Abstract
<jats:p>The future of Very Large Scale Integration (VLSI) architectures is being fundamentally reshaped by the intersecting demands of Neuromorphic computing, edge artificial intelligence (AI), and sustainable electronic systems. As traditional CMOS scaling nears its physical limitations, researchers are exploring bio-inspired Neuromorphic designs that mimic the brain's energy-efficient computing mechanisms using spiking neural networks (SNNs), as demonstrated by platforms like Intel’s Loihi and IBM’s True North (Davies et al., 2018; Merolla et al., 2014). In parallel, edge AI is driving the need for VLSI systems capable of performing high-speed, low-latency inference directly on-device, thus reducing reliance on cloud infrastructure. Hardware accelerators such as Eyeriss and Google’s EdgeTPU offer promising solutions through energy-aware architecture and domain-specific design (Chen et al., 2019; Jouppi et al., 2017). These innovations are increasingly evaluated not just on performance, but also on ecological impact. Emerging sustainable VLSI solutions include non-volatile memory technologies like ReRAM and PCM, in memory computing approaches that reduce data movement, and eco-friendly materials like carbon nanotubes and organic semiconductors (Sebastian et al., 2020; Rajendran et al., 2020). However, despite significant advancements, challenges persist in fabrication scalability, algorithm-hardware co-design, and lifecycle sustainability assessment. This paper explores the synergetic evolution of these domains and highlights the critical need for interdisciplinary collaboration to develop future VLSI architectures that are intelligent, efficient, and environmentally responsible. The convergence of Neuromorphic computing, edge AI, and green hardware marks a paradigm shift, positioning VLSI not only as a medium for computation but as a cornerstone of sustainable technological progress. </jats:p>