FPGAs

Explainer

FPGAs explained

An FPGA, or field-programmable gate array, is a chip whose internal wiring can be changed after it is made. It contains a grid of small logic blocks, memory blocks and programmable connections. Loading a new configuration file, called a bitstream, turns it into a different digital circuit.

Engineers describe the circuit in a hardware description language such as Verilog or VHDL, or in a higher-level tool. Software then maps it onto the chip. The same FPGA can act as a network card today and a video processor tomorrow.

How it differs from the usual approach

A CPU runs software on fixed hardware. An FPGA changes the hardware itself to match the task, so it can process data in parallel with very predictable timing. It is less efficient than an ASIC at the same job and harder to program than a CPU. For low volumes, fast-changing standards or tight latency needs, it is often the practical choice.

Where it is used

• Prototyping ASICs before they are manufactured.
• Telecom and network equipment that must follow changing standards.
• Low-latency tasks such as financial trading and packet processing.
• Industrial control, aerospace and defense electronics.
• Data center acceleration, for example in network cards and some cloud services.

What to watch

• AMD completed its purchase of Xilinx in 2022. Altera, Intel's former FPGA unit, became a separate company in 2025 when Silver Lake bought 51 percent of it. Intel kept 49 percent.
• Other makers include Lattice, Microchip and Achronix, and some firms sell FPGA fabric as intellectual property to embed in other chips, called eFPGA.
• FPGAs increasingly combine programmable logic with CPU cores and AI engines in one device.
• Open-source toolchains for hardware design are improving, though vendor tools still dominate.

Frequently asked questions