What is EEPROM (electrically erasable programmable read-only memory)?

Its name comes from the wider ROM (read-only memory) family that it belongs to, rather than RAM (random access memory), which loses its data the moment power is cut.

On average, an EEPROM memory chip is rated for 100,000 to 1 million write/erase cycles, more than most devices will need over a normal lifespan. Many of these chips are also rated for typical data retention of 10 to 20 years or longer, depending on the manufacturer and operating conditions.  Advanced industrial chips can last up to 100 years or longer.

EEPROM holds small amounts of data that a device needs to remember between power cycles. For instance, a phone might remember a paired Bluetooth speaker or a smart lock might store the code for an authorized user. Enterprise hardware, like network switches and industrial sensors, uses EEPROM the same way, just at a larger scale.

PROM, EPROM and EEPROM sound alike, but they’re not the same. All three are types of programmable read-only memory built between the 1950s and 1980s, each fixing something the last one could not accomplish.

PROM (programmable read-only memory) was invented in 1956 by engineer Wen Tsing Chow while he was working on a US Air Force missile guidance project. It can only be written to once, typically by a technician during manufacturing, after which the data is permanent and can’t be erased or rewritten.

Then, in 1971, Intel engineer Dov Frohman developed EPROM (erasable programmable read-only memory). This new type of ROM built on a proposal two Bell Labs researchers made a few years earlier to repurpose a transistor’s floating gate as a form of reprogrammable memory. EPROM chips can be erased and reused. Doing so requires exposing them to ultraviolet light through a small transparent window on top of the package. The chip must be removed from the circuit board first.

EEPROM arrived in the late 1970s and early 1980s, removing that last physical step EPROM still required. In contrast to EPROM, an EEPROM chip erases and rewrites data with an electrical signal, right on the board, without any UV light or removal needed. It also works byte by byte instead of wiping the whole chip at once, so changing one setting doesn’t mean touching everything else stored on it.

Flash memory is a form of non-volatile memory with ongoing data storage capability. It’s the memory most people interact with in their daily lives, powering USB drives, phone storage, solid-state drives and memory cards. It has also become the dominant choice for large-scale enterprise storage because it is cost-effective to manufacture at scale and fast enough for everyday use.

Flash grew out of EEPROM technology, and historically both were built around the same floating-gate transistor design. Today, they both store charge in an isolated gate structure. Where they diverge is in how each one erases data. Because classic EEPROM puts a select transistor next to every cell, it can update one byte at a time.

Flash memory, like NAND flash, drops this select transistor, enabling far greater cell density. This makes it faster and more cost-effective at scale, but less precise for small changes since it eliminates byte-level erase capability and erases in larger blocks. (NOR flash is closer to EEPROM in that it enables byte-level programming, but it too does not have byte-level erase.

Endurance is another significant difference between the two. EEPROM is typically rated for 100,000 to 1 million write/erase cycles, with industrial-grade versions reaching into the millions.  Flash holds up for fewer cycles per cell and endurance varies with cell type, such as single level (SLC) versus multi-level cells (MLC) or quad-level cells (QLC).

EEPROM and flash memory both have significant roles in the non-volatile memory market, which also includes newer types like MRAM (magnetic-based) and ReRAM (resistance-based).  According to Fortune Business Insights, the market memory types, was valued at USD 78.08 billion in 2025. Their research predicts the market will reach USD 135.37 billion by 2034, at a CAGR of 6.4%. AI data center expansion, industrial automation and connected consumer devices are fueling that growth.

Reading, writing and erasing data on an EEPROM means controlling the electrical charge inside its memory cells. Each cell relies on three parts:

- Floating gate
- Control gate
- Insulating (oxide) layer

Inside each transistor sits a small piece of conductive material, wrapped in a thin insulating layer. Whether it holds a charge or not represents a bit of data, a 1 or a 0, and that charge stays put even without power, the basic mechanism behind EEPROM’s non-volatility.

The control gate sits above the floating gate and manages the voltage applied to the cell during reads, writes and erases. During a read, it checks whether the cell’s threshold voltage falls above or below a reference point, which tells the chip whether that bit is a 1 or a 0.

Separating the floating gate from the rest of the transistor, this layer keeps electrons trapped in place until enough voltage moves them. Manufacturers must calibrate its thickness carefully. Too thin, and the charge leaks out on its own. Too thick, and the tunneling voltage can’t push electrons through it.

Programming and erasing an EEPROM comes down to one thing: moving electrons onto or off the floating gate. The chip applies a relatively high voltage, commonly between 12 and 20 volts, across the cell, pushing electrons through the insulating layer in a process called Fowler-Nordheim tunneling. Reverse the voltage and that same tunneling effect pulls the electrons back off, erasing what was written.

Because that entire process runs on an electrical signal, EEPROM can be reprogrammed in place, without the UV light or board removal that PROM and EPROM require. Manufacturers rely on that flexibility for anything that needs to change after a device ships, like a thermostat’s saved temperature or a car’s odometer reading.

Each cell can be addressed on its own, so an EEPROM can be programmed one byte at a time without touching anything else stored on the chip. In practice, a microcontroller handles this process, sending write and erase commands to the EEPROM over a serial protocol like I2C or SPI. This is common in embedded systems, where a microcontroller manages one specific task rather than running general-purpose software the way a computer’s microprocessor does.

Not every EEPROM communicates this way, though. A parallel EEPROM transfers multiple bits at once over separate data lines instead of one bit at a time. Serial EEPROM has become the more common choice in modern designs, due to its lower pin count and simpler wiring.

EEPROM’s design comes with several advantages, especially when a device needs to hold small amounts of data reliably over a long lifespan.

These benefits include:

- Byte-level access
- Reprogrammable in place
- Highly persistent
- Low power consumption

EEPROM updates data one byte at a time. A device can change a single setting without erasing or rewriting anything else stored on the chip, something flash memory’s block-based erasing can’t do.

EEPROM can be reprogrammed while still installed on the motherboard. There’s no removing the chip, no UV light, no specialized equipment, just an electrical signal sent through the existing circuit.

A standard EEPROM chip holds up for 100,000 to 1 million write/erase cycles and retains data for decades, sometimes over a century, without power. A device can be rewritten thousands of times over its lifetime and still hold onto that data long after it’s unplugged.

EEPROM uses relatively little power, especially for reads, which drives its popularity in battery-powered and embedded devices.

EEPROM plays a key role in devices ranging from consumer electronics to industrial hardware. Here’s a closer look at some common use cases:

- Computers
- Consumer electronics
- Automotive systems
- Industrial and enterprise equipment