NVIDIA-Backed Groq Raises $350 Million To Build the World’s Leading AI Inference Cloud

NVIDIA-backed Groq has raised $350 million to become the world's leading AI inference cloud. The $350 million Series A round was led by Disruptive, with NVIDIA expected to join as an investor. Groq's valuation has now fallen to $3.5 billion, 50% less than the $6.9 billion valuation it received in September 2025. Groq's strategy is to reduce the cost of AI inference by half, as they aim to build the leading platform for this critical layer of AI infrastructure.

Groq's executive chairman and founder, Alex Davis, stated that inference will become the largest and most critical layer of AI infrastructure.

NVIDIA's involvement came after Groq licensed its core inference technology to NVIDIA in December 2025, resulting in a reported $20 billion deal. As part of the agreement, NVIDIA hired Groq's founder, Jonathan Ross, and nearly 90% of its engineering team. Groq currently operates 13 data centers, serves over six million developers, and plans to increase its capacity from 54MW to more than 200MW by 2027.

Despite the significant valuation cut, Groq remains an Nvidia Cloud Partner and operates on top of NVIDIA's stack. The company's core technology, LPU, now sits within NVIDIA's ecosystem. Groq's advantage now lies in its existing data center footprint and developer base. However, the company faces the challenge of building a defensible business after licensing away the technology that made it distinctive, leaving the question of its future as a standalone platform or a dependent supplier of NVIDIA's compute.