A small module sits on many hobbyist workbenches, priced about the same as a sandwich. It runs Wi-Fi, talks to sensors, and ends up inside doorbells, thermostats, and toys. Few people who solder one into a project think about the processor at its center, or the question of who owns the rules that processor follows. That question turns out to be one of the quieter power struggles in modern technology. It decides which companies can build their own chips, which countries can, and who pays a toll to do it.
Why Companies Build Their Own Chips
For most of computing history, buying a processor meant buying someone else’s design.
Photo by The largest technology firms have since decided they’d rather build their own.
The reason is partly performance. A chip tuned for one job tends to beat a general-purpose one at that job. Apple’s M-series processors, designed around its own software, regularly outperform comparable Intel and AMD parts on power efficiency. Amazon designs its Graviton server chips so its cloud business doesn’t have to negotiate pricing with outside vendors on every purchase.
There’s also a quieter motive. A company that can make its own silicon is harder to corner in a negotiation, and less exposed when a single supplier raises prices or runs short. For the average reader, the device in your hand increasingly reflects a business decision about independence, not just an engineering one.
What RISC-V Actually Is
One term does most of the work here: the instruction set architecture, or ISA.
Photo by An ISA is the contract between software and hardware. It defines the basic vocabulary a processor understands, the way a published grammar defines a language.
RISC-V is an open, royalty-free ISA, meaning any company can design a chip that follows its rules without paying a license fee. It began at UC Berkeley around 2010 and is now stewarded by a non-profit called RISC-V International. Companies are also free to customize and extend it [LinkedIn]. One reference puts it plainly: RISC-V is not a single CPU but a standard that devices follow [SEGGER].
Compare that to the dominant alternatives. ARM licenses its design to chip makers for royalties. The semiconductor IP business, which earns revenue through licensing, royalties, and customization services, runs entirely on this model [MarketsandMarkets]. RISC-V removes that recurring toll. It also removes the hand-holding a paid license includes.
The design is modular. A maker picks a base, then adds optional pieces for tasks like floating-point math or cryptography, scaling from tiny microcontrollers up toward supercomputers. For the general reader: RISC-V is a blueprint language anyone may use for free, not a product on a shelf.
The Politics Behind Open Silicon
Openness is precisely what makes RISC-V politically charged.
U.S. export controls can limit who receives ARM or x86 technology. An open standard with no single owner is far harder to restrict the same way. Chinese designers, including Alibaba’s chip unit, have shipped RISC-V processors partly to reduce dependence on licenses that could be cut off. That has prompted debate in Washington about whether open architectures should be treated differently under export law.
Incumbents have their own incentives. When an executive frames an open standard as a security concern, it’s worth reading the commercial interest underneath the claim.
Governance sits with RISC-V International, which moved its legal home to Switzerland in 2020 to signal that no single government steers the roadmap. That neutrality is both a strength and a coordination headache, since no one party can simply decide the future.
For the general reader: the fight over RISC-V is really a fight over who controls the global supply of chips.
Cross-Industry Adoption Patterns
Adoption is uneven, and that unevenness is the most honest part of the story.
Photo by The pattern echoes how open-source software spread: first into unglamorous plumbing, then slowly upward.
-
Embedded and IoT: Already mainstream. Espressif’s low-cost wireless modules and Western Digital’s storage controllers ship RISC-V silicon in volume, where cost and customization decide everything.
-
Data centers and phones: Growing, but ARM and x86 still lead, mostly because their software ecosystems are mature.
-
High-performance computing: Real but early. An independent evaluation found a significant performance gap between tested RISC-V chips and both ARM and x86-64 [arXiv].
The main friction is software. Using nonstandard RISC-V instructions can require a custom toolchain, custom application code, and often a custom kernel [Ubuntu]. Public money is closing that gap, with European and Indian programs funding RISC-V development as a matter of national sovereignty.
One market projection sees the RISC-V processor market growing from about $1.9 billion in 2026 to roughly $26 billion by 2034 [Straits].For the general reader: RISC-V is ready today for small devices, while its role in larger machines is still being built.
Return to that workbench module, the one cheaper than lunch. The interesting thing about it isn’t its price but its lineage: its processor follows a rulebook that no company owns and no government fully controls. Most silicon you touch carries a quiet license fee and a chain of permissions back to a single owner. This one doesn’t. That cheap board running Wi-Fi in a drawer is an early, working piece of an open-chip era that has already shipped, sitting unremarkably in everyday hardware.
🔖
- Straits Research, RISC-V processor market projection 2026 to 2034
- LinkedIn, RISC-V as an open and royalty-free architecture
- SEGGER Knowledge Base, RISC-V is a standard not a single CPU
- arXiv evaluation, performance gap between RISC-V and ARM and x86-64
- Ubuntu, RISC-V custom-instructions toolchain requirements
- MarketsandMarkets, semiconductor IP licensing and royalty model
