NVIDIA Announces Full-Scale Production of Groq 3 LPX Racks for Nebius Deployment

ChainCatcher reports that on Monday local time, Dion Harris, Senior Director at NVIDIA, announced that the Groq 3 LPX rack has entered full-scale mass production and will be deployed in the data centers of the new cloud service provider Nebius, with an expected launch later this year. This marks the commercialization of technology acquired by NVIDIA following its approximately $20 billion technology licensing deal with Groq in December last year. Several key Groq employees have joined NVIDIA, with founder Jonathan Ross serving as NVIDIA’s Chief Software Architect. NVIDIA is accelerating the production of Groq chips and delivering products to customers, underscoring the importance of low-latency inference. The Groq 3 LPX, released in March this year, is an inference accelerator for the Vera Rubin platform, integrating 500 megabytes of high-speed SRAM directly on the chip die to reduce memory bottlenecks. Each LPX rack can integrate 256 Groq 3 chips; according to NVIDIA’s benchmark tests, it can process 3,400 tokens per second. The chip is manufactured by Samsung. Harris stated that low-latency chips are not intended to replace GPUs but rather to use the appropriate processor for different stages of workloads, enabling cloud providers to offer premium pricing tiers to latency-sensitive users. NVIDIA CEO Jensen Huang plans to allocate approximately one-quarter of his data center space for AI programming applications to Groq chips, with the remainder dedicated to Vera Rubin systems, and expects combined sales of Blackwell and Vera Rubin systems to reach $1 trillion by 2027.