Jensen Huang: "Samsung Foundry Manufacturing 'Groq 3'... Production Underway, Shipping in Second Half"[GTC 2026]

"Expressing Gratitude to Samsung"
Production Ramping Up, Shipments Imminent
Building an AI Partnership Following HBM4

Jensen Huang, CEO of Nvidia, officially mentioned Samsung Electronics on the 16th (local time) as a manufacturing partner for the next-generation artificial intelligence (AI) inference chip, the "Groq 3 Language Processing Unit (LPU)."

Jensen Huang, CEO of NVIDIA, is speaking at the 'GTC 2026' keynote held on the 16th (local time) at the SAP Center in San Jose, California, USA. Photo by AFP Yonhap News

During his keynote speech at "GTC 2026," held at the SAP Center in San Jose, California, USA, Huang stated, "Samsung Electronics is manufacturing the Groq 3 LPU chip for us," and added, "We are deeply grateful to Samsung." He continued, "The chip has already entered the production phase, and we are working to ramp up output as quickly as possible." He further explained, "Market shipments are expected to begin in the second half of this year, likely around the third quarter."

Groq is an inference-dedicated chip startup acquired by Nvidia last year, possessing LPU technology that delivers overwhelmingly faster inference speeds compared to traditional graphics processing units (GPUs).

Jensen Huang, CEO of Nvidia, is speaking during the keynote address at GTC 2026 held on the 16th (local time) at the SAP Center in San Jose, California, USA. Photo by Hyunji Kwon

Huang's remarks publicly acknowledged that Samsung Foundry (semiconductor contract manufacturing) is participating in the production of Nvidia's next-generation AI chips, indicating ongoing collaboration between the two companies within the AI semiconductor supply chain.

In particular, with Samsung Electronics having achieved the world's first mass production and shipment of sixth-generation high bandwidth memory (HBM4) last month, and now handling the foundry manufacturing process as well, the two companies have established a "total AI partnership" system that encompasses both memory and foundry operations.

This content was produced with the assistance of AI translation services.