Access Official Technical PDF

Share Summary:LinkedInEmail

ReRAM memory technology, also known as resistive RAM, is being developed as an alternative approach to non-volatile memory (NVM). For industrial automation and electronics teams, the interest is practical: memory must retain data without power while fitting the requirements of increasingly capable devices and systems. A 2020 presentation from Weebit Nano described development work aimed at both embedded memory and standalone memory chips. Its claims and timelines reflect the company’s position at that time, rather than a statement of current product availability.

Where ReRAM fits in the memory landscape

Semiconductor systems use different types of memory for different purposes. SRAM and DRAM are volatile: they lose stored information when power is removed. Flash is non-volatile and is used to retain data, but developers continue to investigate alternatives as requirements for cost, power consumption, reliability and scaling evolve.

ReRAM belongs to the non-volatile category. The company described its approach as using silicon oxide-based materials and positioned it as a potential alternative to Flash. Its presentation also claimed substantial speed, energy-efficiency and endurance advantages over existing Flash technology. Those figures are company claims, not universal characteristics of ReRAM; real-world performance depends on the specific design, process and application.

The potential use cases span two broad market segments. Embedded memory is integrated into a system-on-chip (SoC), alongside processing and other functions. Discrete memory, by contrast, is supplied as a standalone chip and integrated onto a circuit board. Both approaches can be relevant to equipment that needs persistent storage, but they bring different design and integration considerations.

ReRAM memory technology for embedded systems

For embedded applications, a memory array is only one part of the solution. It must be incorporated into a usable memory module and integrated with the surrounding SoC design. That module-level work is important to system developers because a promising memory cell does not, by itself, establish that the technology is ready for an end product.

In its 2020 presentation, Weebit reported that a 1 Mb array had been developed with a partner and that the results were considered market-competitive. The company also reported endurance of 106 cycles and retention of 10 years at temperatures up to 150°C. At the time, development of the memory module around the array was ongoing, with customer demonstrations and agreements described as future targets. These were historical milestones and targets, not confirmation of later outcomes.

For engineers assessing an embedded NVM option, this distinction matters. Array-level results are relevant, but an evaluation also needs to account for the complete memory module, its interfaces, process integration and the qualification needs of the intended application. The presentation’s stated plan to transfer the technology to a production facility likewise highlighted that moving from development to manufacturable technology is a separate stage.

Discrete memory and the role of a selector

Standalone memory devices present a different integration path. The company described initial work toward entering the discrete market by investigating whether its ReRAM could be incorporated into a partner’s products, with NOR Flash replacement identified as an early focus. External laboratory testing of technical parameters was reported in 2019, followed by investigation of possible product integration.

A further part of this work involved a selector. In a memory array, selector devices can be used to control access to individual memory elements and help manage the operation of the array. The presentation described work with research institute Leti to develop an integrated ReRAM-and-selector demonstration. It also noted that the selector was still under development. That dependency is important: a memory technology’s commercial prospects can rely not only on the storage element but also on supporting components and a manufacturable array architecture.

What validation and commercialization require

Third-party testing can provide evidence that a technology’s measured parameters are not limited to the developer’s own laboratory. However, technical validation is not the same as production qualification or commercial adoption. For system integrators and plant technology teams, the broader assessment would include whether the memory meets application requirements, how it fits the semiconductor process, and whether the complete solution can be supplied and supported.

The presentation outlined a development path that included partner testing, memory-module development, selector work and technology transfer to production. It also identified customer agreements as a goal. These steps illustrate the range of work between demonstrating a memory array and delivering a product that equipment manufacturers can design into systems.

ReRAM’s prospective application areas included connected devices, robotics, autonomous vehicles and artificial intelligence. The presentation also proposed that ReRAM’s physical behavior could be useful in brain-inspired computing. These were areas of potential interest, rather than proof that the technology was already deployed in those applications.

Practical implications for automation decision-makers

Industrial automation products often combine controllers, sensors, communications and edge-processing functions, each with its own memory needs. When evaluating an emerging NVM option, decision-makers should separate the technology’s reported cell-level results from the readiness of the full memory product. Useful questions include:

  • Is the target use embedded memory in an SoC or a standalone memory device?
  • Are endurance and data-retention results demonstrated for the relevant operating conditions?
  • What module, selector or other supporting development remains?
  • Has the technology progressed from laboratory validation to production transfer and application qualification?
  • Are integration plans and commercial availability established for the required design timeline?

ReRAM is a developing non-volatile memory approach with potential applications in both embedded and discrete markets. The 2020 Weebit presentation provides a snapshot of the work then underway: array development, partner validation, module integration, selector development and plans for production transfer. For automation engineers and technical decision-makers, the key lesson is to assess the complete implementation and its maturity—not just the promise of the memory cell—before considering it for a system design.

Access Official Technical PDF

About The Author