China Chips Still Behind US, Nvidia in Performance: Baillie Gifford

Key Concepts

• HBM (High Bandwidth Memory): Specialized memory technology essential for AI processing, now shifting from a commodity to a "lock-in" product.
• Semiconductor Supply Chain Complexity: The multi-national, highly interdependent nature of chip manufacturing (design in the US, fabrication in Taiwan, lithography in the Netherlands, memory in Korea).
• Structural Opportunity: Long-term investment thesis focusing on companies that deepen their competitive moat as they scale.
• Bottleneck Investing: The strategy of identifying supply chain constraints and investing in the companies that control those critical nodes.
• EUV (Extreme Ultraviolet) Lithography: Advanced technology used by ASML to produce the most complex semiconductors.

---

1. Nvidia and the China Market Strategy

The current investment strategy for Nvidia assumes zero access to the Chinese market.

• Regulatory Dynamics: While the US government has approved the export of H200 chips to China, the primary barrier is the Chinese government’s internal decision to restrict purchases by domestic firms.
• Long-term Threat Assessment: The speaker argues that China’s attempt to build a domestic chip supply chain is unlikely to threaten Nvidia’s dominance for at least five years due to:
  • Performance Gap: Existing Chinese high-power chips remain a fraction of the performance of Nvidia’s leading-edge products.
  • Supply Chain Complexity: Unlike EVs or batteries, semiconductor manufacturing requires global cooperation and extreme technical precision that cannot be replicated solely through domestic capital injection.

2. Memory Bottlenecks and SK Hynix

The semiconductor industry is currently facing a memory bottleneck, a challenge Jensen Huang (Nvidia CEO) anticipated three years ago by urging memory makers to invest in permanent capacity.

• Shift in Industry Dynamics: Historically, the memory industry was "boom and bust" and highly cyclical. The speaker notes a fundamental shift:
  • Long-term Agreements: SK Hynix is now securing long-term purchase agreements, reducing volatility.
  • Customization: The rise of custom HBM creates "customer lock-in," moving memory away from being a commoditized product.
• Investment Thesis: SK Hynix is viewed as a company that is becoming structurally better as it scales and deepens its position in the AI ecosystem.

3. TSMC: The Ecosystem Coordinator

TSMC remains the largest position in the portfolio due to its unique role in the global supply chain.

• Dominance: TSMC demonstrates "extreme control" over the supply chain, acting as the central coordinator for multiple global actors (e.g., ASML, equipment suppliers, and chip designers).
• Growth Guidance: TSMC has raised its AI-related growth guidance to 56% over the next five years, a figure the speaker cites as a high-conviction indicator of the industry's trajectory.

4. Global AI Investment Perspective

The speaker emphasizes a "dichotomy" in the AI sector: while AI is largely designed in the US, it is manufactured internationally.

• Geographic Diversification: Investors are encouraged to look beyond US borders to find "rich hunting grounds" for AI-related names.
• Key International Players:
  • TSMC (Taiwan): Fabrication and supply chain coordination.
  • SK Hynix (South Korea): Memory integration.
  • ASML (Netherlands): EUV lithography machine production.

5. Market Sentiment vs. Fundamentals

The speaker addresses the recent market volatility (near-correction territory in the chip sector) and the tendency for markets to misinterpret a "lack of news" as a negative catalyst.

• The "Catalyst" Fallacy: The market reacted negatively to the lack of news regarding chip exports during the US-China summit. The speaker argues this is irrelevant to the long-term fundamentals of the companies involved.
• Synthesis: The core investment philosophy is to ignore short-term market noise and focus on companies that possess structural advantages, deep competitive moats, and the ability to navigate an increasingly complex global ecosystem.

Conclusion

The main takeaway is that the AI revolution is a global, interdependent phenomenon. While geopolitical tensions regarding China create uncertainty, the fundamental strength of companies like TSMC, SK Hynix, and ASML—driven by their critical roles in a complex, non-commoditized supply chain—provides a robust long-term investment case that transcends short-term market fluctuations or political headlines.