The Grus Arch Roadmap
Embedded first, general-purpose later. One substrate, two compiler profiles.
The idea
We don’t chase an everything-machine from day one. The approach is to start where the value is concrete and dependable, prove it in a small, constrained setting, and grow the substrate from there.
The foundation is shared: one computing substrate carrying two compiler profiles — one tuned for embedded, one aimed at more general use. We widen the sign as we climb each rung, never before.
Three rungs
First rung — IP core / coprocessor
We begin as a coprocessor (IP core) sitting next to a host MCU / SoC. Existing software assets stay on the host, while the regular, well-structured work is offloaded. The embedded and edge domain — where power and area are tight and predictable behavior matters most — is where the advantage of surveying dataflow statically pays off most directly.
Second rung — Standalone embedded controller
Building on the first rung, we move to a standalone embedded controller that runs without a host.
Third rung — Toward general purpose
Extend the established substrate toward broader computation. This is a direction of vision; it takes concrete form as the R&D advances.
Technical details for each rung will be published in stages at the right time. If you’re interested in collaboration or evaluation, please get in touch.