Hors HBM, Jensen Huang a également déclaré que NVIDIA collaborait avec Samsung Electronics dans le domaine de la fabrication de wafers.

K-LinePoet · 2026-03-28T06:06:51+00:00

Le PDG de Nvidia, Jensen Huang, a remercié Samsung Electronics lors de la conférence GTC pour avoir accéléré la production du processeur d'inférence AI Groq 3 LPU, confirmant la collaboration entre les deux entreprises dans le domaine de la fabrication de wafers. Ce processeur sera intégré à la prochaine plateforme AI de Nvidia, « Vera Rubin », dont la production en série est prévue pour le troisième trimestre de cette année.

K-LinePoet

2026-03-28 06:06:51

Création du résumé en cours

According to a report by the Korean Central News Agency, NVIDIA (NVDA.US) CEO Jensen Huang specifically mentioned Samsung Electronics during his keynote speech at the annual Global Developer Conference “2026 GTC Conference” on March 16, thanking Samsung for accelerating the production of artificial intelligence (AI) inference chips for NVIDIA. This statement confirms that the two companies are collaborating in the wafer fabrication field.

The conference was held on the same day at the SAP Center in San Jose, California, USA. During his speech, Huang referred to AI inference chips and stated that Samsung is producing the “Groq 3 Language Processing Unit (LPU)” chips for NVIDIA and is accelerating the production speed, for which he expressed great gratitude to Samsung. He also mentioned that the chip will be featured in NVIDIA’s next-generation AI chip platform “Vera Rubin,” which is expected to be mass-produced and launched in the third quarter of this year.

From Huang’s remarks, it is evident that Samsung Electronics’ wafer fabrication division is producing the Groq 3 LPU chips. Thus, it can be inferred that, in addition to high bandwidth memory (HBM), Samsung Electronics is also collaborating with NVIDIA in the wafer fabrication field, highlighting the close partnership between the two giants.

On the same day, Samsung Electronics showcased its seventh-generation HBM product “HBM4E” and vertically stacked chip “Core Die” at the conference, actively promoting its collaboration progress with NVIDIA in the storage field.

Voir l'original

Cette page peut inclure du contenu de tiers fourni à des fins d'information uniquement. Gate ne garantit ni l'exactitude ni la validité de ces contenus, n’endosse pas les opinions exprimées, et ne fournit aucun conseil financier ou professionnel à travers ces informations. Voir la section Avertissement pour plus de détails.