
AI Hardware & Accelerator Chips
Last updated · Compiled from public sources with automated enrichment
Loading the map…
Startup designing chips purpose-built for large-model training.
Visit website4 of 4 confirmed by two independent sources · 2 backed by a published source.
Published sources supporting the relationships above.
TechCrunch · 24 Feb 2026 · supports MatX competes with NVIDIA
The company aims to develop processors that are ten times more effective at training large language models compared to Nvidia's GPUs.
unknown publisher · supports MatX competes with Graphcore
unknown publisher · supports MatX competes with Graphcore
unknown publisher · supports MatX competes with NVIDIA