Kora Systems Unveils AI‑Native Linux Desktop OS to Challenge Windows

Kora Systems, the new venture from Cyanogen co‑founder Kirt McMaster, has revealed a Linux‑based desktop operating system that integrates large language models and autonomous agents directly into the user environment. The OS, simply called Kora, lets AI agents operate across applications, access local files, and retain context between tasks without being confined to a single app window.

Slated for a 2027 release, Kora will run on existing Intel and AMD PCs as well as Qualcomm Snapdragon devices. A team of roughly 50 engineers is building the platform, leveraging AI coding assistants such as OpenAI Codex and Anthropic Claude, which McMaster says have accelerated development by at least fivefold over the past six months.

Microsoft’s Windows still commands about 77 % of the global desktop market, but McMaster is not aiming for mass adoption. He estimates 45–50 million heavy AI users today, a figure he expects to surpass 200 million by 2033. Kora’s business model focuses on this high‑value segment willing to pay for premium model access and agent services.

Why it matters for GPU and AI infrastructure

• OS‑level agents require low‑latency, local inference – driving demand for on‑device GPU compute.
• Support for heterogeneous hardware (x86 and Arm) encourages optimized driver stacks and runtime libraries.
• An open, Linux‑based foundation lets cloud providers and OEMs tailor GPU scheduling for AI workloads.

McMaster’s experience at Cyanogen taught him the danger of depending on the incumbent’s services. Kora is built from the ground up to avoid reliance on any single cloud provider, using a modular “Chord” chat interface that can swap between local and frontier models. The project evolved from Playtron’s GameOS, giving the team a three‑year head start on a stable Linux core before pivoting to AI‑centric features.