Answers are grounded in this paper and cite their sources.
Memristors have emerged as a promising technology for in-memory and neuromorphic computing, offering a route around the von Neumann bottleneck. As the field diversifies across material systems, the question has shifted from whether a device can switch to how well, how reliably and how efficiently each kind does so. After a brief account of the memristor’s history and the controversies surrounding it, this performance-centric review surveys the landscape of memristor technology, tracing progress in seven categories of resistive-switching material: four defined by composition or structure (oxides, organics, 2D materials and perovskites) and three by switching mechanism (phase-change, ferroelectric and spintronic). The categories are evaluated against a comprehensive set of figures of merit (FoMs) spanning electrical switching, reliability, scalability and synaptic operation. Of these, five key FoMs and the switching and transport mechanisms are compiled, where reported, for 223 devices, with statistical distributions presented for each category so that typical device performance can be read against the best achieved. The remaining FoMs, including the neuromorphic ones, are discussed where reported. The analyses provide a comparative view across categories, highlight trade-offs within them, and relate performance to intrinsic material properties and underlying physical mechanisms. Optoelectronic operation and other notable advances are also covered, and the readiness of each category for industrial use is assessed. The result is a quantitative cross-category reference that lets researchers orient themselves and benchmark new devices.
No takes yet. Share an insight, caveat, or question.
Khwairakpam et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: