CITIC Securities: How the changes in NVIDIA's executive team will affect Chinese technology assets?
*Author: [CITIC Securities](https://mp.weixin.qq.com/s/IJWVZCONYQVWb-I_P5jCQQ)*
On November 12, U.S. time, NVIDIA's board announced that Jensen Huang resigned from his positions as President, CEO, and Director of the company, effective immediately. CFO Colette Kress will serve as interim CEO; since November 5, NVIDIA's stock price has fallen by more than 22%. We believe that after the management adjustment of leading overseas companies, the likelihood of changes in their customers, supply chains, and overall strategies has increased; meanwhile, according to The Information, the U.S. Department of Commerce is finalizing new AI chip export control rules, planning to include countries in Southeast Asia and others under stricter export licensing management, making it increasingly difficult for the NV series cards to enter the domestic market. Concerns about Chinese CSPs obtaining computing power may lead to short-term performance expectation disturbances. **In the short term, stricter export controls are expected to affect the progress of overseas computing center construction, reducing visibility on CSP growth; in the medium to long term, domestic CSPs are expected to further increase their procurement ratio of domestic computing power, and with the iteration of domestic products and enhanced supply capabilities, domestic computing power is likely to become a mainstay of domestic AI-related Capex, occupying a significant mid-term share.**
### Matters:
Jensen Huang has resigned from all positions at NVIDIA, and the CFO will serve as interim CEO. On November 12, U.S. time, NVIDIA's board announced: Jensen Huang has resigned from his positions as President, CEO, and Director of the company, effective immediately; Executive Vice President and CFO Colette Kress will serve as interim CEO while continuing as CFO. The board has initiated the selection process for a new CEO and appointed Harvey C. Jones as the company's first Chairman of the Board. Jensen Huang, as a co-founder of NVIDIA, has led the company for 33 years, and his deep binding with the company's strategy and customer relationships means that the impact of this management change will far exceed that of a typical personnel change.
### Domestic Semiconductors and Computing Power Chain: Replacement Logic Fully Strengthened, Super Nodes Drive Domestic Chain into Performance Release Period
After the management adjustment of leading overseas companies, along with the iteration of domestic products and improvement in supply capabilities, domestic computing power is expected to become a mainstay of domestic AI-related Capex, occupying a significant mid-term share. On the demand side, the difficulty of NV series cards entering the domestic market continues to increase; we estimate that by 2026, China's existing computing power load will be 30.4GW, and it is expected to reach 97.4GW by 2030. The state-owned asset system, credit costs, and infrastructure carrying conditions are favorable, and China's computing power construction is still in a clear acceleration phase. Under the policy requirements for the localization rate, breakthroughs in performance and capacity improvements of domestic AI acceleration cards are crucial. In this context, domestic computing power chip manufacturers are expected to receive more supply support. Additionally, the ramp-up of domestic advanced process production lines and improvements in yield are also solid support for the volume increase of domestic AI computing cards. In terms of industry rhythm, it is expected that from the end of this year to next year, domestic super nodes represented by Alibaba and ByteDance will accelerate their landing, and both supply and demand for domestic cards are expected to increase significantly. The connection links of super nodes are expected to see a significant explosion in value, and the domestic chain is expected to gradually release performance starting from Q4 2026.
### Overseas Computing Power Chain: Both Investment and Financing Already Under Pressure, Management Change Adds Disturbance, Maintain Caution on Overseas Chain in the Short Term
We estimate that by 2026, the global (excluding China) existing computing power load will be 69.3GW, and at the current pace, it is expected to reach 322.4GW by 2030. However, under rapid growth, North America's computing power construction faces challenges on both the investment and financing sides: On the investment side, at the macro level, AI capital expenditure will reach a peak as a proportion of GDP for a single industry, while at the micro level, CSP terminal revenue is only 1/7 of the investment, making breakthroughs on the application side crucial; on the financing side, considering the financing gap for CSPs and the demand from new cloud factories, it is expected that funding pressure will remain relatively controllable until 2027, but the gap will become significant after 2028. In this context, Jensen Huang's departure will weaken NVIDIA's ecological integration and credit guarantee capabilities, significantly increasing the likelihood of changes in NVIDIA's customers, supply chains, and overall strategy, leading to a slowdown in technological iteration in the overseas computing power chain.
### Domestic CSPs: Increased Difficulty in Obtaining Computing Power Disturbs Short-Term Performance Expectations
According to The Information, the U.S. Department of Commerce is finalizing a new AI chip export control rule, planning to include countries in Southeast Asia and others under stricter chip export licensing management.
In the short term, stricter export controls are expected to affect the progress of advanced computing center construction, impacting the continuity of growth. The visibility of short-term CSP growth is reduced, but considering that the demand for computing power has not changed significantly, and after supply constraints, CSP margins are expected to improve significantly, while the ROIC from previous chip purchases will also further increase. Secondly, for model companies, the exploration of the intelligent limits of cutting-edge models still heavily relies on support from advanced computing power. Under limited computing power supply, we believe model companies will further optimize AI infrastructure for training and inference, promoting cost reduction and efficiency improvement.
In the medium to long term, we believe that domestic CSPs will further increase their procurement ratio of domestic computing power. Full-stack platforms with strong heterogeneous computing capabilities are expected to achieve higher efficiency in utilizing domestic computing power, gaining more market share.
### AI Applications: Chip Structure Still Favorable, Application Diffusion is the Main Increment for Model Commercialization in 2027
In our annual strategy, we judge that the certainty of AI applications is increasing: AI Coding has already crossed the critical point, becoming a general foundation for intelligent agents. Leading products are experiencing explosive growth in ARR, with the daily workload of agents at OpenAI equivalent to 3.1 human workloads. The number of active users surged after Codex was merged into the ChatGPT desktop version; in terms of enterprise-level AI, enterprise AI spending and LLM application rates continue to rise, with localized deployment dominating enterprise-level AI demand. Leading manufacturers are promoting the formation of a "deployment---feedback---iteration" flywheel through front-line deployment (FDE) models; in terms of personal intelligent agents, products like Qianwen Go and Meituan Xiaotuan have entered the real task execution phase, and WeChat has opened two AI access modes, supporting intelligent agents to call mini-program services; AI4S and vertical AI have also entered an explosive phase.
In the media direction, we remain optimistic about the prosperity of AI content, especially AI long dramas and AI movies, as domestic video model capabilities iterate and costs decrease. A batch of high-quality content is expected to go online, and we are particularly focused on the industry catalysis brought by the release of the ByteDance World Model in December; on the platform side, WeChat is accelerating its entry into the short drama sector, and upstream IP and content providers are expected to enjoy dividends in stages, with a focus on companies holding quality IP and content. From the perspective of chip structure and transaction volume, the media sector currently has room compared to the first half of 2023 and early 2026, and the chip structure of the sector remains favorable. Fluctuations in other directions, from a funding perspective, are beneficial for the subsequent rise of the media sector.
### Investment Strategy:
From a configuration perspective, the A-share AI hardware industry chain currently does not have sufficient odds to hedge against uncertainty. The proportion of active public TMT holdings remains at 98.4% since 2010, and the sector will shift from being driven by incremental funds to stock or even reduction games. The restoration of the funding ecosystem will take time, with momentum and prosperity strategies weakening and rotation accelerating.
Within the technology sector, we recommend seizing structural opportunities in the following order: 1) Domestic computing power direction, overseas management adjustments and export controls reinforce the logic of localization, and the domestic chain is expected to gradually release performance starting from Q4 2026; 2) Physical AI direction, with giants entering the field and 2027 being the year of implementation, focus on orders from embodied intelligent supply chains and the mass production verification of L4 and L3 autonomous driving; 3) AI application diffusion direction, AI Coding has already crossed the critical point, with AI film and television, AI4S, enterprise-level AI, and personal intelligent agents entering the task execution phase, focusing on the diffusion of application scenarios and commercialization realization.
Additionally, pay attention to the broad structural repair opportunities in non-AI sectors: Non-AI sectors have previously endured pressures from material price increases, rising long-term interest rates, and foreign exchange losses due to the negative externalities of AI. The main process of valuation contraction has been completed this year. As AI investment slows down, interest rates and material prices decline, and the appreciation of the RMB slows, K-shaped differentiation is expected to converge. We recommend a stable return strategy represented by PB-ROE (focusing on white goods, logistics, energy, oil transportation, pharmacy, overseas brokers, and financially stable real estate leaders), as well as a reverse layout and expectation interpretation strategy represented by domestic consumption (focusing on food and beverage, catering supply chains, airport duty-free, health products, and leading medical device companies).