Speaker: Jason Smith, reporting from Beijing, China.
Core Argument: China has already won the AI race—not in five or ten years, but now—by solving the fundamental bottleneck of affordable, clean electricity at a scale the United States cannot match.
The video’s central thesis is that artificial intelligence is, at its core, an energy problem. Every AI breakthrough—ChatGPT, DeepSeek, Grok—is “electricity in disguise.” The nation that controls the most abundant, cheap, and reliable power will ultimately dominate the AI landscape. China, the speaker argues, has quietly built that foundation while the rest of the world remains distracted by model benchmarks and quarterly earnings.
⚡ Energy as the decisive bottleneck
- In 2025 alone, China added ~540 GW of new electricity capacity—eight times the US total for the same year.
- Over 84% came from clean sources (solar, wind, hydro). China now leads the world in all three categories.
- In the first half of 2025, renewables exceeded China’s growing demand, cutting fossil fuel generation by 2%—meaning China has already passed peak fossil fuel emissions.
🇺🇸 US challenges and structural disadvantages
- US data centers are driving up local electricity bills. Goldman Sachs notes that middle-class families in Virginia, Texas, and Georgia are effectively subsidizing the AI boom through higher power costs and food prices.
- When communities resist, data centers simply relocate to poorer areas that lack resources to fight back—a pattern that is neither sustainable nor equitable.
🌊 China’s engineering innovations
- Commercial underwater data centers off Hainan Island use constant cold seawater for cooling, reducing energy consumption by 40–60%.
- Pods are placed next to offshore wind farms, eliminating transmission losses and land-use conflicts.
🏗️ Investment priorities: war vs. infrastructure
- The US spent $8 trillion on post-9/11 wars, with an annual defense budget exceeding $900 billion.
- China directed comparable resources into high-speed rail, 5G/6G, battery gigafactories, quantum labs, and massive energy grids.
🔓 Open-source software as a force multiplier
- China’s top models—DeepSeek, Alibaba’s Qwen series—are fully open source, creating a global developer ecosystem.
- US frontier labs remain locked behind paywalls and NDAs. The video argues that open collaboration accelerates iterations and network effects, while restrictive models favor short-term profits.
💰 Wall Street’s blind spot
- US stock markets, especially the “Magnificent Seven,” trade on the assumption of permanent American AI leadership.
- When markets finally price in China’s cheaper energy, greener data centers, and faster innovation loops, the repricing could be severe.
🤝 A possible path forward for the US
- The speaker suggests pragmatic partnerships: allowing Chinese solar companies to colocate factories in Arizona, Texas, or Nevada; modernizing the US grid with Chinese inverters and battery storage.
- The real obstacle, he claims, is not capability but entrenched energy lobbies and outdated national-security theater.
Final takeaway: The AI future belongs to the country that can keep the lights on tomorrow. China has already built that foundation—abundant clean power, underwater data centers, open models, and disciplined long-term investment. The question is no longer whether China has won, but whether the rest of the world will acknowledge the reality and choose to run alongside rather than remain in denial.