The momentum behind RISC-V is often framed as a technical fight with ARM. The more important difference is the business model behind each instruction set.
Every modern processor speaks an instruction set, the basic vocabulary of commands the chip understands. For years the conversation about instruction sets centered on one dominant architecture in mobile and embedded devices and another in personal computers and servers. Now a third name comes up constantly: RISC-V. The momentum behind it is real, but the way that momentum is usually explained, as a head-to-head technical contest, misses the point. The decisive difference is not the engineering. It is the licensing.
What an instruction set actually is#
An instruction set is the agreement between hardware and software about what the chip can do and how to ask it. Software is compiled to target a specific instruction set; a chip is designed to implement one. This is why instruction sets are sticky. Once a large body of software targets one, moving away is costly, because everything has to be recompiled, retested, and in some cases rewritten.
That stickiness is the whole game. A new instruction set is not competitive on day one no matter how clean its design, because the software ecosystem has to grow around it. This is the hill RISC-V has to climb, and it is a steep one.
The licensing split#
Here is the difference that matters. The established architecture is owned by a company that licenses it. To build a chip using it, you pay, and you operate within the terms of that license. RISC-V is an open standard. The base specification is free to use, and anyone can design a chip around it without paying a license fee for the instruction set itself.
That changes the incentives for chip designers in concrete ways:
Discussion