As Nvidia's annual developer conference, GTC 2026, which drew the global tech industry's attention, wrapped up on the 19th local time, the digital twin platform Omniverse drew as much interest as its next-generation artificial intelligence (AI) chips. Core affiliates of Korea's four major conglomerates—Samsung, SK, Hyundai Motor, and LG—have already finished applying the product on manufacturing floors.
Samsung Electronics and SK hynix, which supply high-bandwidth memory (HBM) to Nvidia, shared the results achieved through Omniverse adoption and their future plans in this year's GTC announcements. Hyundai Motor Group, which released that it would jointly develop next-generation Autonomous Driving solutions with Nvidia, also formalized its plan to adopt Omniverse last year. LG Electronics, introduced as a robotics partner during Nvidia CEO Jensen Huang's GTC 2026 keynote, is currently building a Smart Factory using Omniverse.
Omniverse is a product that uses digital twin technology to precisely reproduce real spaces in virtual environments. It can create "digital twins" from products on a small scale to factories and even entire cities. Moving reality into digital makes analysis through AI and big data easier. Changing the layout of production equipment or the flow of logistics in the real world costs a lot of expense, but those constraints disappear in virtual space. It means you can freely run simulations. It is often used to boost product performance or find ways to improve factory operating efficiency. Nvidia's AI chips are used for the computation required in this process.
◇ Samsung Electronics and SK hynix: "Achieve 'autonomous factories' by 2030 with Nvidia technology"
Samsung Electronics and SK hynix gave back-to-back presentations on the second day of GTC 2026. Song Yong-ho, vice president and head of the AI Center in Samsung Electronics' DS (semiconductor) institutional sector, presented on "Semiconductor manufacturing innovation using Agentic AI," while Do Seung-yong, vice president and head of the digital transformation (DT) institutional sector at SK hynix, joined a panel on "Designing the future of manufacturing." Although they differed somewhat on specifics, both cited Omniverse as a key theme. Both companies said they plan to make semiconductor manufacturing sites autonomous by 2030 through Omniverse. The aim is not simple automation of processes but to maximize semiconductor manufacturing efficiency by establishing factories where AI understands the site and executes optimal decisions on its own.
Song shared a case of implementing "Pyeongtaek Plant 1," the world's largest memory semiconductor production base, in virtual space with Omniverse, saying it "has enabled preemptive responses, such as process optimization and risk prediction." He also released for the first time a video showing what outcomes could be achieved by introducing Humanoid Robot units to semiconductor production plants.
Do noted that surging demand for memory semiconductors due to the expansion of AI services requires "manufacturing innovation" to expand production capacity. To that end, SK hynix said it has adopted Omniverse to validate production flow, material movement, layout, and more in advance by performing simulation, AI training, and operations optimization without production stoppages.
A semiconductor industry official said, "Samsung Electronics and SK hynix emphasized at GTC 2026 that they are both memory suppliers and customers that use Nvidia's digital twins," adding, "They effectively delivered the message that they are 'partners' with Nvidia in autonomizing semiconductor manufacturing processes."
◇ Expansion of "Omniverse adoption" in step with Jensen Huang's visit to Korea
What Samsung Electronics and SK hynix revealed at GTC is closer to an "interim results announcement" than a new collaboration. Samsung Electronics first disclosed its adoption of Nvidia's Omniverse in Mar. 2024. The Omniverse adoption plan revealed at GTC 2024 was further detailed when CEO Huang visited Korea in October last year for the Asia-Pacific Economic Cooperation (APEC) summit. At the time, Samsung Electronics said it would build a digital twin manufacturing system spanning "memory, logic, foundry (semiconductor contract manufacturing), and packaging" using 50,000 Nvidia graphics processing units (GPUs). The "Pyeongtaek Plant 1" case announced at this event indicates that this strategy is being realized sequentially.
SK hynix likewise announced, timed to Huang's visit to Korea, a plan to convert its Icheon campus into a digital twin environment using 2,000 units of Nvidia's Blackwell GPUs. SK hynix plans to sequentially expand this to the Yongin Semiconductor Cluster as well. The first-phase fab of the Yongin Semiconductor Cluster is being built to the height of a 50-story apartment in the Wonsam-myeon area of Cheoin-gu, Yongin, and SK hynix plans to invest about 31 trillion won there by the end of Dec. 2030. SK hynix's ambition is to prepare an environment close to an "autonomized process" from initial operation.
Hyundai Motor Group did not separately announce results related to Omniverse at GTC 2026. However, it released a strategic partnership with Nvidia in Jan. 2025, thereby formalizing its adoption of Omniverse. Like Samsung Electronics, Hyundai Motor Group further detailed the plan last October by announcing, among other things, the introduction of "50,000 Blackwell GPUs." At the time, Hyundai Motor Group said it "will enable testing and verification of Autonomous Driving vehicle software across infinite driving scenarios."
LG Electronics, which is developing Humanoid Robot units using Nvidia's inference AI models, is also actively using Omniverse. In October last year, LG Electronics said it was combining Omniverse with manufacturing and production data accumulated over 60 years. It is building a virtual space that encompasses factory equipment units to streamline logistics operations and maintenance.
Nvidia is shifting its business position from a simple AI chip supplier to a "solutions provider." On last month's earnings conference call, CEO Huang said, "We want to put as many ecosystems as possible on top of Nvidia." The idea is to deliver high-performance solutions powered by the most efficient AI chips to customers and produce results.
Omniverse is considered a core product for realizing this strategy. In particular, its importance has grown in the era of "physical AI," in which AI that once operated in digital spaces like chatbots advances to being embedded in devices and products, recognizing the real world and making its own judgments. Nvidia has continued to advance Omniverse, which it released in open beta in Oct. 2020. In Jan. last year, it expanded service areas to robots, Autonomous Driving, and vision AI, and more recently it has been improving the platform to "do more in digital space" by partnering with companies that have digital twin technology, such as Siemens and Dassault Systèmes.