D-Matrix Adopts NVIDIA NVLink Fusion for Rack-Scale XPU Deployment
AI Executive Summary
AI inference chipmaker d-Matrix announced it will integrate NVIDIA's NVLink Fusion to connect its next-generation Raptor XPUs to the broader NVIDIA AI infrastructure platform.
The collaboration leverages NVIDIA MGX rack architecture, scale-up networking, and Spectrum-X scale-out networking, allowing d-Matrix to incorporate Vera CPUs, ConnectX-9 SuperNICs, BlueField-4 DPUs, and Spectrum-X Ethernet into a unified cooling and rack system.
This integration provides d-Matrix with a lower-risk deployment path for ultralow-latency inference alongside GPU-based systems like the NVIDIA Vera Rubin NVL72.
Why It Matters
Strategic TakeawayAdopting NVLink Fusion allows custom silicon developers to bypass proprietary rack and network validation bottlenecks by plugging directly into a mature, standardized NVIDIA AI factory stack. This architectural synergy standardizes liquid-cooled rack designs across heterogeneous processors, bridging specialized inference XPUs with mainstream GPU and CPU ecosystems to optimize datacenter performance-per-watt.
Multi-Vector Implications
- TECHNICALd-Matrix will integrate NVIDIA NVLink scale-up and Spectrum-X scale-out networking with its Raptor XPUs, alongside Vera CPUs and BlueField-4 DPUs in MGX racks.
- MARKETCustom XPU vendors gain a lower-risk, accelerated path to AI factory deployment, reducing time-to-market by leveraging NVIDIA's pre-validated supply chain.
- GOVERNANCEDatacenters adopting unified MGX rack architectures must ensure power, liquid-cooling, and security compliance across multi-vendor XPU, CPU, and GPU deployments.
Strategic Outlook
12-18M HorizonOver the next 12-18 months, custom AI inference chipmakers will increasingly adopt open interconnect standards like NVLink Fusion to embed specialized accelerator directly into standardized NVIDIA MGX hyperscale racks. This trend will drive the deployment of heterogenous, disaggregated inference systems where XPUs operate seamlessly alongside NVIDIA GPU and CPUs in unified AI factories.
Referenced Coverage & Sources
Read the full coverage below for original reporting, technical benchmarks, and complete primary source details.
Physical AI Takes the Wheel: How the World's Robotaxi Leaders Are Building with NVIDIA Technologies
The global robotaxi market - physical AI's first commercial breakthrough - is projected to reach $400 billion by 2035, with over 6 million commercial.
NVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference V6.1 Debut
System performance, efficient infrastructure scaling and continuous software optimization are key levers that determine AI inference economics.
Emerald AI, Google and NVIDIA Launch Alliance to Advance Flexible AI Data Centers
AI factories are the infrastructure of the intelligence era.
From Megawatts to Tokens: How NVIDIA Maximizes AI Factory Production
On a sweltering August evening in Silicon Valley, as the sun dropped and air conditioning loads spiked, Silicon Valley Power sent a signal to an AI factory.
Inference
Inference is the process of using a trained AI model to make predictions or generate text based on new inputs. During inference, data flows forward through the neural network to produce an output, without modifying the model's weights.
NVIDIA
NVIDIA is a pioneer of GPU computing, dominating the hardware market for AI acceleration, training, and inference with its high-performance Hopper and Blackwell architectures.
AI Infrastructure
AI Infrastructure refers to the hardware compute, vector databases, network fabrics, orchestration layers, and MLOps platforms required to train, evaluate, and serve AI models at scale.
Explore technical glossaries, weekly market briefings, and editorial research articles related to this story:
Get top 5 high-signal AI news, venture funding rounds, and research papers auto-routed to dedicated channels every 3 hours.