Chris Painter, president of METR, argued that if one frontier AI developer pulled far ahead, the rest of American industry would act as an offshore balancer and provide large subsidies to sustain a competitive second developer. Andrew Trask of Google DeepMind replied that concentration of power risks are enormous and could extend across industries, political systems, geographies, and ideologies. Trask questioned whether even a duopoly would suffice given AI's broad reach and called for evaluation regimes designed with sensitivity to avoid further concentrating power. He attached charts expressing skepticism about the balancing scenario.