NVIDIA Begins Full Production of Groq 3 LPX as AI Chip Race Intensifies
NVIDIA has begun full production of its Groq 3 LPX inference system, a major development in the competition to provide the computing power needed for increasingly demanding AI applications.
The system comes from technology obtained through NVIDIA's $20 billion acquisition of Groq, and the new LPX racks are designed specifically for AI inference—the stage where trained AI models actually generate responses and perform tasks for users.
NVIDIA says Nebius will be the first cloud customer to deploy Groq 3 LPX systems alongside its Vera Rubin platform. The racks can contain up to 256 LPX accelerators.
Why this matters
|
|
|
The AI race is increasingly shifting from simply training bigger models to making AI faster and cheaper to run.
That means specialized inference chips could become extremely important as AI agents, chatbots, coding assistants and other applications handle billions of requests.
