GigaIO has sold its SuperNODE platform and FabreX PCIe Gen 5-based AI fabric to d-Matrix, marking a strategic transition as the company pivots toward edge computing solutions.
The sale builds on a year-long collaboration between the two companies, during which d-Matrix integrated its Corsair inference platform with GigaIO’s SuperNODE architecture. The combined solution enabled scalable, rack-level deployments capable of supporting dozens of accelerators within a single node, addressing the growing demand for low-latency AI inference in data center environments.
With the acquisition, d-Matrix also gains access to GigaIO’s engineering expertise in rack-scale system design, strengthening its ability to deliver integrated, high-performance AI inference platforms. The move positions d-Matrix to expand its capabilities in deploying complete, production-ready systems tailored for enterprise and hyperscale environments.
GigaIO’s FabreX technology, based on PCIe Gen 5, has been central to enabling high-speed interconnects between compute and accelerator resources. Its integration into d-Matrix’s systems is expected to enhance data throughput and reduce latency, key factors for real-time AI workloads.
Following the divestment, GigaIO will focus on advancing edge AI computing. The company’s flagship offering, Gryf, developed in collaboration with SourceCode, represents a new approach to portable high-performance computing. Designed as a compact, suitcase-sized AI supercomputer, Gryf delivers datacenter-class capabilities in a field-deployable form factor.
The shift reflects increasing demand for localized processing, as organizations seek to analyze data closer to its source to minimize latency and improve operational efficiency. Industries such as defense, energy, and media are among the early adopters exploring such solutions for real-time analytics and decision-making.
GigaIO stated that Gryf enables dynamic deployment of applications in the field, allowing users to process critical data without relying on centralized infrastructure. The system’s rugged design and scalability make it suitable for environments where traditional data center resources are not practical.
The transaction highlights a broader industry trend toward specialized AI infrastructure, where companies are aligning their strategies around either centralized data center performance or distributed edge intelligence.
