Groq

Groq is an AI infrastructure company founded in 2016 by Jonathan Ross, former lead architect of Google TPU. The company builds the worlds fastest AI inference chips called Language Processing Units or LPUs. Groq chips can run large language models at speeds 10x faster than traditional GPUs with significantly lower latency and energy consumption. The company has raised over $640 million in funding at a $2.8 billion valuation from investors including BlackRock, Neuberger Berman, and Tiger Global. Groq is headquartered in Mountain View with approximately 300 employees. The company offers GroqCloud API providing access to models like Llama 3, Mixtral, and Gemma at speeds exceeding 500 tokens per second. Major customers include enterprises requiring real-time AI inference for applications like customer service, content moderation, and voice assistants. Groq competes with NVIDIA in the AI chip market and is considered a potential disruptor in AI infrastructure.