The Crowd Supply Baochip Dabao update is the kind of post that shows up in a Lobsters feed and makes you sit back for a moment. It is an origin story from a hardware maker who decided that using someone else’s chip was no longer sufficient, and who went ahead and started building their own. The post covers the usual ground of these announcements, the backstory, the motivation, the moment of deciding to commit, but what it is really describing is a broader shift in what is now possible for independent hardware developers.
That shift did not happen overnight. It is the product of roughly a decade of infrastructure work by the open silicon community.
The Infrastructure That Made This Possible
For most of computing history, chip design was the exclusive domain of companies with eight-figure engineering budgets and access to proprietary EDA tooling. TSMC and similar foundries required commitments measured in wafer lots. The software to design the chips, tools from Synopsys and Cadence, cost more annually than most startups raise in seed rounds.
That started changing around 2020 when Google partnered with SkyWater Technology to release the Sky130 process design kit as open source. For the first time, a production-grade semiconductor process node (130nm) was fully documented and accessible to anyone. SkyWater’s fabrication was made available to independent designers through the efabless platform’s Multi-Project Wafer (MPW) shuttle program, where multiple designs are tiled onto a single wafer to share fabrication costs and reduce per-designer cost to a few thousand dollars.
The companion toolchain, OpenLane, brought together open-source EDA tools including Yosys for synthesis, OpenROAD for place-and-route, and Magic for layout verification, into a reproducible RTL-to-GDS flow. The Caravel harness provided a pre-built SoC wrapper with GPIO, UART, SPI, and a management RISC-V core, so designers could focus their effort on the custom logic blocks they actually wanted to build rather than rebuilding the same I/O scaffolding from scratch.
By 2022 and 2023, a wave of MPW shuttle tapeouts from academic groups, hobbyists, and small companies demonstrated that this pipeline actually worked. Chips came back from the fab and ran code. The success rate was uneven, but the existence of working silicon at all, designed with free tools and fabricated on a shared wafer run, was a meaningful threshold.
RISC-V as the Architecture of Choice
Running in parallel with the open PDK work, the RISC-V ISA emerged as the architecture of choice for custom silicon precisely because it imposes no licensing costs or restrictions. Unlike ARM, which requires a license agreement for even the simplest Cortex-M core, RISC-V lets anyone implement the ISA, add custom extensions, and tape out a core without a royalty conversation.
The ecosystem around RISC-V has matured substantially. VexRiscv, written in SpinalHDL, is a configurable, pipelined implementation that has been used in dozens of research and production designs. PicoRV32 is a more minimalist option suitable for resource-constrained FPGAs or simple custom silicon. The CVA6 core from the OpenHW Group implements the full application-class RV64 ISA and runs Linux. All of these are available under permissive licenses, synthesizable with open tools, and well-documented.
This combination gave independent hardware makers a path that did not previously exist: design a custom chip around an open architecture, using open tools, on a publicly documented process node, fabricated at a cost that is within the range of a crowdfunding campaign.
Where Dabao Fits
The Baochip Dabao project represents one developer deciding to take this path seriously. The creator’s framing of “why now” and “how it came about” is honest about what makes this kind of project both harder and easier than it looks from the outside.
Harder, because the gap between “I have a working simulation” and “I have a chip that boots” is enormous. Timing closure, physical verification, ESD protection, power domain planning, and the mechanics of actually testing a fresh die are all specializations unto themselves. Even with modern tooling, a solo designer is navigating problems that teams of ten engineers used to handle.
Easier, because the tooling has genuinely improved. OpenLane 2, restructured around a plugin architecture that makes the flow more composable and debuggable, has lowered the friction on iterating through design changes. Formal verification tools like SymbiYosys are open source and increasingly practical for catching RTL bugs before committing to silicon. The documentation around Sky130 and the efabless process has expanded with each shuttle run as the community accumulates institutional knowledge.
There is also a community dimension that the early open silicon era lacked. The Zero to ASIC course by Matt Venn has graduated hundreds of designers through the full process of getting a design from RTL to tapeout. The course community maintains shared infrastructure, shared documentation, and shared experience about what typically goes wrong at each stage. A creator like Baochip is almost certainly drawing on this accumulated knowledge rather than solving every problem from first principles.
The Chinese Hardware Context
The name “Baochip” and “Dabao” carry a Chinese inflection worth noting. China’s domestic semiconductor ecosystem has been under sustained pressure since US export restrictions on advanced nodes began in 2019 and tightened through 2022. The policy response has pushed significant capital into domestic chip companies, but it has also energized a broader cultural interest in hardware independence among Chinese engineers and makers.
Projects like the WCH CH32V series, which produced RISC-V microcontrollers cheap enough to undercut STM32 parts on price while offering competitive peripherals, or the Sipeed Tang Nano FPGA boards, reflect an ecosystem where bottom-up hardware development is taken seriously. An indie creator building custom silicon in this environment is not an anomaly; it is a coherent response to a set of pressures and opportunities that are specific to this moment in Chinese hardware culture.
The tension between dependence on Western-designed chips and the desire for sovereignty has filtered down from national policy to individual makers in interesting ways. Where a Western hobbyist might reach for custom silicon out of curiosity or craft, a Chinese maker might have additional motivation rooted in supply chain anxiety. Both end up in the same place: designing their own thing.
What Comes After the Campaign
The Crowd Supply model works well for this kind of project because the platform has a history with open hardware and with campaigns that are as much about sharing a design process as selling units. The typical campaign of this type succeeds not on volume but on community: the people who back it are also the people most likely to contribute test cases, port firmware, write documentation, and push the design forward after the initial run.
That community dynamic is what separates a hardware project with a story from a hardware product. The Baochip update is written by someone who wants to share the work, not just sell the output. The best Crowd Supply campaigns are the ones where the interesting material is in the updates and the forum threads, not just the delivery confirmation. Dabao looks like one of those.