Bria Platform Gives Enterprises Control Over GenAI
Bria, a visual generative AI platform, is offering organizations granular control over model training and deployment. The platform includes mechanisms for enforcing authorship, data provenance, and brand compliance rules. The company's founder described the system as a way for enterprises to maintain creative and ethical standards in automated visual workflows.
- Bria's core technology is a patented attribution engine that traces generated outputs back to the original licensed training data, allowing for programmatic compensation to content creators. The revenue-sharing model is designed to function similarly to Spotify's, providing micropayments to artists based on how their work influences new creations. - To ensure its models are "commercially safe," Bria trains them exclusively on 100% licensed data from over 30 partners, including major stock image providers like Getty Images, Alamy, and Envato. This approach provides users with full IP and privacy indemnity and ensures compliance with regulations like the EU AI Act. - The company has raised a total of $65 million, including a $40 million Series B funding round led by Red Dot Capital with participation from Intel Capital and GFT Ventures. These funds are earmarked for scaling its platform and expanding its attribution technology to other media formats, including music, video, and text. - For builders, Bria provides extensive developer access, including APIs, SDKs, and even open-source models with full access to the source code and weights. This allows for deep customization and integration into various development environments and creative workflows. - The platform is expanding its presence in multi-tool creative workflows through integrations with professional software. Bria's models are now available in tools like Photoshop, ComfyUI, Houdini, and Nuke, embedding its capabilities directly into existing production pipelines. - Founder and CEO Dr. Yair Adato has a background that includes scaling the retail AI company Trax to a unicorn and holding a PhD in Computer Vision. He states that this experience shaped Bria's focus on solving real-world production constraints like governance and business accountability. - In June 2025, Bria released its third-generation text-to-image model as its first fully open-source offering. The company claims it matches the performance of leading open-source models with 66% fewer parameters and is twice as easy to fine-tune, requiring 50% less data and compute power.
