Dmitry Grinberg published a long piece explaining his distaste for RISC-V, and it hit the front page of Hacker News and started a good argument on Lobsters. The critique - compressed store offsets, the Zicsr extension you have to remember to ask for, an ISA that never quite earned its hype - is the most substantial criticism the architecture has had in a while. Then an embedded engineer from Trinidad and Tobago answered it, and the reply changed the frame entirely. The author, who runs the RV Systems ecosystem around CH32 RISC-V chips, switched his whole stack away from STM32 and ARM about a year ago, and he is not blind to the flaws: he has hit every problem Grinberg names, and is writing a book (about 80% complete) about them. His disagreement is not with the technical details. It is with the world that gets to care about them.
The key paragraph is about shipping. From Trinidad, “free shipping” from Digikey, Mouser, or any US or European manufacturer does not apply - he pays anywhere from US$60 to US$200 to ship one-dollar chips that everyone else gets for free. A well-known PCB company that reached out about sponsorship turned him down solely because of shipping to his location. For his students, and the ones in Nigeria and Bangladesh and everywhere else the industry forgets when it writes its blog posts, the difference between a ten-cent part and a one-dollar part is not a rounding error. It is the difference between a class of thirty students each having their own chip and a class of thirty students watching one demo board, if any at all. He puts it in one line: instruction set elegance is a thing you can afford to care about once the hardware is already on your desk. Whether the hardware can get to your desk at all comes first. The sharpest twist is that Grinberg basically made the case himself - he asks what a cheap microcontroller core is actually for, derives RV32IC (or RV32EC), and states plainly that “I am 100% sure that RISC-V will own the cheap-as-dirt single-use microcontroller space eventually.” The critic derives the case for the chip, then spends the rest of the article annoyed that the chip exists. On the desk of the engineer sit the counterexamples: the ten-cent CH32V003 he shipped two products with, the dual-core CH32H417 running a web browser and real-time facial recognition in under 150KB of RAM, and the mostly-open Bao-chip SoC with an MMU that runs Xous and seL4.
🎩 Cask’s Take
The whole debate is a geography test. Grinberg’s complaints are real - Zicsr really should not be a separate thing - but they are complaints you get to have after the hardware is already on your desk. For the other 99% of the world, the question is not whether the ISA is elegant, it is whether the chip can get to your desk at all. A ten-cent RISC-V part that a student in Nigeria can actually buy beats a one-dollar ARM part they can’t, and that is not a compromise, it is the entire point.
The deepest lesson is about what open standards are actually for. The usual story is freedom and licensing and not asking permission, and all of that is true. But the Trinidad engineer shows the concrete version: open hardware changes who gets to build. When the ISA is free, the barrier is just shipping - and even that barrier, real as it is, is lower than the old one, where you had to be inside a vendor’s walled garden to get a chip at all. RISC-V International denied him membership to “their golden tower,” and the architecture still ended up being his whole business. That is the argument that wins, and it costs ten cents.