Nvidia's Groq 3 LPX Racks to Launch This Year Following Major Acquisition

Aug 24, 2026

AI Summary

Nvidia has announced that its Groq 3 LPX racks are now in full production and will be deployed later this year. This follows the company's $20 billion acquisition of Groq, emphasizing the importance of low-latency inference in AI applications.

Nvidia's Groq 3 LPX rack is in full production and will be operational later this year, according to senior director Dion Harris.
The Groq racks will work with Vera central processors and Rubin graphics processors at neocloud Nebius.
The Groq architecture features 500 megabytes of SRAM on the chip to minimize memory bottlenecks, and 256 Groq 3 chips are packaged into each LPX rack.
The Groq 3 LPX rack can deliver 3,400 tokens per second, enhancing low-latency inference capabilities for AI applications.
Nvidia's acquisition of Groq for $20 billion is its largest purchase to date, highlighting the competitive landscape in AI chip manufacturing.
Other companies, like Advanced Micro Devices and Cerebras, are also focusing on low-latency inference technologies.
Nvidia CEO Jensen Huang has projected significant sales growth from the new Vera Rubin systems and Groq chips through 2027.
Nvidia is ramping up shipments of its Vera Rubin systems, which began production earlier this year.

nvidiagroqai chipslow-latencyinference

Nvidia's Groq 3 LPX Racks to Launch This Year Following Major Acquisition

Related Stories

Apple Reduces Workforce in AI Shift, Impacting Vision Pro and Siri Teams

Nvidia Customers Anticipate Price Increases for AI Systems

Nvidia Customers Face Price Increases Amid Major AI Investments

US Markets Prepare for Key Events with Focus on Bessent, Nvidia, and Trade Issues