Maverick Silicon

Tenstorrent builds AI processors, computing systems and licensable chip designs for developers, enterprises and infrastructure operators.

Develops modular three-dimensional compute and memory systems for AI infrastructure.

CScale develops optical interconnects that keep large AI computing systems communicating reliably.

Celero develops coherent digital signal processors and analog technology for optical links that connect AI accelerators and data centers. Its validated 2nm silicon architecture is designed to move more data across scale-up, scale-out and data-center links with lower power and a path from 1.6T to next-generation 3.2T connectivity.

Fab2 develops chipmaking equipment and design software to shorten the path from chip design to fabrication.

Lyte develops perception systems for robots and autonomous machines used in inspection, logistics and manufacturing.

iPRONICS builds silicon-photonics optical circuit switches that let AI data centers reconfigure high-bandwidth connections between compute systems across and within racks.

Develops analog integrated circuits using AI-based digital twin design automation

Develops ultra-low-power silicon for AI data centers and Physical AI systems

Develops RF-based interconnect technology for AI data-center and automotive networking

Builds MEMS-based precision timing chips for electronics and AI data centers

Provides open-standard AI networking infrastructure for ultra-low-latency training and inference

Provides AMD-powered AI cloud infrastructure for high-performance workloads

Provides open-standard AI networking infrastructure for ultra-low-latency training and inference

Develops high-speed energy-efficient chip-to-chip connectivity solutions for AI systems

Provides open-standard AI networking infrastructure for ultra-low-latency training and inference

Develops Ethernet-based packet processors for software-defined autonomy.

Provides semiconductor services including chip design, manufacturing, and supply chain visibility