From RuntimeWire · RuntimeWire · · 30s
d-Matrix launches 1,024-XPU inference clusters
d-Matrix turns inference chips into 1,024-XPU data-center clusters.
In brief
d-Matrix turns inference chips into 1,024-XPU data-center clusters. d-Matrix sells complete inference clusters for data centers, extending Sid Sheth's 2019 bet on inference compute. Originally reported by RuntimeWire.
d-matrix
d-matrix turns inference chips into 1,024-xpu data-center clusters.
sid sheth
sid sheth, the co-founder and ceo of d-matrix, spent years arguing inference would become a larger compute opportunity than training.
why it matters
d-matrix sells complete inference clusters rather than accelerators alone, extending sid sheth's 2019 bet into data-center infrastructure.
The one thing
d-Matrix sells complete inference clusters for data centers, extending Sid Sheth's 2019 bet on inference compute.
