Codasip announced in early April that it is pivoting toward cyber-resilient SoC architectures and divesting its low-end RISC-V processor design business to an undisclosed U.S. public semiconductor company. That acquirer also picks up a broad license to Codasip Studio, the EDA toolchain used for rapid processor customization. For the EU's DARE project -- a 240M euro initiative to develop RISC-V processors and accelerators for European data centers -- this is a significant disruption to both the architectural roadmap and the tool ecosystem.
The DARE project is explicitly about European digital autonomy. RISC-V was the chosen architecture precisely because it is open and not controlled by U.S. IP. Now one of its key partners has sold its core RISC-V design business and the associated EDA tooling to a U.S. company. The irony is sharp: an initiative designed to reduce dependency on U.S. technology has just had its foundational tool license transferred to an American acquirer, identity undisclosed.
Barcelona Supercomputing Center director Mateo Valero was direct with EE Times: "It is very concerning that a company such as Codasip has severe difficulties." EuroHPC is aware and exploring alternatives, but finding a European company with equivalent RISC-V processor customization capability and EDA tooling is not a fast process. The CHERI security pivot Codasip is making is technically interesting -- building capability-based security into the ISA level rather than patching it in software -- but it does not help DARE's near-term timeline.
The broader signal here is about the fragility of European semiconductor sovereignty efforts when they depend on commercial entities that face their own market pressures. The DARE project had the right instinct about open ISA independence. The execution exposed how quickly a single partner's commercial decision can cascade into a project-level crisis. Any hardware team building on DARE-adjacent open RISC-V tooling should be watching this closely.