kausable is a deep-tech startup based in Heidelberg, Germany, developing what it calls "reasoning-first" AI. Its core product is a horizontal AI foundation built on causal world models trained on synthetic data. The company says these models learn cause-and-effect from minimal data, adapt to new situations without retraining, and handle problems they have not seen before - bypassing the retraining loop characteristic of conventional AI systems.
The technical work spans AI foundation models, causal world models, synthetic-data training, machine reasoning and generalisation to novel problems, with applications in robotics and applied machine learning. Target verticals include robotics, energy, finance and medicine, as well as environments marked by rapidly changing conditions and rare events.
The team numbers about nine people and is led by physicists; the company describes itself as research-led and staffed by "scientists at heart with a builder's mindset". It is moving from research toward placing models in front of real customers, and is hiring across research, engineering, product and operations. kausable has raised €12m in seed funding, led by UVC Partners and Entourage, with angel investors from OpenAI, Google DeepMind and Black Forest Labs.
