Nvidia and Groq Enter Landmark Technology Deal

We’re excited to share a major milestone involving Groq, a portfolio company of our sister fund. On December 24, 2025, Groq announced a non-exclusive inference technology licensing agreement with Nvidia. As part of the agreement, founder Jonathan Ross, President Sunny Madra, and other team members will join Nvidia to help advance and scale the licensed technology, while Groq will continue operating as an independent company under new CEO Simon Edwards.

Strategically, this agreement underscores the accelerating demand for inference and lays the groundwork for a hybrid AI infrastructure architecture, where different compute engines are optimized for different stages of serving models. Groq’s deterministic, low-latency inference model augments the scale, memory bandwidth, and throughput of Nvidia GPUs, improving speed and cost efficiency for real-time, long-context, multi-step reasoning systems in areas like robotics, autonomy, and edge deployment.

The reported near tripling of Groq’s valuation from $6.9B to $20B in three months validates our approach in identifying the opportunities with lasting value in noisy, bubble-like markets. Groq exemplifies that winner archetype: (i) differentiated and durable moat as a “nuts-and-bolts” player in AI, (ii) aggressive infrastructure buildout (12 data centers across multiple regions in 2025 with expansion plans underway), (iii) and a team of unmatched engineering talent at the frontier of AI hardware (ex-Google TPU).

We’re proud to have supported Groq’s journey to this milestone and be a part of the paradigm shift towards an inference-first era for AI.