First Strike

The piece argues that electricity, not chips alone, is the foundation of computing power and geopolitical advantage. It cites a recurring trend suggesting China could have ten times the United States’ electricity supply within a decade. In response, it proposes a long-term US strategy: sell AI chips to China, use the proceeds to acquire solar-panel and battery factories, and rapidly expand domestic power capacity. With robotic labor, immigration, a business-friendly culture, and abundant land, the United States could remain competitive over the long term while generating prosperity for both countries.

Instead, the author sees short-term, zero-sum thinking driving the AI race and asks when that race can end if rivals will eventually match one another’s capabilities. The feared answer is that some decision-makers believe artificial superintelligence could enable a decapitating first strike. The author doubts such a strategy could overcome an opponent’s ability to produce steel and drones at scale, but warns that success would be catastrophic. NVIDIA’s nationalization is cited as a development that has sharply raised the author’s concern about an imminent world war. The central warning is that pursuing decisive military advantage through AI could lead to existentially dangerous choices rather than cooperation and investment in energy capacity.