Nvidia (NVDA) is collaborating with artificial intelligence chip company d-Matrix to build systems designed for ultra-low latency AI inference, d-Matrix said Thursday.
Financial terms were not provided.
D-Matrix said it will integrate its Raptor chips into Nvidia's rack systems using NVLink Fusion to enable AI labs, hyperscalers and neoclouds to offer premium-level token services.
The company said it expects the first systems to become available in Q4 of 2027.
Shares of Nvidia were down 2.2% in Thursday trading.
Price: 218.73, Change: -4.94, Percent Change: -2.21