Neurorium was born from a clear research insight: biological neural networks (BNNs) can be trained to perform computational tasks, potentially enabling a new path for energy-efficient intelligence. The company’s website describes how, in 2023, a joint research project between FinalSpark and the AI4Science Group at the Free University of Berlin - under the supervision of co-founder Mohsen Sadeghi - demonstrated that training BNNs is feasible. Building on this evidence, Neurorium was founded in 2025 to develop the technology required to unlock the potential of ‘Post-Silicon Intelligence’ - a future where living neurons complement, and in some cases replace, silicon-based computing systems.
At the core of Neurorium’s work is ‘Organoid Intelligence (OI)’. The company develops methods and approaches that translate digital information (numeric or visual data) into spike patterns intelligible to neurons, and then trains BNNs in closed-loop ways to achieve quantifiable computational objectives. It also provides an API that encapsulates the full toolchain so users can invoke capabilities analogous to traditional AI and machine-learning frameworks. Neurorium positions its technology as a practical, real-world solution for highly energy-efficient processing using living neural systems.