Physix Frontier · News Briefing Card (IT Home · Oct 9, 2026)
Nvidia to Invest in AI Inference ASIC Firm d-Matrix
KEY FACTS
- The Information reported that Nvidia plans to invest in AI inference ASIC startup d-Matrix.
- d-Matrix's next-generation inference XPU will support Nvidia's NVLink Fusion ecosystem.
- The XPU will also adopt Nvidia's MGX universal rack reference architecture.
- Nvidia had previously completed a de facto acquisition of Groq, and this is another move.
- The two companies had already established a technical partnership before the investment news.
PHYSIX OBSERVATION
While developing its own GPUs, Nvidia is bringing inference ASIC upstarts into its investment portfolio, effectively locking down the inference race through both capital and technical standards. For companies like d-Matrix, connecting to NVLink Fusion and MGX means picking a side, and it also means narrower bargaining room. The signal is clear: independent paths for inference chips are being absorbed into Nvidia's ecosystem, and it will become increasingly difficult for latecomers to go it alone.
Source: IT Home report
Physix Frontier