Stable Diffusion

Stable Diffusion is an open-source image generation model that uses diffusion-model architecture to create images from text prompts, released in 2022 and widely credited with catalyzing the consumer AI image-generation wave . Its open-source release inverted the prior dynamic where closed models like DALL-E and Imagen dominated, quickly attracting a developer community that built an extensive plugin ecosystem including ControlNet, T2I-Adapter, and LoRA—tools that gave users far more control than closed alternatives . In the view of one 5Y Capital piece, this made Stable Diffusion "truly a productivity tool" rather than "just a toy," though Midjourney remained superior on raw image quality and ease of use .

The model's impact extended beyond product adoption to community composition: per Yunqi Capital, Stable Diffusion marked an inflection where open-source AI communities expanded beyond core ML engineers to include "grassroots researchers" and hobbyist contributors . Its training cost—just "a few million dollars" in GPU time—was orders of magnitude below contemporaneous large language models, making it a paradigmatic case of efficient capability deployment . In subsequent years, the term "Stable Diffusion moment" became shorthand in Chinese VC discourse for any open-source release that democratizes access to a previously researcher-only technology, applied by one Zhenge Fund contributor to the agent framework OpenClaw in 2026 .