AIChipMapExport Controls

Groq · SUPPLY CHAIN TRACE

CHIP DESIGNERS🇺🇸 US

Groq designs the LPU (Language Processing Unit), a custom inference accelerator that achieves the fastest publicly-benchmarked LLM inference speeds through deterministic dataflow execution — eliminating the memory bandwidth bottleneck that constrains GPU inference. Groq's chips are fabricated by TSMC and its GroqCloud service has become a reference benchmark for inference speed in the AI industry.

Open full company page

UPSTREAM — SUPPLIERS

↓ all supply to Groq ↓

Groq

AI inference chip company designing LPUs (Language Processing Units); GroqCloud holds the fastest publicly benchmarked LLM inference throughput

UPSTREAM — SUPPLIERS ↑ · DOWNSTREAM — CUSTOMERS ↓
→ View company page