Jensen Huang foresaw the future of AI a decade in advance, yet once instructed NVIDIA’s executives to never think about one question: What should the company do if Taiwan is blockaded or invaded? This is not a joke—it’s a detail revealed by journalist Stephen Witt in his book, *The Thinking Machine*. NVIDIA’s most advanced chips rely heavily on TSMC for manufacturing. If conflict erupts across the Taiwan Strait and semiconductor fabs are unable to operate normally, NVIDIA’s core supply chain could suffer catastrophic damage. When Witt asked NVIDIA’s Executive Vice President of Operations, Deb Shoquist, how the company prepared for such a scenario, her response was: “It’s like asking me what I’d do if California sank into the ocean.” She also revealed that Huang had explicitly told her to “never think about this question.” Even more startling to Witt was that, at least at the time of his interview, NVIDIA had no contingency plan whatsoever for such a scenario. This episode has recently been revisited by diplomatic policy scholar Julian Gewirtz, who first discussed it earlier this year in *The New York Review of Books*. Interestingly, it was precisely Huang’s decade-ahead strategic insight that positioned NVIDIA at the center of the AI era. Yet when it comes to the company’s most lethal geopolitical risk, he appears to have adopted a different logic: if you can’t control it, don’t think about it. Of course, no contingency plan can magically create a second TSMC. But the purpose of risk management isn’t to guarantee that disaster won’t happen—it’s to know in advance what options remain when disaster strikes. Now, NVIDIA’s board has launched an independent investigation following reports that Huang may be targeted by foreign intelligence operations. While there is no confirmed direct link between the two events, they compel a renewed examination of the company’s governance. When a company’s most critical strategic decisions revolve entirely around a single visionary CEO, his judgment becomes its most valuable asset—and the questions he refuses to consider may become the organization’s blind spots. Jensen Huang can decide which future NVIDIA bets on—but he cannot decide which future will not occur. Julian Gewirtz’s original post · Original article in *The New York Review of Books*