Nvidia Chief Executive Jensen Huang toured the exhibition hall at the annual developer conference "GTC 2026," visiting the booths of Samsung Electronics and SK hynix in succession to underscore their partnerships. As high bandwidth memory (HBM) emerges as a key variable in the race for artificial intelligence (AI) performance, the moves directly highlighted the strategic standing the two companies hold in Nvidia's ecosystem.

Samsung Electronics says on the 17th that it strengthens its global AI leadership by showcasing its ability to be the only supplier of total memory solutions that implement next-generation HBM4E technology and the Vera Rubin platform at Nvidia GTC in San Jose, United States, from the 16th to the 19th (local time). The photo shows Nvidia CEO Jensen Huang visiting the Samsung Electronics booth at GTC 2026 and posing for a commemorative photo. From left, Hwang Sang-joon, executive vice president of Memory Development, Nvidia CEO Jensen Huang, and Han Jin-man, president and head of the Foundry Business. /Courtesy of Samsung Electronics

According to the industry on the 17th, Huang first stopped by the Samsung Electronics booth at around 5 p.m. on the 16th (local time) at the GTC exhibition held at the San Jose Convention Center in California. Meeting with Executive Vice President Cho Sang-yeon, head of DSA, he called it a "great partnership" and said, "Samsung is the best in the world."

When Cho showed Huang the core die of the recently mass-produced "HBM4," introducing it as "the world's best HBM4 in existence," Huang said, "I must approve it. Do you need approval?" and signed the product.

Huang then signed a wafer of Groq AI inference chips produced at Samsung's foundry. It was a symbolic moment showing that cooperation is expanding beyond memory into manufacturing. As he left, he also shouted, "Go, Samsung."

Nvidia CEO Jensen Huang (left) and SK Group Chairman Chey Tae-won /Courtesy of SK hynix

About 30 minutes later at the SK hynix booth, the atmosphere grew more familiar. Huang chatted with SK Group Chairman Chey Tae-won and said, "You are perfect," "I'm proud." He then wrote the phrase "Jensen♡SK hynix" on a prototype AI system equipped with HBM4 and server low-power memory modules, stressing their close partnership.

On his first visit to GTC, Chairman Chey, who met with Huang over "chimaek" in Silicon Valley in Feb., met him again about a month later to continue the partnership. As he moved through the exhibition hall, Chey also greeted Madison Huang, Nvidia's head of product marketing and Huang's daughter.

Huang visited the major memory corporations' pavilions in succession that day, personally checking on supply chain partners. With competition in AI Semiconductor intensifying and securing HBM to resolve data-processing bottlenecks emerging as a key factor, his moves are seen as an effort to publicly highlight ties with major partners such as Samsung Electronics and SK hynix.

The same message continued in the keynote. Huang specifically mentioned the Samsung Electronics foundry, saying, "Samsung is producing LPU chips for us," and expressed thanks. It signaled that Samsung Electronics is expanding its role to cover not only the memory needed for next-generation AI platforms but also manufacturing.

The axis of cooperation expanded beyond semiconductors to mobility. Huang mentioned cooperation with Hyundai Motor and outlined plans to develop Level 4 Autonomous Driving technology. The idea is to combine Nvidia's AI platform with vehicle technology and expand into robotaxis and other areas.

The GTC venue this year showed that the AI supply chain centered on memory and foundry is being restructured around Nvidia, according to industry assessments. Analysts say Samsung Electronics has solidified its role as a "total supplier" covering both memory and manufacturing, while SK hynix has cemented its position as a key partner focused on high-performance memory. An industry official said, "Jensen Huang's very path probably revealed the priorities of the supply chain in the AI era."

※ This article has been translated by AI. Share your feedback here.