GQGQ
Groq
The fastest inference on the planet
AISeries DEst. 2016200–500 employeesMountain View
Groq builds the Language Processing Unit (LPU), a custom chip delivering inference speeds 10x faster than GPUs. Their cloud API serves open-source models like Llama and Mixtral at speeds that feel instantaneous — enabling real-time AI applications previously impossible on GPU infrastructure. Groq is the infrastructure layer for latency-sensitive AI products.
Total Raised$640M
BlackRockCisco InvestmentsSamsung Catalyst
Founders
J
Jonathan RossD
Douglas Wightman
