Asterinas
A memory-safe, Linux-compatible OS kernel written in Rust, aiming to be a production-grade Linux alternative.
Asterinas is a general-purpose operating system kernel written from scratch in Rust with a Linux-compatible system call interface. It implements more than 240 Linux system calls, enough for Linux binaries and container images to run on it, and it is built by an independent, community-led project with over 100 contributors and roughly 250,000 lines of code. Rather than incrementally adding Rust to the Linux kernel, the project starts over so that every design decision can assume a memory-safe language.
Its defining idea is the framekernel architecture: the whole OS lives in a single address space like a monolithic kernel, but only a small framework called OSTD is allowed to use unsafe Rust. System calls, filesystems and device drivers must be written in safe Rust, which shrinks the memory-safety trusted computing base while keeping the performance of direct function calls between kernel components. On top of that the project ships OSDK, a toolkit for building, running and testing Rust kernels, and a novel memory-management design called CortenMM. x86-64 is the Tier 1 platform; 64-bit Arm, RISC-V and LoongArch are supported at lower tiers, and Intel TDX confidential VMs are a Tier 2 target. Version 0.19.0 added a DRM graphics subsystem that runs Xorg with Mesa software OpenGL, hypervisor support in OSTD, a Rust-native device model, a rewritten overlayfs and a more complete virtio-fs driver, aimed at using Asterinas as the guest kernel for Kata Containers and Confidential Containers.
The project positions itself against commercially supported Linux and commercial RTOSes: it is MPL-2.0 licensed, with no subscription, support contract or safety certification attached. Its experimental distribution, AsterNixOS, is NixOS with Linux swapped for Asterinas and is explicitly not ready for production use and not affiliated with the NixOS Foundation. The stated 2026 goal is production readiness for standard and confidential virtual machines on x86-64, with data centers, autonomous vehicles and embodied AI as the target deployments. You can try it from the AsterNixOS installer ISO or build the kernel and distribution yourself inside the project's Docker development environment.
