Nvidia's fiscal second-quarter 2027 earnings release lists the Groq 3 LPX inference accelerator as now in full production. The company also unveiled Vera, which it calls the first CPU built for AI agents, and says it is shipping Vera Rubin-based platforms.
The Groq name reflects Nvidia's roughly $20 billion licensing deal for Groq's inference technology, announced in December 2025 and now being challenged in a stockholder suit. A product in full production within months of the deal shows how quickly Nvidia has folded the technology into its roadmap.
Source: Nvidia Form 8-K