German AI startup kausable has successfully secured €12 million in a seed funding round to develop AI that adapts without constant, costly retraining. The investment was led by European venture capital firms UVC Partners and Entourage, with support from HTGF and Mätch VC. The company's "reasoning-first" approach aims to create more efficient and autonomous intelligent systems for complex industries.
A New Approach to AI Learning
Kausable is challenging the conventional AI paradigm, which relies on massive datasets and frequent retraining to keep models current. The company's technology is built on a "world model" that learns from abstract cause-and-effect relationships. This allows the AI to develop a core understanding of how systems work, enabling rapid adaptation to new scenarios.
Instead of learning from language, the models are trained on synthetic causal data, a method that enhances privacy and data efficiency. This foundational knowledge allows the AI to transfer its learning across different domains with very little new information. It can therefore solve problems it has never encountered during its initial training phase.
Strategic Investment and European Ambition
The €12 million investment underscores strong confidence in kausable's vision, with UVC Partners and Entourage leading the round. The funding is also supported by prominent angel investors from organizations like OpenAI, Google DeepMind, and Black Forest Labs. This capital injection arrives as Europe seeks to strengthen its digital sovereignty by fostering homegrown frontier AI companies.
Andreas Unseld of UVC Partners noted that kausable's technology can turn AI into a scalable solution for industrial companies. Pieterjan Bouten of Entourage emphasized that the startup is building AI that reasons about the future rather than just remembering the past. Both investors see the company's approach as a foundational shift in artificial intelligence development.
From Research to Real-World Impact
Founded in 2025 by Johannes Haux, Dr. Benjamin Herdeanu, and Gregor Ramien, kausable originated from research at Heidelberg University. The team has already demonstrated its model's ability to predict rare, high-impact events in diverse fields. For example, it can forecast power grid blackouts or anticipate epileptic seizures from EEG data with minimal examples.
The new capital will be used to expand the nine-person team and accelerate the transition from research to commercialization. Kausable plans to launch pilot projects with early customers in sectors like robotics, forecasting, and industrial systems. These industries are ideal for technology that can adapt quickly to dynamic and unpredictable environments.
Vision for a Foundational Intelligence Layer
The company's long-term goal is to create a foundational intelligence layer that other specialized AI systems can utilize. This causal reasoning core would handle decision-making and predictions, working in tandem with vision and language models. This modular approach could significantly improve the overall capabilities and efficiency of future AI applications.
By focusing on efficiency, kausable also aims to dramatically reduce the computational cost and environmental impact associated with AI. This vision positions the technology not only as more powerful but also as more sustainable and economically viable. It offers a path toward making advanced AI more accessible across a wider range of industries.
Kausable's €12 million funding marks a significant step for European deep tech and the pursuit of more adaptable AI. The company's focus on causal reasoning addresses fundamental industry challenges, promising more efficient and intelligent systems. This investment accelerates its mission to redefine AI's capabilities and drive innovation across critical global sectors.