D
d-Matrix builds specialised AI inference hardware using digital in-memory compute to reduce data movement and improve performance for generative AI workloads.
Tensordyne
Tensordyne builds custom silicon, hardware and software for low-power, high-performance multimodal generative AI inference, including its flagship Napier inference system.
Tensordyne designs and builds high-performance, low-power generative AI inference systems. The company applies novel AI research techniques to create custom silicon, hardware and software aimed at accelerating multimodal generative AI inference at scale. Its technical work spans logarithmic mathematics for AI, scale-up networking architecture, custom silicon and hardware design, and software development for low-power, high-performance computing.
Napier is the company's flagship AI inference system. It combines purpose-built silicon with logarithmic mathematics and an optimised scale-up networking architecture, and is positioned as a new class of inference technology designed for speed and cost efficiency. The approach reflects a research-led focus, drawing on recent AI research to inform both the underlying mathematics and the system architecture.
Tensordyne is headquartered in Sunnyvale, California, with an office in Munich, Germany. It serves hyperscaler and neocloud data centres, application layer companies, foundation model builders and enterprises. The company states a focus on safe and sustainable AI inference solutions and on working with customers at the leading edge of AI.
Open jobs
No open jobs right now.