Bria Secures Fresh Funding for AI Models Trained on Licensed Data

AI-powered image generators are currently at the heart of numerous copyright lawsuits against AI companies, primarily because they're trained on vast datasets pulled from public websites. These companies often defend their practices by claiming protection under the fair use doctrine, yet many copyright holders challenge this assertion.

As a result, some startups, including New York and Tel Aviv-based Bria, founded in 2020 by Yair Adato and Gal Jacobi, are taking a different approach. Bria opts to train its image generators solely on licensed content, collaborating with around 20 partners, such as Getty Images, to ensure compliance and content integrity. CEO Yair Adato emphasized that Bria compensates image owners "programmatically" based on their "overall influence" on the platform.

"Bria foundation models house one billion visuals and millions of videos," Adato shared with TechCrunch. He further noted that by training on globally representative datasets, Bria has successfully reduced biases often found in AI-generated visuals. This approach enables their models to produce diverse and inclusive content, ideal for a wide range of creative uses.

Bria enhances user experience by offering plug-ins compatible with popular image editing and design software like Photoshop and Figma. Additionally, they provide a fine-tuning API that allows customers to tailor Bria's models for specific needs. Users have the flexibility to operate these models on Bria's platform or on external computing environments, like public clouds. Adato assured that in all scenarios, customers retain ownership of both the data and the outputs.

"Enterprise customers can access our source code and [models] through payment," Adato explained. "We offer more than 30 specialized APIs for visual creation and modification, available via subscription and usage-based pricing. Companies can also fine-tune our generative AI models with their brand assets, creating custom engines that preserve their visual identity."

Bria's vision extends beyond current capabilities. Adato shared with TechCrunch that the company, with its 40 employees, aims to build an "IP ecosystem" that allows businesses to use licensed images from media conglomerates for commercial purposes, complete with "built-in compliance."

The company is not stopping at images; Bria plans to expand its platform to support other media types like music, video, and text, as well as on-device applications. Despite challenges in the broader tech industry, such as market maturation and macroeconomic pressures, Adato believes these factors only strengthen Bria's market position.

While other ventures like Adobe, Spawning AI, and Shutterstock are also venturing into licensed media generators, Bria has carved out a significant position in this emerging market. On Thursday, they announced a successful Series B funding round, raising $40 million, led by Red Dot Capital and supported by Maor Investments, Entrée Capital, GFT Ventures, Intel Capital, and IN Venture. This brings Bria's total funding to approximately $65 million, with most of the new funds earmarked for product development.

"We're experiencing rapid growth with our 40 customers, seeing over 400% annual recurring revenue growth last year," Adato noted. He also mentioned plans to expand the team with experts in generative AI for music and video, global sales and marketing, IP and copyright law, and AI consultancy, aiming to double the team size by year's end.

Updated 3/15 11:22 p.m. Pacific: The original version of this story incorrectly stated that Bria was founded in 2023, not 2020, and failed to mention co-founder Gal Jacobi. We apologize for these errors.