Overview of emerging electronics technologies for artificial intelligence: A review

Abstract

This paper shows the short- and long-term electronics technologies emerging as the enablers of next-generation AI systems and focuses on rapidly developing technologies with promise toward enabling the new AI revolution, such as neuromorphic, quantum computing and edge AI processors. These technologies are key to improving the computational power, energy efficiency, and scalability required in AI solutions across healthcare, autonomous systems, and better endeavours. Neuromorphic computing works similarly to the brain's neural configuration to build a more energy-efficient AI system by simulating biological functionality, while quantum computing is ubiquitous as the next stage of problem-solving systems in AI and exponentially increases computational speed and functionality. Finally, Edge AI processors play an important role in real-time AI decision-making, especially in environments with limited power and space, as they allow data to be processed at the original point of generation. Of course, although these technologies demonstrate great potential, there are still obstacles to overcome for subtle hardware-software integration, architecture scalability and high energy consumption. This study highlights sustainable hardware design as an essential solution to these challenges, discussing low-power chips, AI accelerators and energy-efficient designs that allow devices to run at scale without performance liabilities. The paper also highlights quantum and neuromorphic computing—which mimics the structure and function of biological brains—as an important focus for overcoming limitations regarding scalability, allowing for novel architectures equipped to deal with the extremely large amounts of data required for future, more advanced AI models. We also discuss how these progressions can facilitate the creation of effective and scalable AI systems that support AI in addressing global challenges like environmental deterioration and resource limitations. Lastly, the paper highlights the importance of ongoing research and innovation in such areas to promote the evolution of AI systems that are resilient, scalable and energy-efficient in a way that ensures the long-term sustainability of AI and its implementation in various domains.