30-year semiconductor veteran; started career at Intel, previously built a $3B business at Inphi (sold to Marvell). Founded D-Matrix in 2019 to build inference-specific accelerators; recently raised $275M at a $2B valuation targeting enterprise AI workloads.
no scored calls yet — needs a stated position or a categorical verdict, with a matured window vs SPY
Dmatrix's inference-optimized chip delivers 10x speed, 3x lower cost and 3-5x better energy efficiency than GPUs by avoiding high-end process nodes and HBM memory, targeting enterprise sub-100B parameter models where Nvidia is less optimized.
D-Matrix targets inference workloads for enterprise smaller models (<100B params) with a dedicated accelerator architecture delivering 10x speed, 3x cost and energy efficiency versus GPUs, while avoiding supply-constrained HBM and advanced packaging by using older process nodes.