Sacramento: TECHz – News Desk
Nvidia announced a licensing agreement with AI chip startup Groq, known for its Language Processing Unit (LPU) technology. The deal positions Nvidia to expand its dominance in AI inference, the process of running trained models, where competition from AMD, Cerebras, and Google is intensifying.
Nvidia confirmed the agreement’s scale at $20 billion, structured as a licensing and talent acquisition deal rather than a full merger. This approach allowed Nvidia to absorb Groq’s competitive edge while avoiding regulatory hurdles that often accompany large semiconductor acquisitions. Analysts described the move as a decisive step to secure the future of inference hardware.
Earlier in the month, Nvidia launched the RTX PRO 5000 Blackwell GPU (72GB), expanding desktop memory options for professional developers working with agentic AI workloads. Nvidia introduced the Nemotron 3 family of open AI models – Nano, Super, and Ultra – designed for efficiency and accuracy in generative AI tasks. These models positioned Nvidia as a stronger player in open-source AI, competing directly with Meta’s Llama and Alibaba’s Qwen.
The December announcements highlight Nvidia’s strategy to reinforce its leadership in both AI hardware and open-source models, ensuring its position as the world’s most valuable tech company heading into 2026.
