Korean AI Expert Day [Spoken in Korean, 한국어 진행 ...
Google AI summary of how horean intelligence has supported nvidia and world deep learning ai
In recent years, the relationship between Jensen Huang (CEO of Nvidia) and South Korea has strengthened, leading to significant partnerships and investment activity that have generated numerous Korean-language documents.
Key collaborations and investments
- SK Hynix: The South Korean memory chip giant is a major supplier of High-Bandwidth Memory (HBM) chips for Nvidia's AI processors. SK Hynix has committed to building new fabrication plants in South Korea to meet the rising demand driven by Nvidia and other AI customers.
- Samsung: After an extended qualification process, Samsung secured approval from Nvidia in September 2025 to supply its 12-layer HBM3E chips. This is a significant move for Samsung and strengthens the Korean role in Nvidia's supply chain.
- KT: The Korean telecom company KT collaborated with Nvidia to develop its AI-powered smart speaker, GiGA Genie. KT used Nvidia's DGX SuperPOD and NeMo framework to train its large language models (LLMs) to master the complex Korean language.
- South Korean government: In May 2025, the South Korean government announced a $1 billion plan to purchase 10,000 high-performance Nvidia GPUs for a new national AI computing center. This initiative seeks to boost the nation's AI capabilities and promote research and development.
- KT invests in sovereign AI: While not directly involving Nvidia, major Korean tech firms like KT have received government funding to build "sovereign AI" models to reduce reliance on foreign technologies. This creates a complex dynamic where Korean companies are both key partners and emerging competitors to Nvidia.
Jensen Huang's presence in Korea
CEO visits: In August 2025, Korean media reported that Jensen Huang was expected to visit Korea in October for the APEC summit in Gyeongju. His visits often generate significant media attention in the country.Examples of Korean-language PDFs
Due to the nature of search results, finding publicly available Korean-language PDF documents on Nvidia's website is challenging. However, search results reveal that Korean-language documents are created for specific purposes, such as:
Technical papers: ResearchGate contains conference papers in Korean related to Nvidia's CUDA platform, such as one on an efficient implementation of the Korean Digital Signature Algorithm.summary 2 search ai korea partners jensen or "data sovereignty"
Recent events show a connection between Nvidia CEO Jensen Huang, South Korea's AI industry, and the concept of data sovereignty. South Korea is pursuing "sovereign AI," a strategy to develop national AI capabilities using domestic data and infrastructure, and has entered into strategic partnerships with Nvidia to achieve this.
Jensen Huang and South Korea AI partners
- APEC Summit in Korea: Nvidia CEO Jensen Huang is scheduled to attend the Asia-Pacific Economic Cooperation (APEC) summit in Gyeongju, South Korea, in October 2025. His participation highlights deepening US-Korea cooperation on semiconductors and AI.
- Meetings with conglomerates: In August 2025, Huang met with leaders of major South Korean firms at a US-Korea business roundtable in Washington D.C., including Samsung Electronics Chairman Lee Jae-yong and SK Group Chairman Chey Tae-won.
- Semiconductor suppliers: Nvidia's dominance in AI hardware relies heavily on high-bandwidth memory (HBM) chips supplied by South Korean firms SK Hynix and Samsung Electronics.
- Partnership with Naver: Earlier in 2025, Naver executives met with Huang to strengthen their partnership on building sovereign AI systems and data center operations.
- Gaming and robotics collaboration: In April 2025, Huang met with the CEO of Korean game studio Krafton to discuss future collaboration in AI and robotics.
- Rivalry from FuriosaAI: South Korean chipmaker FuriosaAI is positioning itself as a challenger to Nvidia by partnering with LG and developing more cost-effective and efficient AI hardware.
South Korean data sovereignty
- National initiative: South Korea has launched a "Sovereign AI" initiative to develop homegrown AI models, aiming to become one of the top three global AI powers by 2027.
- Reducing foreign reliance: The effort is a move to reduce reliance on foreign-controlled AI models and ensure national data control and security.
- Cultural and linguistic relevance: A key motivation for the initiative is to train AI on Korean data to ensure models understand Korean culture and language. This concern gained prominence after a global AI model gave an ambiguous answer about the disputed Dokdo islands.
- Government funding and partnerships: The Korean government has invested hundreds of millions of dollars in the project and selected five domestic teams, including Naver Cloud, SK Telecom, LG AI Research, NC AI, and Upstage, to compete in building national foundation models.
- New AI law: To regulate this growing field, South Korea's AI Basic Law is set to take effect in January 2026, applying regulations not only domestically but also to foreign companies serving Korean users.
Feedback
2 pages
Missing: korean | Show results with: korean
---------------------------------
alter4native
Major partnerships between Nvidia and Korean companies like SK Telecom and Naver are focused on building sovereign AI infrastructure to ensure data sovereignty. While the "GTU" acronym was not found in relation to these partnerships, it is clear that Korean companies are building large-scale, domestically controlled AI computing facilities using Nvidia's latest GPU technology.
Key Korean partners and initiatives with Nvidia
SK Telecom (SKT)
- Haein Cluster: SKT launched Korea's largest sovereign AI infrastructure in August 2025, using over 1,000 Nvidia Blackwell GPUs. The cluster provides GPU-as-a-Service (GPUaaS), with the goal of supporting the development of national AI foundation models and expanding Korea's overall AI capabilities.
- VAST Data partnership: SKT collaborated with VAST Data, which provides an AI operating system to help run the AI cloud platform. This partnership aims to balance speed, compliance, and cost for the national AI infrastructure.
- Merger with Sapeon: In August 2024, SKT's AI chip subsidiary, Sapeon Korea, announced a merger with another Korean AI chipmaker, Rebellions. The move is a national effort to challenge global AI chip leaders like Nvidia.
Naver Cloud
- Morocco AI Data Center: Naver, in partnership with Nvidia and others, is building a large-scale AI data center in Morocco to serve the Europe, Middle East, and Africa (EMEA) regions. The facility will use Nvidia's Blackwell GPUs and is designed to create a sovereign AI and cloud infrastructure that aligns with Europe's focus on data sovereignty. Government-backed GPU project: In July 2025, Naver Cloud was selected as one of three cloud service providers for a South Korean government initiative to procure and distribute Nvidia GPUs, including 3,056 H200 units.
Other partnerships and context
Government investment: The South Korean government has a major initiative to build a sovereign AI ecosystem. In 2025, the Ministry of Science and ICT budgeted $1.03 billion to procure 10,000 high-performance Nvidia GPUs for public and private use.=======alt
At the GTC 2025 conference, Nvidia announced partnerships with several Korean companies related to sovereign AI, including Naver Cloud, SK Telecom, Medical IP, and memory chip makers like SK Hynix and Samsung. The collaborations focus on building localized AI ecosystems, developing new AI applications, and supplying high-performance GPUs to address Korea's data sovereignty needs.
Naver Cloud
Naver Cloud announced a groundbreaking partnership with Nvidia to build a sovereign AI ecosystem specifically for the East Asian market.
- Localized AI system: The collaboration focuses on creating AI systems developed and controlled within a country to ensure data security, regulatory compliance, and alignment with cultural nuances.
- HyperCLOVA X integration: Naver Cloud CEO Kim Yoo-won highlighted that the effort will leverage its advanced AI model, HyperCLOVA X, which specializes in non-English languages.
- Application in elderly care: As an example of sovereign AI addressing national challenges, Naver's CLOVA CareCall service for the elderly was adapted for the Japanese market. The service uses Nvidia's ACE platform and Naver's AI technologies to provide emotional care.
- Financial sector deployment: The Bank of Korea also partnered with Naver to develop a high-performance generative AI platform on dedicated cloud infrastructure, training Naver's HyperCLOVA X model on proprietary financial data.
SK Telecom (SKT)
VAST Data collaborated with SK Telecom to create Korea's largest AI infrastructure, which is based on Nvidia supercomputers.
- Sovereign-grade control: The infrastructure is designed to enable AI services to operate entirely within South Korea's borders.
- Secure, multi-tenant services: It provides secure, multi-tenant AI services, ensuring data privacy and guaranteed performance for isolated workloads.
- GPU as a Service (GPUaaS): The infrastructure was designed for national-scale GPUaaS to meet the growing demand from both public and private sector AI developers.
Medical IP
Medical IP, a Korean medical imaging company, was highlighted at GTC 2025 as Nvidia's sole biohealthcare partner in the country.
- DeepCatch X: At the conference, the company debuted DeepCatch X, an AI tool that creates 3D anatomical models from 2D chest X-rays. The technology is designed to infer over 100 anatomical structures with high accuracy.
- Apple Vision Pro integration: The models created by DeepCatch X can be viewed through Apple's Vision Pro headset.
Memory chip manufacturers
In addition to the GTC partnerships, Korea's memory chip giants play a crucial role in supporting the AI hardware ecosystem.
- SK Hynix and Samsung Electronics: Both companies were involved in discussions related to supplying Nvidia-optimized chips for AI development.
- Investment in AI chips: In July 2024, SK Hynix announced a $6.8 billion investment to build a new plant to meet the surging demand for advanced AI chips.
- National initiative: The Korean government, through its Ministry of Science and ICT, announced a $1 billion GPU deal in May 2025 to acquire high-performance GPUs, with Nvidia likely a major supplier.
No comments:
Post a Comment