NewKerala.com

Samsung Unveils HBM4E Chip at Nvidia Event, Deepening AI Partnership

Samsung Electronics unveiled its next-generation HBM4E memory chip during Nvidia's GTC 2026 conference, showcasing a significant performance leap over the previous HBM4. Nvidia CEO Jensen Huang publicly thanked Samsung for manufacturing the Groq 3 LPU chip, confirming an expansion of their collaboration into the foundry business. The new HBM4E is designed to support Nvidia's Vera Rubin AI platform with speeds of 16 Gbps per pin. Samsung also highlighted its advanced hybrid copper bonding technology to improve packaging for future AI memory solutions.

Key Points: Samsung Unveils HBM4E, Nvidia CEO Highlights AI Partnership

  • Samsung unveils physical HBM4E chip
  • HBM4E offers 4.0 TB/s bandwidth
  • Nvidia CEO confirms Samsung makes Groq 3 LPU
  • Partnership expands to foundry business
  • Hybrid copper bonding tech showcased
3 min read

Nvidia CEO highlights closer ties as Samsung unveils HBM4E chip

Samsung reveals HBM4E chip at Nvidia GTC, boosting AI performance. CEO Jensen Huang confirms expanded foundry cooperation for AI platforms.

"I want to thank Samsung, who manufactures the Groq 3 LPU chip for us, and they are cranking as hard as they can. - Jensen Huang"

Seoul, March 17

Samsung Electronics unveiled its seventh-generation high bandwidth memory, namely HBM4E, during an annual technology conference hosted by Nvidia, where the U.S. tech giant highlighted its expanding partnership with the Korean chipmaker beyond memory chips.

Samsung Electronics provided updates on the development of its HBM4E products as it showcased its capabilities as a total memory solution provider for Nvidia's Vera Rubin AI platform during Nvidia GTC 2026, which kicked off on Monday (U.S. time) for a four-day run in California, reports Yonhap news agency.

It marked the first time Samsung Electronics unveiled the physical HBM4E chip, which is expected to support speeds of 16 gigabits per second per pin and bandwidth of 4.0 terabytes per second.

The performance is an improvement over the HBM4, which has speeds of 13 gigabits per second per pin and bandwidth of 3.3 terabytes per second.

During his keynote speech, Nvidia CEO Jensen Huang expressed gratitude to Samsung Electronics for its production of the Groq 3 language processing unit (LPU), which will be used for Nvidia's AI platform to enhance its performance.

"I want to thank Samsung, who manufactures the Groq 3 LPU chip for us, and they are cranking as hard as they can. I really appreciate you guys," the CEO said, confirming that Samsung Electronics' foundry division manufactures the chip.

Huang's remark suggests Samsung Electronics and Nvidia have broadened their cooperation in the AI sector to include the foundry, or chip contract manufacturing, business.

Last month, Samsung Electronics began its first commercial shipments of sixth-generation HBM, or HBM4, designed for Nvidia's Vera Rubin platform, which the chipmaker says offers the "ultimate performance" for AI computing.

Samsung Electronics also introduced hybrid copper bonding (HCB) technology, which enables stacking of more than 16 layers while reducing thermal resistance by 20 percent compared to thermal compression bonding (TCB), highlighting its packaging capabilities for the next-generation HBM.

"In order for innovation in the AI industry, a strong AI system, such as the Vera Rubin platform, is essential," the South Korean tech giant said.

"Samsung Electronics plans to continue supplying high-performance memory solutions supporting the Vera Rubin platform," it added.

The chipmaker added the two companies intend to lead a transition in the global AI infrastructure paradigm on the back of such ties.

During the event, Samsung Electronics set up an exhibition booth comprising three zones -- AI Factories, Local AI and Physical AI -- introducing the firm's next-generation chips meeting demand from the AI industry.

- IANS

Share this article:

Reader Comments

P
Priya S
The specs are mind-blowing! 4.0 TB/s bandwidth is insane. This kind of tech will power the next generation of AI models. Makes me wonder when we'll see Indian semiconductor fabs producing something even close to this level. The government's push is there, but execution is key. 🤔
R
Rohit P
Great news for tech, but a small note of caution. This deep partnership between a US giant and a Korean leader further consolidates power in very few hands. We need more players to ensure healthy competition and prevent monopolistic pricing, especially for developing economies like ours.
S
Sarah B
The hybrid copper bonding tech reducing thermal resistance is a game-changer for packaging. Heat dissipation is a massive challenge in our Indian data centers, especially in the summer. If this tech trickles down, it could really improve efficiency and reduce cooling costs here. 👍
V
Vikram M
Jensen Huang thanking Samsung publicly is a big deal. It shows they're moving beyond just a buyer-supplier relationship to a true co-development partnership. This is the model Indian electronics manufacturing needs to aspire to - moving from assembly to core innovation and IP creation.
K
Kavya N
Exciting, but let's not forget the environmental impact. Manufacturing these advanced chips is incredibly resource and energy-intensive. I hope part of their "transition in the global AI infrastructure paradigm" includes a strong focus on sustainability and green manufacturing processes. 🌍

We welcome thoughtful discussions from our readers. Please keep comments respectful and on-topic.

Leave a Comment

Minimum 50 characters 0/50