gap on their own — through better instruction-following, inpainting, or controllable diffusion — the ceiling for a node-based tool gets lower. Foundation labs have every incentive to absorb that workflow into their own products rather than cede it to a third party.
There is also a distribution question. ComfyUI's 4 million users are weighted toward technical creatives who already tolerate a steep learning curve. Translating that into the kind of seat-based enterprise revenue that justifies a $500M valuation will require either a managed cloud product, a team tier, or both — none of which the company has detailed publicly.
Still, the round is a clean read on where investors think generative media is heading. The money in image and video AI has mostly chased model labs and consumer apps. ComfyUI is a wager that the durable margin sits one layer up, in the tooling that turns a probabilistic model into a deliverable a client will actually accept. If that thesis holds, the next round will not be the last.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
