d-Matrix AI chip startup d-Matrix announced plans to connect its next-generation Raptor AI inference processors to NVIDIA’s infrastructure using NVLink Fusion, while also working with Astera Labs on high-speed data paths. The partnerships give the Santa Clara-based chip startup an infrastructure path for deploying its processors in large-scale AI systems.
Key Takeaways
- d-Matrix will use NVIDIA NVLink Fusion to connect its Raptor XPUs to NVIDIA’s AI infrastructure platform.
- Raptor processors are designed for AI inference workloads, including chatbots, coding assistants and voice agents.
- d-Matrix plans to complete the final Raptor design by the end of 2026.
- Integrated systems using Raptor are expected to become available in 2027.
- d-Matrix is also working with Astera Labs on high-speed data paths for its systems.
d-Matrix Adopts NVIDIA NVLink Fusion for Raptor
d-Matrix is adopting NVIDIA NVLink Fusion to connect its Raptor processors with NVIDIA’s AI infrastructure. The arrangement places the startup’s next-generation inference hardware within an infrastructure platform designed to support AI systems.
The partnership centers on Raptor XPUs, which d-Matrix is developing for AI inference workloads. The processors are intended to work within larger AI computing systems rather than operate as standalone components.
NVLink Fusion provides the connection between d-Matrix’s processors and NVIDIA infrastructure. For the startup, the technology forms part of its planned approach to deploying Raptor in integrated AI systems.
The development represents a specific hardware and infrastructure partnership between d-Matrix and NVIDIA. Rather than focusing only on processor development, d-Matrix is also establishing connectivity with infrastructure used for AI computing.
Other AI startups have also been building businesses around infrastructure and enterprise deployment. For example, AI infrastructure startup expansion has involved software designed to manage AI workloads within existing cloud environments, illustrating a separate approach to supporting enterprise AI systems.
The planned integration is tied to the company’s next-generation Raptor processors. d-Matrix expects to complete the final Raptor design by the end of 2026.
Raptor’s Role in AI Inference
Raptor processors are designed for inference, the stage of AI computing associated with running AI models and producing responses or outputs.
The Raptor platform is intended for workloads that include chatbots, coding assistants and voice agents. These applications require computing systems capable of processing AI workloads after models have been developed.
d-Matrix’s focus on inference distinguishes the Raptor processors by the workloads they are designed to address. The company’s current development plans therefore center on hardware intended for the execution side of AI applications.
The processors are being developed as part of an integrated system strategy. NVIDIA connectivity and additional high-speed data-path work with Astera Labs are both part of that approach.
NVIDIA Infrastructure Supports Raptor Deployment
NVIDIA NVLink Fusion is central to d-Matrix’s planned integration of Raptor into NVIDIA’s AI infrastructure. The technology provides the connection required for the startup’s processors to operate within that infrastructure.
The arrangement gives Raptor a defined path toward integration with NVIDIA systems. d-Matrix is developing its processors while preparing them for use within a larger infrastructure environment.
The planned deployment involves Raptor XPUs rather than a general-purpose processor platform. Their stated purpose is AI inference, including workloads associated with conversational systems, coding assistants and voice-based AI agents.
The infrastructure relationship is also connected to d-Matrix’s development schedule. The company plans to complete the final Raptor design by the end of 2026, followed by integrated systems using the processors.
Raptor-based integrated systems are expected to become available in 2027. The schedule establishes the next major stage of the startup’s hardware development and deployment plans.
The NVIDIA relationship therefore covers more than processor development. It connects d-Matrix’s hardware plans with an established AI infrastructure platform and provides the technical framework for integrating Raptor into larger systems.
The role of infrastructure partnerships is also evident among other AI startups. AI startups using external infrastructure have described using established computing platforms rather than building every infrastructure component internally, providing another example of partnerships supporting small technical teams.
NVIDIA MGX Infrastructure Integration
The Raptor processors are being positioned for integration into NVIDIA infrastructure through NVLink Fusion. This connection is part of d-Matrix’s strategy for bringing its inference hardware into integrated AI systems.
The planned infrastructure integration also places the processor within a broader system architecture. Raptor is being developed for specific inference workloads while the NVIDIA connection addresses the infrastructure required to deploy those processors.
The approach links processor development with system compatibility. d-Matrix’s development schedule consequently includes both the completion of its Raptor design and preparation for integrated systems.
d-Matrix Works With Astera Labs on Connectivity
d-Matrix is also working with Astera Labs on high-speed data paths for its systems. The connectivity work adds another partnership to the startup’s Raptor development strategy.
High-speed data paths are part of the infrastructure required for AI computing systems. d-Matrix’s work with Astera Labs therefore complements its planned connection to NVIDIA infrastructure.
The two partnerships address different parts of the system surrounding Raptor. NVIDIA NVLink Fusion provides the connection between Raptor processors and NVIDIA’s AI infrastructure, while Astera Labs is involved in high-speed data paths.
The combined work supports d-Matrix’s effort to develop Raptor as part of an integrated AI system. The startup is addressing processor design alongside the connectivity needed for deployment.
For Raptor, system connectivity is directly linked to its intended use in AI inference. The processors are designed for workloads including chatbots, coding assistants and voice agents, making the surrounding infrastructure an important component of the planned system.
d-Matrix’s partnership with Astera Labs therefore forms part of the same development effort as its NVIDIA relationship. Both involve infrastructure required to connect and deploy the company’s AI inference processors.
The partnerships also give the Raptor platform multiple infrastructure components as development progresses toward the company’s planned 2027 availability.
Other technology startups have similarly used partnerships to support specialized AI products. AI startups adapting business strategies have pursued relationships with cloud infrastructure providers and enterprise technology companies to improve compatibility and accelerate AI deployment.
Raptor Development Moves Toward 2027 Availability
d-Matrix plans to complete the final Raptor design by the end of 2026. The design milestone precedes the expected availability of integrated systems using the processors in 2027.
The development schedule establishes a sequence from processor design to integrated system availability. NVIDIA NVLink Fusion and the work with Astera Labs are components of that process.
Raptor’s intended applications remain centered on AI inference. The processors are designed for systems handling chatbots, coding assistants and voice agents, giving the platform a defined workload focus.
The company’s infrastructure partnerships are tied directly to that deployment plan. NVIDIA provides the technology for connecting Raptor XPUs with its AI infrastructure, while Astera Labs is working with d-Matrix on high-speed data paths.
The approach places connectivity alongside processor development as d-Matrix prepares its next-generation hardware. The company is not only developing Raptor processors but also establishing the infrastructure relationships required for their planned integration.
The expected 2027 availability of integrated Raptor systems will follow the planned completion of the final processor design in 2026. Until then, d-Matrix’s announced development work remains focused on completing Raptor and its supporting system architecture.
Frequently Asked Questions
What is d-Matrix developing?
d-Matrix is developing Raptor, a next-generation processor platform designed for AI inference workloads. The processors are intended for applications including chatbots, coding assistants and voice agents.
What is NVIDIA NVLink Fusion?
NVIDIA NVLink Fusion is the technology d-Matrix plans to use to connect its Raptor XPUs to NVIDIA’s AI infrastructure platform.
What are d-Matrix Raptor processors designed to do?
Raptor processors are designed for AI inference workloads, including systems supporting chatbots, coding assistants and voice agents.
When are Raptor-based integrated systems expected to become available?
d-Matrix expects integrated systems using Raptor processors to become available in 2027. The company plans to complete the final Raptor design by the end of 2026.
What is d-Matrix’s partnership with Astera Labs focused on?
d-Matrix is working with Astera Labs on high-speed data paths for its AI systems. The connectivity work is part of the infrastructure supporting the company’s Raptor processors.
