'Technological revolutions make geopolitical revolutions.' From that opening axiom, Hal Brands assembles eighteen chapters that ask whether artificial intelligence will follow the geopolitical script of steam, steel, and the atom or tear it up. Brands, who writes both bookends, answers firmly: stability concepts inherited from the Cold War 'conceal strategic traps', and 'the bedrock of stability in the AGI era is US advantage'. The strongest chapters carry real evidentiary weight. Jacquelyn Schneider distills experimental and wargaming evidence, her own included, into an anatomy of escalation in which 'fear is far less likely to lead to preemption than anger or overconfidence', and chapters on first-mover advantage, export controls, and cyber operations supply empirical ballast unusual in a policy collection. Yet the volume argues with itself, and its contributors often complicate the editor's stability-through-advantage thesis. Mark Leonard's penultimate chapter challenges its load-bearing premise: 'AI will get commodified, erasing any first-mover advantage', while the deeper contest concerns 'which conception of the human gets encoded into the digital infrastructure of the twenty-first century'. If the lead cannot be held, the bedrock dissolves. Brands's introduction registers Leonard only as an alliance-rupture scenario, not the commodification challenge; his concluding chapter itself holds that first movers are 'eventually overtaken', turning that erosion logic on rivals' leads while sparing the bedrock claim. The conversation also never leaves the Atlantic policy world: no Chinese, Indian, or Global South voice appears in the book.. Even so, this is the most serious single-volume mapping of AI's strategic stakes yet assembled, and its unresolved quarrel, whether American advantage underwrites order or merely renames the contest, is the question peace research now inherits.