AI chip startup Groq announced that it raised $750 million in new funding, giving the company a post-money valuation of $6.9 billion. This funding round exceeded earlier expectations, as reports in July suggested Groq would raise about $600 million at a valuation near $6 billion. Groq has now raised over $3 billion to date, according to PitchBook estimates. Previously, in August 2024, Groq raised $640 million at a $2.8 billion valuation, making this latest round more than double its valuation in just a year. The company sells data center computing power and develops custom AI chips, which it calls LPUs (Language Processing Units). Unlike GPUs traditionally used in AI systems, Groq's LPUs are optimized as inference engines — specialized hardware designed to run AI models quickly and efficiently. Groq’s offerings are available as both cloud services and on-premises hardware clusters, supporting popular open-source AI models from companies such as Meta, DeepSeek, Qwen, Mistral, Google, and OpenAI. The company claims its systems deliver equal or better AI performance at significantly lower costs compared to alternatives. Founder Jonathan Ross previously worked at Google, where he helped develop the Tensor Processing Unit (TPU), a specialized chip for machine learning workloads. Ross launched Groq in 2016, the same year Google announced its TPU. Today, Groq’s technology is positioned as a strong competitor to Nvidia, aiming to disrupt the dominance of GPUs in AI computing. Source: Groq Newsroom