SANTA CLARA, Sept 10: AI chip startup d-Matrix has announced plans to integrate its processors directly into Nvidia’s data centre systems, seeking to meet growing demand for faster and more efficient artificial intelligence applications.
The company said it will adopt Nvidia’s NVLink Fusion technology to connect its Raptor processors with Nvidia’s data centre platforms. The combined systems are designed primarily for AI inference, the process through which trained models respond to user requests and perform tasks in real time.
Unlike AI model training, which requires large amounts of computing power to develop models, inference focuses on running those systems efficiently after they have been trained. Demand for this capability is increasing as AI-powered applications become more widely used.
d-Matrix said its Raptor chips are being developed for applications where low response times are important, including chatbots, coding assistants and voice agents.
The Nvidia compatible systems are expected to become available in 2027, while the final design phase of the Raptor processors is scheduled to be completed by the end of 2026.
The startup is also working with connectivity company Astera Labs on customised high-speed data pathways designed to improve communication between components inside the systems.
Financial details of the partnership were not disclosed. d-Matrix, which is based in California, has received backing from Microsoft and was valued at around $2 billion after raising $450 million last year.
The development highlights the growing technology industry’s shift toward specialised hardware designed to handle the expanding volume of AI workloads beyond the training stage.