Jeff Bezos built Amazon’s competitive shield by choosing slim margins over fat ones. He framed low profits not as a weakness but as a barrier to big rivals like IBM or Microsoft, who couldn’t risk depressing their overall earnings. In 2007, when analyst Scott Devitt labeled Amazon a buy despite its razor-thin margins, portfolio managers laughed— they saw a nonprofit sham, not a fortress under construction. Behind the scenes, Amazon was flexing its cost-cutting muscle, turning lean operations into a hard-to-replicate engine. That engine isn’t a static barrier. It’s a system that demands constant upkeep: ruthless cost control, process optimization, relentless reinvestment. Unlike patents or brand loyalty, these operational muscles fade if neglected. The same principle shows up in AI. Anthropic’s compute constraints forced sharper engineering, squeezing more performance per watt. Competitors could copy the code, but matching years of efficiency tuning takes real commitment. Traditional moats look like walls you build once. Bezos’s version is more like digging deeper trenches every day. Fat profits invite rivals to spend on R&D and jump in. Low margins scare off most newcomers. Elon Musk dismisses moats in favor of rapid innovation, but real defensibility comes from engines you maintain at full throttle, not walls you hope never break down.