Nvidia plans to invest in d-Matrix, an AI inference accelerator startup, per The Information, which cited three people with knowledge of the deal. The move is part of Nvidia’s effort to make its technology compatible with rival chip designs: a month ago d-Matrix said it would adopt Nvidia’s NVLink Fusion interconnect to connect its Raptor processors with Nvidia hardware in rack-scale solutions.
D-Matrix, founded in 2019 and based in Santa Clara, builds AI inference accelerator cards, JetStream NICs, and Aviator software. It claims 10 times faster performance, three times lower cost, and three to five times better energy efficiency than GPUs. It closed a $275 million Series C in November 2025 at a $2 billion valuation and was reported in July to be targeting a $5 billion valuation in new financing.
NVLink Fusion, unveiled in May 2025, already counts Arm, Fujitsu, Intel, MediaTek, Qualcomm, SiFive, and Synopsys as partners. An investment in d-Matrix would be the latest in Nvidia’s strategy of financing the ecosystem around its interconnect, ensuring that even rival inference silicon runs through Nvidia’s plumbing.