NVIDIA Acquiesces To Being The Plumbing Of AI

In a predictable display of hardware pragmatism, AI inference chipmaker d-Matrix has announced it will adopt NVLink Fusion to connect its upcoming Raptor XPUs to NVIDIA's ubiquitous infrastructure platform. Rather than building a parallel universe, d-Matrix is simply bolting its custom silicon directly onto the prevailing monopoly's plumbing, integrating with NVIDIA MGX architecture and Spectrum-X networking.

• Adopting NVLink Fusion allows d-Matrix to plug straight into the dominant enterprise hardware stack without forcing buyers to rethink their entire data centre architecture.
• By hitching Raptor XPUs to NVIDIA's scale-up and scale-out networking, d-Matrix avoids the lonely fate of building a brilliant chip that nobody can actually plug into anything.
• This partnership further cements NVIDIA not just as a chip designer, but as the mandatory tollbooth operator for every aspiring AI hardware vendor on the market.

Read the original: d-Matrix Adopts NVIDIA NVLink Fusion for Rack-Scale XPU Deployment