D-Matrix Targets Nvidia Rack Integration by 2027
AI inference chip startup D-Matrix plans to integrate its Raptor XPU into Nvidia’s MGX rack architecture by Q4 2027. D-Matrix expects the first Raptor tape-out before the end of 2026 and is designing the chip to work with Nvidia’s NVLink Fusion interconnect. The company says up to 144 Raptor XPUs could operate within a single NVLink fabric.
Raptor combines a TSMC 4nm logic die with 3D-stacked DRAM. It targets more than 100 TB/s of memory bandwidth and 32 GB of memory per card. D-Matrix claims Raptor can deliver about 4.7 times the throughput of HBM-based alternatives for generative AI inference, although the figure remains a company estimate.
D-Matrix’s existing Corsair platform is already in production and uses SRAM-based digital in-memory computing. The startup has raised an estimated $450 million to $500 million, including a $275 million Series C round that valued it at about $2 billion. Alchip, Andes and Astera Labs are among the technology partners linked to the company and its Raptor integration plans.
Neutral
The direct cryptocurrency-market impact is likely neutral. The announcement concerns AI inference hardware rather than a cryptocurrency, blockchain network or token, and it provides no immediate change to crypto revenues, regulation, liquidity or network activity. As a result, traders are unlikely to reprice BTC, ETH or major altcoins solely on this news.
There could be a modest indirect effect through Nvidia and the wider AI-equipment trade. Nvidia-linked hardware announcements have historically supported AI-related equities and, at times, AI-themed crypto tokens when they strengthen expectations for data-centre demand. However, this project is still at the roadmap stage: first tape-out is targeted for 2026 and rack-scale deployment for Q4 2027. The reported 4.7-times throughput advantage is also an unverified company claim, leaving execution, manufacturing and customer-adoption risks.
In the short term, the story may create sentiment-driven volatility in AI and semiconductor markets, but it is unlikely to produce a sustained crypto move. Over the longer term, successful Raptor deployment could reinforce demand for AI infrastructure and indirectly benefit AI-related digital assets. Traders should monitor Nvidia’s share price, semiconductor-sector flows, AI token volume, product-validation milestones and broader risk appetite rather than treating the announcement as a direct crypto catalyst.