- Nvidia (NVDA:US) CEO Jensen Huang announced in Seoul that their newly developed Vera central processing unit (CPU) will fully utilize SK Hynix's (000660:KS) dynamic random-access memory (DRAM) chips. The two companies are preparing to significantly expand their cooperation over the next year.
- As Nvidia's first standalone data center microprocessor, the Vera chip aims to directly compete with Intel's (INTC:US) Xeon series, AMD's (AMD:US) EPYC series, and Amazon's (AMZN:US) self-developed Graviton chips, marking a reshaping of the high-performance computing hardware ecosystem.
- Jensen Huang is intensively visiting key technology companies in Korea. In addition to meeting with senior executives from SK Group and SK Telecom, he plans to meet with Samsung Electronics (005930:KS) Vice Chairman Kyung Kye-hyun, as well as leaders from Hyundai Motor and LG Group, to engage in in-depth discussions on next-generation AI network infrastructure.
Revolutionary CPU Deeply Integrated with Memory Supply Chain
Jensen Huang clearly stated in Seoul that the newly launched Vera CPU is Nvidia's core product in the standalone data center microprocessor field. This chip will deeply integrate SK Hynix's memory technology to meet the stringent requirements of future generative AI for extreme bandwidth and data processing latency. Nvidia and SK Hynix not only dominate the current graphics memory field but also plan to extend their cooperation to the next-generation general computing core. Senior executives from both companies reiterated that the scale of supply chain collaboration will see significant marginal growth from the second half of this year to next year, which could have a profound impact on the supply-demand balance in the high-bandwidth memory market.
Reshaping the Data Center Chip Competition Landscape
The official debut of the Vera microprocessor directly enters the server computing market previously dominated by traditional architectures. Market positioning indicates that Nvidia's move is not only to reduce reliance on external computing architectures but also to achieve optimal coupling of CPUs and GPUs at the system level. This strategy will exert direct market share pressure on Intel's Xeon series and AMD's EPYC chips. Meanwhile, as cloud service providers expand their self-developed chip portfolios, Nvidia is building a more tightly integrated software and hardware ecosystem to defend and counteract. If competitors fail to launch products with equivalent integration in time, their data center business pricing power may face a phase of reevaluation.
Comprehensive Multilateral Meetings with Core Suppliers
In addition to consolidating its storage supply alliance with SK Hynix, Nvidia's itinerary in Korea demonstrates its clear intention to restructure the core semiconductor chain in Asia. Jensen Huang is scheduled to hold a closed-door meeting with Samsung Electronics Vice Chairman Kyung Kye-hyun, who is responsible for the semiconductor business, on Monday. The market widely expects that the two parties may engage in substantive discussions on the progress of high-bandwidth memory validation and potential cooperation in the wafer foundry field. Furthermore, meetings with senior executives from Hyundai Motor and LG Group indicate that Nvidia is accelerating the extension of its AI computing power from traditional cloud to diverse edge scenarios such as smart mobility and smart manufacturing, with related supply chain companies' capital expenditure plans also under close scrutiny.
Positioning for AI Transformation in Telecom Networks
In discussions with senior executives from SK Telecom and other operators, Jensen Huang focused on the technological upgrade of future telecom network infrastructure. Nvidia believes that as global communication networks gradually evolve towards intelligence, traditional telecom network nodes are transforming into distributed AI computing centers. If telecom operators can successfully integrate AI technology into the underlying communication architecture, Nvidia's microprocessors and communication computing platforms will gain entirely new large-scale application scenarios. This potential cross-industry collaboration could become a long-term new variable driving revenue and profit changes in the related technology industry chain, with global analysts closely monitoring the progress of related technology validation.
