Neuromorphic hardware for the next generation of artificial intelligence.
A revolution in artificial intelligence, enabled through fundamental physics and electricity.
Today's artificial neural networks are simulated on digital hardware, like GPUs and CPUs. These simulations incur a massive cost overhead, and training the largest neural networks can cost tens of millions of dollars.
Our neuromorphic hardware utilizes analog physics to achieve massively scalable and efficient artificial intelligence. This stands in stark contrast to the digital simulations that abstract neural networks away from their original inspiration, the brain.