d-Matrix Raptor NVLink Fusion rack: 144 XPUs, 2.3 TB 3D-DRAM, 7.2 PB/s bandwidth
firstadopter · x · 2026-10-13
A slide from a d-Matrix presentation shows the Raptor NVLink Fusion AI rack specs, targeting data-center-scale ultra-low-latency inference.
- 18 Raptor trays + 9 NVLink Switch trays per rack
- 144 Raptor XPUs total
- 2.3 TB 3D-DRAM memory capacity
- 7.2 PB/s 3D-DRAM memory bandwidth
Notably, d-Matrix is adopting NVIDIA's NVLink Fusion interconnect to cluster its digital inference chips at rack scale.
More from Infra
- Rodolfo Rosini on 2040: cheap compute everywhere, devices jumping orders of magnitude — MikePFrank · 2026-10-13
- AI inference chips are a $225bn market: Vaire on why startups avoid silicon and its Nvidia-adjacent plan — MikePFrank · 2026-10-13
- Vaire CEO Explains Reversible Computing Chips That Recycle Energy Like a Tide Mill — MikePFrank · 2026-10-13
- China's $165B AI buildout rests on $11B model revenue and 4% of Nvidia's chip output — demian_ai · 2026-10-13
- New llama.cpp release adds multimodal model support and Metal performance gains — vboykis · 2026-10-13
- Broadcom optical roadmap: TH6 hits 102.4T in 2026, next-gen OCI by 2028 — firstadopter · 2026-10-13