Neuromorphic Computing Chips: Challenges and Trends

Over the past decade, artificial intelligence (AI) has made unprecedented advancements in various fields. However, with the emergence of large-scale models in recent years, the energy consumption associated with AI computing has become a critical issue that urgently needs to be addressed. Furthermore, as Moore's Law reaches its limits, increasing computational power has become extremely challenging. The scarcity of energy and computational resources presents significant challenges to the further development and application of current AI technologies. These challenges provide an unprecedented opportunity for the introduction of neuromorphic computing chips which show great potential in addressing the problems currently faced by AI. We present in this article the evolution of brain-inspired neuromorphic computing chips, tracing their evolution from the early artificial retina to advanced designs incorporating millions of artificial neurons. Also, we explore future opportunities in key applications such as brain-computer interfaces, which can facilitate more efficient neural communication; embodied intelligence, which seeks to replicate human cognitive functions; and large-scale models, where these chips' energy-efficient processing capabilities can greatly enhance AI system performance and scalability.