MA
MatX
High-throughput chips for LLMs
About MatX
MatX designs high-throughput chips for large language models, targeting frontier labs with products like the MatX One chip.
MatX builds high-throughput chips optimized for the largest AI models, delivering superior FLOPS and latency for training, inference, and reinforcement learning. Their MatX One chip achieves higher throughput than any announced product while matching best-in-class latencies. The company focuses on frontier AI labs, supporting large MoE and dense models with efficient memory and interconnect architectures. Revenue is generated through direct sales of custom hardware solutions to AI research organizations.
Open roles at MatX
Work at MatX? Claim this page to keep it up to date.