RISC-V Supreme

Massively parallel RISC-V at data-center scale — many small cores presented as one coherent operating story.


Domain

Compute

Stage

Research

Started

2026

Products


Many cores

Small RISC-V cores in parallel

One story

A coherent operating model

Data-center

The scale it targets

The problem

Parallelism is easy to buy, hard to operate.

Filling a data center with small, efficient RISC-V cores is now a purchasing decision. Operating thousands of them as one system is not — scheduling, memory, failure and observability were designed around a smaller number of larger machines.

Without a coherent operating story, massive parallelism turns into massive operational surface. The cores are cheap; the coordination is where the cost moves.


The approach

Make the fleet legible.

01

Many small cores

The design assumes a large population of small RISC-V cores as the substrate, rather than a handful of large processors — betting on parallel efficiency at data-center scale.


02

One coherent operating model

On top of that substrate sits a single operating story for scheduling, memory and failure, so the fleet is administered as one system instead of thousands of independent machines.


Where it stands

Research, stated plainly.

MilestoneStatusEvidence
Operating-model designIn progressInternal study
Small-scale clusterNext
Scheduling & failure modelNext
Data-center pilotNext

Figures on this page are targets and internal measurements at proof-of-concept stage. We share the methodology with anyone who wants to run it.


What's next

A cluster to argue against.

The near-term goal is a small cluster that makes the operating model concrete — enough cores to expose the scheduling, memory and failure questions the design has to answer.

RISC-V Supreme spins out of Research ByIQ if the operating model holds as the core count grows.