Analysis of Investment Opportunities in Sino-US AI Competition Under Jensen Huang's AI 'Five-Layer Cake' Theory

NVIDIA CEO Jensen Huang proposed the AI ‘Five-Layer Cake’ Theory, which provides a systematic framework for understanding the current Sino-US AI competition landscape. The theory divides the AI industry chain into five key layers: Energy, Chips, Infrastructure, Models, and Applications, each containing unique investment opportunities and risks.

Energy has become a key bottleneck in the AI race. In an interview with the Financial Times, Jensen Huang clearly stated that one important reason China will beat the US in the AI race is China’s lower energy costs [1]. Microsoft CEO Satya Nadella also emphasized: “The core contradiction we face now is not excess computing power, but energy bottlenecks” [1].

• Clean Energy Infrastructure:
  Stable power supply systems such as nuclear power, photovoltaic, and wind power
• Energy Efficiency Technology:
  Data center energy-saving solutions, cooling technology
• Energy Storage Systems:
  Grid-level energy storage technology, battery systems

NVIDIA has an obvious leading position in GPU technology; currently, its stock price is $176.12 and its market capitalization reaches $4.29 trillion [0]. The company has more than 80% share in the AI chip market, and its gross profit margin remains at a high level of around 80%.

Facing US chip export controls, China is accelerating the process of domestic substitution, promoting open-source chip architectures and independent R&D.

• GPU Industry Chain:
  Chip design, manufacturing, packaging and testing
• ASIC Chips:
  Custom chips for specific AI applications
• Chip Materials and Equipment:
  Key materials and equipment for semiconductor manufacturing

Jensen Huang emphasized that NVIDIA has transformed into an “AI infrastructure company” and is no longer a pure chip manufacturer [3]. This transformation reflects the dependence of AI development on large-scale infrastructure.

• Large tech companies are expected to invest about $500 billion in AI infrastructure by 2025 [4]
• Supply-demand imbalance persists; NVIDIA stated that “cloud services are sold out” [2]

• Data Center REITs:
  Real Estate Investment Trusts with high-quality data center resources
• Cloud Service Providers:
  Enterprises providing AI computing power leasing services
• Network Infrastructure:
  High-speed network equipment, optical fiber communications

• US:
  Focuses on closed-source large models, represented by the GPT series
• China:
  Leads in open-source models with more abundant application scenarios [1]

China’s open-source model strategy may bring faster application popularization and commercialization processes.

• Vertical Industry Applications:
  Medical AI, fintech, autonomous driving
• Enterprise-level Services:
  AI-driven productivity tools
• Consumer Applications:
  Personal assistants, content generation tools

1. Energy Cost Advantage:
   China’s relatively low energy costs provide a cost advantage for AI training and deployment [1]

2. Manufacturing Foundation:
   A complete manufacturing ecosystem supports large-scale production of AI hardware

3. Abundant Application Scenarios:
   A huge domestic demand market and diversified application scenarios provide broad space for AI commercialization

1. Technological Leadership:
   Maintains leadership in core chip technology and basic research
2. Mature Ecosystem:
   Has a complete AI development and deployment ecosystem
3. Sufficient Capital:
   Venture capital and public markets provide sufficient funds for AI development

• Energy infrastructure stocks: Benefit from energy demand growth driven by AI development
• Chip industry chain leaders: High technical barriers, deep moats
• Data center REITs: Stable cash flow and long-term growth potential

Focus on innovative companies in chip design and manufacturing technology
Focus on commercialization progress in AI application fields
Deploy data center and cloud service providers

1. Policy Risk:
   Sino-US technological competition may increase regulatory uncertainty
2. Valuation Risk:
   Some AI-related stocks are at historically high valuations
3. Technology Iteration Risk:
   AI technology develops rapidly, making it difficult to select investment targets

Jensen Huang’s “Five-Layer Cake” Theory reminds investors that AI investment requires a full industry chain perspective. Under the current Sino-US competition landscape, each layer contains different investment opportunities; the key is to identify the core value drivers and competitive advantages of each layer. Investors should make balanced allocations among different layers according to their own risk preferences and investment cycles to maximize the investment returns brought by the AI revolution.

[1] Jinling AI Data - Real-time stock quotes and financial data
[2] DeepTech - “The Second Half of the Sino-US AI Race: Energy Bottlenecks Become a More Urgent Challenge Than Chips” (https://www.mittrchina.com/news/detail/15470)
[3] TechWalker - “Jensen Huang’s Global Press Conference: NVIDIA is Transforming into an AI Infrastructure Company” (https://www.techwalker.com/2025/0320/3164511.shtml)
[4] Caixin Global - “NVIDIA Earnings Call Transcript: Three Transformations Drive $500 Billion AI Infrastructure Investment Target” (https://www.cls.cn/detail/2205700)
[5] Financial Times - “FT Person of the Year: Jensen Huang” (https://www.ft.com/content/11a018f4-95e0-41c2-99d8-aff105328a0b)