Insight | Falling behind China on deep tech and chips
September 29, 2026 · 18 min read
Ryan Graves warns about China's deep-tech push, Dylan Patel and Nathan Lambert explain chips, export controls and Taiwan, and Musk frames U.S.-China competition. Technology leadership is now a core question of national security.
The rules, numbers and steps worth keeping in front of you.
- America's AI lead depends on Nvidia's designs and TSMC's factories in Taiwan (Dylan Patel).
- Nathan Lambert: export controls limit how much AI China can run, not whether it can train models.
- Patel: controls are a timing bet; if AI is not transformative in five to ten years, they help China long term.
- Patel: DeepSeek's low cost figure covered only a final training run, not research or experiments.
- Patel estimates 200,000 to 300,000 GPUs were routed to China in a year through third countries.
- Patel: bringing chipmaking fully home would take a decade and a trillion dollars.
- Subsidy gap, per Patel: about 50 billion dollars under the CHIPS Act against roughly 200 billion a year in China.
- Patel: TSMC's Arizona fabs would stop working within a year or two without Taiwan.
- Patel: cutting China off may push it toward a blockade or takeover of Taiwan.
- Elon Musk: China's economy may reach two or three times America's.
- Musk's bottleneck sequence: chips, then voltage transformers, then electricity.
- Patel: data centers could reach 10% or more of U.S. electricity by the end of the decade.
- Ryan Graves: state money funds long-shot science that venture capital avoids; his examples are the most speculative.
- U.S. firms still lead in total compute; the worry is a fragile lead, not a lost one.
This is the short form. The full insight has every expert's reasoning, their own words and the sources.
