Stanhope AI has raised $8M in seed funding to develop brain-inspired AI based on the Free Energy Pri...
The AMW Read
The article introduces a non-transformer, neuroscience-grounded architecture (Active Inference) specifically for embodied AI, which meaningfully updates the baseline for how physical agents learn.
Stanhope AI has raised $8M in seed funding to develop brain-inspired AI based on the Free Energy Principle and Active Inference, moving beyond traditional LLM approaches. Founded by leading neuroscientists from UCL and King's College London, their "Real World Model" enables autonomous systems to learn and adapt continuously in dynamic physical environments like robotics and drones. This approach promises more efficient decision-making without the hallucination issues plaguing current AI systems. The funding will accelerate field trials in defense and industrial automation throughout 2026. This signals a strategic shift toward neuroscience-grounded AI architectures for embodied applications.
#AI #Neuroscience #Robotics #ArtificialIntelligence #DeepTech #ActiveInference #Funding
