All-In Summit: In conversation with Graham Allison

Allison sees U.S.–China war as structurally dangerous but not imminent. He said 12 of 16 historical power transitions ended in war, yet put the odds of avoiding war at 99% over one year and 90% over four, while warning that the longer-term trajectory is far darker.

His most plausible path to war is an unintended Taiwan escalation, not a Chinese decision to manufacture conflict for economic reasons. He compared it with 1950 Korea, when U.S. forces approached China’s border and Mao intervened despite neither side beginning the year seeking war.

Allison says Washington’s central China error was failing to imagine how successful China could become. He recalled China rising from roughly 5–10% of U.S. GDP around 2000 to about three-quarters, while post-Cold War policymakers assumed integration would not produce a peer-scale challenger.

He thinks avoiding war will eventually require the United States to tolerate some Chinese sphere of influence in Asia. His argument is geographic: China already has greater nearby military weight, while Taiwan lies about 90 miles from China but halfway around the world for U.S. forces.

Allison’s balancing strategy is alliances, because he doubts the United States can outmatch China alone if China’s rise continues. He said America would lose a purely bilateral long-run contest, but India, Japan, Australia, South Korea and other aligned states could shift the balance.

He flatly said the United States is not equipped for a hot war with China. Allison added that a clean-sheet Pentagon could, in his view, deliver twice the military effect for half the money—an unusually severe criticism from a former senior Defense Department official.

Allison rejected the idea that U.S. foreign-policy mistakes reflect a single, uncontested establishment consensus. He cited Iraq: Brent Scowcroft, George H.W. Bush’s closest adviser, warned George W. Bush against invasion and published the case; Allison said Bush 41 or Al Gore would not have invaded.

Allison sees Xi Jinping’s concentration of power as a structural weakness for China, not simply a source of strength. He said loyalty can outrank competence and described Chinese AI as strong in science and math but constrained from answering political questions that conflict with Xi Jinping Thought.