Embedded Basics

Firmware vs Software: What's the Difference?

Firmware vs software explained: how they differ in purpose, storage, updates and hardware ties, why the line blurs, and what firmware means in embedded systems.

On this page
  1. The short answer
  2. Comparison table
  3. What is firmware?
  4. What is software?
  5. The key differences explained
  6. Firmware in embedded systems
  7. Where the line blurs
  8. Examples of firmware and software in everyday devices
  9. How firmware gets onto a device and how it’s updated
  10. Frequently asked questions

The short answer

Firmware is a type of software. It’s the code built into a specific piece of hardware that makes that hardware work. It’s stored in non-volatile memory on the device, usually flash, and it handles low-level jobs: starting the device, controlling its components and talking to other hardware.

Software, in everyday use, means the operating systems and applications you install, run and update on a general-purpose device: a web browser, a word processor, Windows or Android.

So the difference isn’t really about what the code is made of. It’s about its role and how closely it’s tied to hardware. Firmware is written for one device or component and mostly works out of sight. Application software is written to run on many devices and is usually what users interact with.

Comparison table

Feature Firmware Software (OS and applications)
Main purpose Controls specific hardware: booting, low-level operation, device functions Provides features for users or manages a general-purpose computer
Tied to hardware? Usually written for one device, chip or board Usually runs on many devices that share an OS or platform
Where it’s stored Non-volatile memory on the device, typically flash Storage drives (SSD, hard disk), app stores, the cloud
How it runs Often runs straight from on-chip flash, or is loaded at power-on Loaded from storage into RAM by the operating system
User interaction Mostly invisible; little or no direct interaction Usually designed for direct interaction
Update frequency Less frequent, in many products Often frequent
Update risk A failed update can leave a device unable to start (“bricked”) A failed update can usually be reinstalled
Typical size Kilobytes to megabytes for many devices; larger on complex ones Megabytes to gigabytes
Common languages Mostly C and C++, with some assembly; Rust is growing Almost any language
Examples PC UEFI/BIOS, SSD controller code, microwave control code, router firmware Web browsers, office suites, games, desktop and mobile operating systems

These are general tendencies, not strict rules. There are exceptions in both columns, as covered below.

What is firmware?

Firmware is the program that lives inside a device and makes it do its job. A TV remote, a washing machine, a keyboard and a car’s engine controller all run firmware. So do many parts inside a computer, including the motherboard, the SSD, the network card and the graphics card.

Firmware usually does some or all of the following:

  • Starts the hardware when power is applied and sets up its components
  • Reads inputs, such as buttons, sensors or data from another chip
  • Controls outputs, such as motors, displays, LEDs or radio transmitters
  • Talks to other hardware using communication protocols
  • On a PC, prepares the system and hands control to the operating system

Where the term came from

The word “firmware” is generally credited to computer scientist Ascher Opler, who used it in 1967. He meant code that sat between hardware and software: harder to change than ordinary programs, but not fixed in the circuitry like hardware. At the time, that kind of code was often stored in memory that could only be written once.

Where firmware lives

Early firmware was stored in true read-only memory (ROM), so it was set at the factory. Later came EPROM and EEPROM, which could be erased and reprogrammed. Today most firmware is stored in flash memory. Flash keeps its contents without power and can be rewritten, which is why firmware can now be updated in the field.

In a microcontroller, the flash is on the same chip as the processor. In a PC, the UEFI firmware is stored on a flash chip on the motherboard.

What is software?

“Software” is the broad term for any instructions a computer runs. Strictly speaking, firmware is included. In everyday comparisons, though, “software” usually means two groups:

  • System software: operating systems such as Windows, macOS, Linux, Android and iOS, plus utilities and drivers
  • Application software: programs people use to get things done, such as browsers, email, games, photo editors and spreadsheets

This kind of software is usually stored on a drive, loaded into RAM when it runs, and managed by an operating system. It’s designed to run on many different machines, as long as they share a compatible operating system and processor architecture.

The key differences explained

Purpose and who interacts with it

Firmware exists to make hardware work. Most people never see it, except when a device asks to install an update. Application software exists to provide features to a user, so it’s built around a user interface.

How tightly it’s tied to the hardware

This is probably the most useful way to tell them apart. Firmware is written for particular hardware: a specific chip, its pin connections, its memory layout and the components on the board. Firmware for one printer model generally won’t work on a different model. Application software sits on top of an operating system that hides most hardware details, so the same app can run on thousands of different computers.

Storage and how it runs

Firmware is stored on the device itself in non-volatile memory. On a microcontroller it often runs directly from flash as soon as power is applied. Application software is installed onto a storage drive. The operating system loads it into RAM and schedules it to run alongside other programs.

Updates and the risk of bricking

You can update both, but the risks differ. If an app update fails, you can usually reinstall it. If a firmware update is interrupted, for example by a power cut halfway through, the device may be left without working startup code. It may then fail to boot. This is called bricking, as in the device is now as useful as a brick.

Many well-designed devices guard against this. They keep two copies of the firmware, check an update before switching to it, or keep a small recovery bootloader that is never overwritten.

Firmware also tends to be updated less often. That’s partly because of the risk and partly because many devices are expected to keep doing one job without change.

Size and languages

Firmware for small devices is often compact. A simple microcontroller may have only tens of kilobytes of flash for the whole program. Firmware for complex devices, such as routers, phones’ modem chips or PC motherboards, can be much larger.

Most firmware is written in C or C++, with some assembly language where very fine control is needed. Rust is increasingly used too. Application software can be written in almost any language, including Python, Java, JavaScript, C# and Swift.

Firmware in embedded systems

In embedded systems, “firmware” usually means the entire program running on the device. On a typical microcontroller there is no separate operating system and no application installed on top. The developer’s code, any libraries and any hardware drivers are compiled together into one binary file. That file is written into the chip’s flash memory.

For example, when you upload a sketch to an Arduino board, you’re writing firmware to its microcontroller. Our article on whether Arduino is an embedded system covers that in more detail. Some embedded firmware includes a small real-time operating system (RTOS), but it’s still built and flashed as a single image.

That’s why embedded developers are often called firmware engineers. Their work involves reading datasheets, configuring peripherals, handling interrupts, managing memory carefully and testing on real hardware. The differences between microcontrollers and microprocessors explain why the environment is so different from desktop programming.

Where the line blurs

The firmware/software split is a convention, not a hard boundary:

  • Firmware images can contain a full OS. Many routers, smart TVs and network storage devices run embedded Linux. Their “firmware update” replaces the operating system and the applications in one package.
  • Phones use the term loosely. A phone “firmware” update may include low-level code for the modem and other chips plus a new version of the operating system.
  • Some firmware is large and complex. Modern PC UEFI firmware includes drivers, network support, a graphical setup screen and security features.
  • Drivers sit in between. A device driver is OS software that talks to hardware. Many drivers also upload firmware into the device they control when the computer starts.

A practical rule: if the code is built for a specific piece of hardware and ships as part of that device, people usually call it firmware. If it’s installed by the user on a general-purpose platform, they usually call it software.

Examples of firmware and software in everyday devices

Device Firmware Software
Desktop or laptop PC UEFI/BIOS, SSD controller, keyboard controller, Wi-Fi card Windows, macOS or Linux; browsers, office apps, games
Smartphone Bootloader, modem (baseband) firmware, sensor and power controllers Android or iOS; apps from the app store
Home router Usually the whole system image, often embedded Linux Phone apps or web pages used to manage it
Microwave oven Control program on its microcontroller None for the user to install
Car Firmware on dozens of electronic control units (engine, brakes, airbags) Infotainment apps, phone integration
Smart thermostat Device firmware, often updated over Wi-Fi Companion phone app

How firmware gets onto a device and how it’s updated

There are several common methods, and many products support more than one:

  • At the factory: chips are programmed before or after they’re soldered onto the board.
  • Debug or programming interface: developers use a hardware probe, often over JTAG or SWD, to write firmware directly into a microcontroller and to debug it.
  • Bootloader: a small program kept in the chip that can receive new firmware over USB, serial or another link and write it to flash. Arduino uploads typically work this way.
  • Over the air (OTA): connected devices download an update over Wi-Fi, cellular or Bluetooth, check it, and install it, often with a fallback copy in case something goes wrong.
  • User-run update tools: for example, updating PC motherboard firmware from a file on a USB drive.

Frequently asked questions

Is firmware a type of software?

Yes. Firmware is software in the sense that it’s a set of instructions a processor executes. The separate name reflects its role: it’s built for specific hardware, stored on the device and mostly hidden from users.

Is BIOS firmware or software?

The BIOS, and its modern replacement UEFI, is firmware. It’s stored on a flash chip on the motherboard, it runs when the computer powers on, it prepares the hardware, and it starts the operating system.

Is an operating system firmware?

On a PC, no. The operating system is software installed on a drive. On many embedded devices, such as routers or smart TVs, the operating system is shipped as part of the firmware image, so in that context it’s often described as firmware.

Can firmware get viruses or malware?

It’s possible. Researchers have shown attacks against PC firmware, peripherals and IoT devices. Malicious firmware can be hard to detect and can survive reinstalling the operating system. Protections include signed firmware updates, secure boot and installing updates only from the manufacturer.

Should I install firmware updates?

In general, updates from the device’s manufacturer are worth installing, especially when they fix security problems. Follow the manufacturer’s instructions. Make sure power won’t be interrupted during the update (keep a laptop plugged in, for example), and don’t install firmware meant for a different model.

The complete guide to embedded basics

What Is an Embedded System? Definition, Features and Examples
Read the guide

Keep exploring

Guides from other topics

All topics →