NVIDIA’s 35x AI Inferencing Leap Arrives Early as Foxconn Fast-Forwards Groq 3 LPX Racks for Trillion-Parameter Models

Foxconn is reportedly the main and exclusive supplier for key components of NVIDIA’s Groq 3 LPX AI inference racks.

Shipments are now expected in Q3 2026, earlier than planned, signaling a faster production ramp.

The rollout is tied to NVIDIA’s broader Vera Rubin AI infrastructure and larger rack targets.

The early schedule suggests stronger demand for AI inference hardware and a bigger manufacturing role for Foxconn.