Stateful Robotics, an embodied AI company spun out from the University of Oxford, has successfully raised $4.8 million in a pre-seed funding round. The investment, led by Amadeus Capital Partners and Oxford Science Enterprises, will advance the company's mission to equip robots with long-term memory and planning capabilities. This new platform aims to solve a critical bottleneck that hinders robotic deployment in unpredictable, real-world environments.
Addressing the Memory Gap in Modern Robotics
Despite significant progress in AI, most robotic systems operate without a memory of past events, treating each decision in isolation. This "stateless" approach makes them vulnerable to disruptions such as unexpected obstacles, changing light conditions, or altered workflows. Consequently, even minor deviations from pre-programmed instructions can lead to operational failure and require human intervention.
This fundamental limitation has stalled the widespread adoption of autonomous systems in complex industrial settings like warehouses and infrastructure sites. While foundation models have improved perception, they do not inherently retain historical context about recurring issues or site-specific behaviors. The inability to plan over longer time horizons remains a critical barrier to achieving reliable and scalable robotic solutions.
Introducing Stateful AI for Enhanced Autonomy
Stateful Robotics directly confronts this challenge with its innovative AI platform that provides a persistent intelligence layer. The system continuously integrates real-time data, task progress, and historical performance into a dynamic, shared model. This approach effectively gives robots a memory, allowing them to recall past incidents and understand the flow of work over time.
By maintaining this continuously updated model, the platform enables robots to move beyond reactive, short-term decisions. They can anticipate potential issues, adapt to disruptions, and plan missions more effectively over extended periods, such as hours or days. This capability ensures that single robots, fleets, and human-robot teams can operate with greater consistency and reliability.
A Team Rooted in Oxford's Research Excellence
The company is spearheaded by a team with deep expertise in both commercializing AI and pioneering academic research. CEO Kirsty Lloyd-Jukes previously served as CEO of Latent Logic, another Oxford spinout acquired by Alphabet's Waymo in 2019. Her leadership is complemented by the technical vision of her co-founders from the University of Oxford.
The technology is built upon a decade of foundational research into autonomy, probabilistic verification, and decision-making under uncertainty. The co-founding team includes Chief Scientist Professor Nick Hawes, Professor David Parker, and Dr Bruno Lacerda. Their collective work forms the scientific backbone of the company’s stateful intelligence platform, bridging research with industrial application.
Strategic Funding to Accelerate Growth
The $4.8 million pre-seed investment was co-led by prominent venture capital firms Amadeus Capital Partners and Oxford Science Enterprises. The round also saw participation from angel investor Stan Boland, a serial entrepreneur who founded the autonomous vehicle platform Five. This strategic backing underscores investor confidence in the company's approach to solving a core industry problem.
This new capital will be instrumental in scaling the company's operations and accelerating its go-to-market strategy. The funds are earmarked for expanding the engineering team, further developing the core performance engine, and strengthening partnerships with industrial clients. Stateful Robotics is already testing its platform with pilot customers in the logistics and infrastructure sectors.
With this significant pre-seed funding, Stateful Robotics is well-positioned to bridge the gap between the potential of robotics and its practical, large-scale implementation. By endowing machines with memory and long-term reasoning, the company's platform could unlock a new era of reliable automation in complex environments. This innovation promises to transform robots from single-task tools into adaptive, continuously learning systems ready for widespread industrial adoption.