d-Matrix

Parasail and d-Matrix describe pairing Corsair with NVIDIA Hopper and Blackwell infrastructure for heterogeneous inference, with Parasail workload routing across hardware types.

d-Matrix's June 2026 release states Corsair entered full production with fabrication and supply services secured through TSMC and Alchip and volume shipments scheduled to begin that summer.

d-Matrix is bringing digital in-memory compute accelerators into AI inference deployments, with scale still below incumbent compute platforms.

d-Matrix develops digital in-memory compute architecture optimized for generative-AI inference.

d-Matrix develops digital in-memory compute architecture optimized for generative-AI inference.

d-Matrix Corsair is in full production and combines purpose-built digital in-memory compute with 3D stacked DRAM and JetStream I/O; in 2026 d-Matrix and Infinity demonstrated complete stateful Qwen3 inference across the platform rather than isolated kernels, validating full-model deployment on its memory-centric architecture.

d-Matrix announced in June 2026 that Corsair entered full production and identified TSMC and Alchip as manufacturing partners using TSMC N6, adding a concrete supply-chain execution milestone beyond its Series C funding.

d-Matrix announced closing $275 million Series C in November 2025 at a $2 billion valuation, bringing total disclosed funding to $450 million. This capitalization event is independent of generic semiconductor commercialization capability.

d-Matrix demonstrates semiconductor design, systems software and hardware commercialization capability.

The previous product documentation describes Aviator using PyTorch, Triton DSL and MLIR, alongside distributed inference and host/on-chip runtimes. This provides concrete evidence for an inference of compiler and runtime skill portability, while the accelerator remains specialized.

AI accelerator, compiler and systems skills transfer across semiconductor, cloud and inference infrastructure employers.

AI accelerator, compiler and systems skills transfer across semiconductor, cloud and inference infrastructure employers.