Volatile vs Non-Volatile Memory: What Is the Difference?

Volatile memory needs a constant supply of power to hold its data, and loses it within moments of the power being removed. Non-volatile memory keeps its data without power, for years in most cases, until it is deliberately erased or overwritten.

The difference comes from how each type stores a bit. Volatile memory holds data as a charge on a capacitor (DRAM) or as the state of a powered circuit (SRAM). Take the power away and the charge drains or the circuit resets. Non-volatile memory traps the bit in a physical state that stays put on its own: electrons held in an insulated gate in flash, or a magnetic or ferroelectric orientation in newer types.

Every computer uses both. Volatile memory is fast and can be rewritten almost without limit, so it holds the working data. Non-volatile memory is slower and wears with use, but it keeps the operating system, applications and records safe while the system is off.

Some designs blur the line. Battery-backed SRAM, used for decades in PLCs and for the CMOS settings beside a PC real-time clock, behaves as non-volatile for as long as its battery lasts. NVDIMMs pair DRAM with flash and a backup energy source, so server memory survives a power cut, and Intel Optane persistent memory, now discontinued, sat between the two.

The table compares the memory types found in industrial and embedded computers, from the fastest to the most persistent.

Figures are typical and vary by maker and grade. Flash endurance depends heavily on how many bits each cell stores: SLC lasts longest, with MLC, TLC and QLC each trading endurance for capacity.

A system start shows how the two work together. When power is applied, the processor runs firmware from a small NOR flash chip on the board. The firmware finds the boot drive, an SSD or eMMC built on NAND flash, and loads the operating system from it into DRAM. From then on, the processor works from DRAM and its own SRAM caches, and writes results back to flash when they need to be kept.

Even the drive mixes both. NVMe, the interface used by fast SSDs, stands for Non-Volatile Memory Express, yet many NVMe and SATA SSDs carry a DRAM cache to speed up writes. Our FAQ on NVMe vs SATA on M.2 explains how to tell the two interfaces apart.

Volatile memory is the reason an unexpected power cut can damage data. Anything still in DRAM, in the operating system’s file cache or in the drive’s own cache is lost, and a write that was half finished can leave a file, or the file system itself, corrupted. In industrial systems the power can drop at any time, from a vehicle ignition switching off to a cabinet breaker tripping, so the design has to expect it.

Non-volatile does not mean permanent. Flash cells slowly leak their charge, faster when hot and faster as they wear. JEDEC’s JESD218 standard rates client SSDs to hold data for around one year unpowered at 30 °C at the end of their rated life, and enterprise SSDs for around three months at 40 °C. A new drive typically retains data far longer, but spare drives and archive media should not sit unpowered in a hot store for years. Choose flash with endurance matched to the write load, and keep logs and frequently changed data off small eMMC chips where possible.

Our industrial PCs are built with industrial-grade storage selected for endurance and temperature, rather than consumer drives.

We fit industrial-grade storage, set up write filters or read-only images and add backup power as standard where the site needs it, so each system we build survives an unplanned power cut.