Standing inside At the HumanX conference at the Moscone Center in San Francisco, it’s strenuous not to feel like you’re at the center of the AI universe. Technology leaders swarm the building, with OpenAI and Anthropic’s headquarters right next door. However, the 70-person startup based 8,000 km away in Germany’s Black Forest – a region famed for ham – has become the main competitor of Silicon Valley’s leading AI image generation laboratories.
In December, Black Forest Labs raised funds on the website: a $3.25 billion quote after signing contracts to support AI image generation functionality in Adobe and the Canva graphic design platform. It has even struck deals with major AI labs like Microsoft, Meta, and xAI to provide similar features in their products.
Nearly two years after its launch, Black Forest Labs can afford to be picky about who it works with. In 2024, Elon Musk’s xAI leveraged Black Forest Labs into power Grok’s first image generator. This partnership put Black Forest Labs on the map, but caused a lot of controversy due to the chatbot’s confined security features. This ended a few months later when xAI developed an internal AI image model.
In recent months, xAI asked Black Forest Labs to relicense the startup’s technology, sources familiar with the matter told WIRED. Sources say that this time Black Forest Labs refused, considering cooperation with xAI, which is famed for its messy work environment, as too operationally hard. xAI did not immediately respond to WIRED’s request for comment.
In September, Black Forest Labs struck A multi-year agreement worth USD 140 million to provide Meta with access to AI image generation technology.
These AI labs want to partner with Black Forest Labs because their image generators are among the best in the world, ranking just below OpenAI and Google’s offerings for this third-party Reference points for artificial analysis. The startup also offers some of the most downloaded text-to-image models Face Huggingindicating that many AI imaging tools on the market likely apply a free version of Black Forest Labs’ technology.
This is particularly impressive because the company has historically had significantly fewer resources than its competitors. This has led to a more proficient line of research called latent diffusion, which essentially involves an AI model first sketching a coarse plan of the image and then painting in more detail.
Latent dissemination “enabled us to release very efficient models that required orders of magnitude less resources than our competitors’ models,” co-founder Andreas Blattmann told WIRED on stage at HumanX this week.
Despite its success, Black Forest Labs believes image generation is just the beginning. Blattmann said the startup plans to unveil a robot powered by one of its artificial intelligence models later this year. (He did not reveal which company produces this equipment). This venture is part of a larger opportunity the company sees in building artificial intelligence that can perceive the physical world and take action.
“Visual intelligence is much more than just content creation. Content creation is just the first step of this whole technology,” Blattmann said. “What I’m personally very excited about — and this is a pattern throughout this conference — is physical AI.”
WIRED sources say Black Forest Labs is also in talks with several hardware companies to power features for products such as shrewd glasses and robots.
A building in the Black Forest
Blattmann and his co-founders, Robin Rombach and Patrick Esser, made a name for themselves when they published groundbreaking research on AI image models in 2021. In 2022, they were hired by Stability AI and released Stable Diffusion, a popular open-source AI image generator based on their previous research. However, two years later they announced their departure and launched Black Forest Labs.
Instead of moving to San Francisco, the trio chose to maintain headquarters near their hometowns in Freiburg, Germany. Blattmann said this decision was key to the company’s success.