Rangduru, Inc.

Founded in 2011 by CEO Euipil Kwon and headquartered in San Jose, CA, Rangduru is a semiconductor company specializing in half-transistor non-volatile memory (0.5T NVM) for embedded and standalone memory markets. Rangduru’s revolutionary 0.5T NVM is a sophisticated memory platform more effective and cost-efficient than conventional OTP memory, NOR Flash memory, DRAM, and SRAM. These competitive advantages stem from two crucial factors: high areal density and non-volatility. Rangduru’s 0.5T DX-OTP™ memory is four times denser than 2T OTP alternatives. And its 0.5T1R DX-NVRAM™ — encompassing DX-MRAM™, DX-PCRAM™, DX-RRAM™ — is two to ten times denser than 1T1R NVM and significantly denser than 1T1C DRAM. As a result, Rangduru’s 0.5T NVM has a lower cost per area, a lower rate of power consumption, and lower latency than alternatives. Furthermore, unlike DRAM and SRAM, Rangduru’s memory is non-volatile, which means that data stored in the cells are not lost when it loses power. These efficiency-enhancing, cost-saving innovations are crucial to the future of AI. Machine learning algorithms — especially deep learning algorithms — require massive amounts of data and power to train. Power inefficiency inhibits AI from expanding into the edge, instead largely confining AI to datacenters where access to power is ample. But even there, the glut of data remains a problem; the computational demands placed on AI servers vastly outstrip the memory bandwidth and storage capacity of their hardware components. Consequently, the number, size, and power consumption of servers continue to proliferate, training times last for weeks, even months, and AI algorithms are unable to reach their full compute potential. Rangduru is offering a promising solution to these many problems, helping us expand AI into the edge, enhance AI on the cloud, and move forward into the future of the information age.