OXMIQ offers a modular GPU stack that separates silicon, software, and interconnect layers, enabling RISC‑V‑based nano‑agent cores, chiplet‑native designs, and near‑memory compute. Its unified software suite (OxCapsule, OxPython, OxCore, OxQuilt) allows native Python execution and seamless code portability from edge devices to cloud, while licensable IP lets OEMs and AI hardware developers customize acceleration for agent‑centric workloads.
Funding
$20M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Current GPU architectures are tightly coupled, proprietary, and optimized for fixed workloads, limiting flexibility for emerging AI agents, heterogeneous compute, and rapid innovation. This rigidity hampers developers who need to move code across edge devices, data centers, and renewable-powered compute platforms without extensive rewrites.
Solution
Oxmiq provides a modular GPU stack that separates silicon, software, and interconnect layers, enabling open, RISC‑V‑based nano‑agents, chiplet‑native designs, and near‑memory compute. The platform supports native Python and other high‑level languages, allowing code to run unchanged from edge to cloud. By offering licensable silicon IP and a unified software stack—including OxCapsule, OxPython, OxCore, and OxQuilt—Oxmiq lets hardware designers and AI developers customize acceleration, integrate heterogeneous containers, and scale from single devices to gigawatt‑scale renewable data centers. The open standards and dynamic heterogeneous containers reduce integration effort and accelerate time‑to‑market for AI workloads.
Target Audience
Primary customers are semiconductor OEMs, data‑center builders, and AI hardware developers seeking customizable, high‑performance GPU solutions for agent‑centric and multimodal AI workloads.
Features
- RISC‑V‑based nano‑agent cores embedded in silicon for fine‑grained, programmable acceleration
- Chiplet‑native architecture supporting scalable SIMT and tensor units across heterogeneous containers
- Near‑and‑in‑memory compute blocks to minimize data movement and improve energy efficiency
- Full‑stack software suite (OxCapsule, OxPython, OxCore, OxQuilt) enabling native Python execution and seamless code portability
- Open interconnect standards for modular integration of compute, storage, and cooling subsystems
- Licensable IP model allowing OEMs and system integrators to tailor GPU designs to specific AI agent workloads