I/O Fund's Kindig Reacts to More China Chip Curbs

Key Concepts:

• Technology restrictions on China
• NVIDIA's revenue and China market
• National security implications of AI technology
• Onshoring of semiconductor manufacturing
• AI chip demand and supply chain
• Enterprise vs. consumer technology
• GPU generation and product roadmap
• Custom silicon and inference
• Astera Labs

1. US Restrictions on Technology Access to China:

• Despite NVIDIA's desire to continue selling into China, the US government is imposing restrictions on technology access due to national security concerns.
• The US government views this technology as having far-reaching implications beyond consumer or enterprise use.

2. NVIDIA's Perspective and Investor Sentiment:

• NVIDIA doesn't want to lose the revenue from the massive Chinese market.
• From an investor standpoint, the US government's efforts to control this technology are seen as a bullish signal, indicating strong demand.
• Demand concerns are minimal, regardless of NVIDIA's ability to sell into China.

3. Supply Chain and Onshoring Efforts:

• The US has been preparing for potential supply chain disruptions for years, as evidenced by Taiwan Semiconductor Manufacturing Company (TSMC) building a fab in Arizona.
• The US is considered to be more prepared than assumed to onshore semiconductor manufacturing and maintain domestic supply chains.

4. Impact on American Companies:

• Despite concerns about the impact on companies in Malaysia and Thailand, the potential initiative primarily affects American companies.
• There is little concern over the China impact on this specific technology due to its outsized demand.

5. Enterprise vs. Consumer Technology:

• Unlike consumer technologies like Tesla and Apple, which heavily rely on the Chinese market, AI technology is primarily enterprise-focused and includes sovereign AI at the government level.
• The need for China is not as critical as it is for consumer-driven technologies.

6. Demand Signals and Tracking:

• Key demand signals are tracked through GPU generation, product roadmaps (AI and video), and supply chain partners.
• Current data suggests NVIDIA will remain on top in 2023 and 2024, with strong demand extending into the second half of 2025.

7. Oracle and Cloud Offerings:

• Oracle's cloud offering in Malaysia may face challenges due to potential limitations on access to GPUs, requiring political negotiation.

8. Hardware and Custom Silicon:

• Hardware remains a promising area, with a resurgence of custom silicon for inference.
• Astera Labs is positioned between NVIDIA and companies like Amazon, providing custom silicon solutions.

9. Astera Labs as an Example:

• Astera Labs is highlighted as a company benefiting from the demand for custom silicon solutions, bridging the gap between NVIDIA's offerings and specific needs of companies like Amazon.

10. Conclusion:

• Despite US restrictions on technology exports to China, the demand for AI chips remains strong, particularly in the enterprise sector. Onshoring efforts and the rise of custom silicon solutions are mitigating potential supply chain risks. Companies like NVIDIA and Astera Labs are well-positioned to capitalize on this trend.