This paper presents a system-level quantitative analysis of energy efficiency across three computational paradigms: GPU-based deep learning, neuromorphic silicon, and hybrid bio-digital platforms exemplified by the CL1 system (Cortical Labs). Using publicly available data, we show that while biological neurons operate at ~10⁻¹⁵ J per spike, the full system power envelope is dominated by life-support infrastructure (30–80 W). We propose standardized benchmarking criteria (BioMark) and a scalability roadmap for bio-digital computing.
Henrique Peixoto (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: