Biological neurons form the foundation of the human brain, one of the most efficient and sophisticated information-processing systems known.

Every computing era is defined by its processor.

In the biocomputing era, that processor is the BPU.

Biological Processing Unit (BPU) is a new kind of computing system that uses living biological neurons as its computational core.

~20W

The human brain performs complex information processing using roughly the power of a light bulb.

Massively parallel

Biological neural networks process information simultaneously across vast, interconnected networks.

Adaptive by nature

Neurons continuously change in response to activity and feedback, enabling dynamic computation.

BPUs are built to train and run intelligent models.

BPUs combine biological neurons, engineered interfaces, and software control into a functional computing system.

Efficient training

Conventional intelligent model training is resource intensive, requiring vast amounts of compute, power, and investment. BPUs offer a new substrate for building intelligence more efficiently.

Scalable inference

Inference at scale consumes significant power and generates substantial heat, making deployment expensive and infrastructure-heavy. BPUs offer a more efficient way to run intelligent models, with lower energy demands and minimal heat generation.

Intelligence for everyone

When intelligence becomes cheaper to build and easier to run, it can become more accessible, more personalized, and more private.

Input

Output

A task, such as a game or text prompt, is translated into electrical signals and delivered to the neurons through the BPU interface.

The resulting neural activity is recorded and translated back into digital output, such as a game action, a prediction, or a text response.

The neurons respond by performing the computation required for the task. With feedback over time, they adapt and learn, improving how they play a game, solve a problem, or generate the desired response.