Groq and Nvidia: deciphering an unprecedented coopetition in the AI Cloud market
Article excerpt
The American startup Groq officially announced on August 17, 2026, a Series A fundraising of 350 million dollars, led by the Disruptive fund with anticipated participation from Nvidia. This operation values Groq at 3.5 billion dollars, compared to 6.9 billion a year earlier, and brings the total raised by the company to one billion dollars in two months, after the 650 million obtained in June 2026. It follows a 20 billion dollar technology licensing agreement signed in December 2025, through which Nvidia acquired the intellectual property of the LPU (Language Processing Unit) chips designed by Groq, as well as several of its key talents, including founder Jonathan Ross. Groq currently operates 13 data centers in North America, Europe, the Middle East, and Asia-Pacific, serving six million developers, and plans to increase its computing capacity from 54 to over 200 megawatts by 2027 thanks to these new funds. Summary and translation produced by Le Fil IA from Le Big Data. Read the original article This operation marks a strategic shift: Groq is giving up directly confronting Nvidia in the field of proprietary chips to reposition itself as a cloud infrastructure provider specializing in inference, by deploying mixed clusters combining its own LPUs with its former rival's GPUs. For IT departments, this case provides a useful framework for understanding an AI computing market...
Keep reading with a free account
The rest of this article, and every signal for Groq, is in your free account.
Extracted from this sentence
Rather than financially exhausting itself against this near-monopoly, the company chose to become a certified cloud partner of Nvidia, by shifting towards the neocloud model, these infrastructure providers specializing in intensive computing for AI.
