AMD (Advanced Micro Devices)

AMD's stock was at $3 in 2014, its CPU architecture was failing, and extinction seemed inevitable. CEO Lisa Su made the ultimate company-betting decision: kill the roadmap, commit entirely to an untested architecture called Zen, and ask customers to wait years for the payoff. She told them: 'I'm stopping my old roadmap because it's not good enough.' The result: 70x stock return, server market share from <1% to 33%, and Intel's first serious competition in decades.

The victory wasn't just engineering - it was strategic jiu-jitsu. AMD went fabless, outsourcing manufacturing to TSMC and achieving process parity with Intel for the first time in 40 years. The chiplet approach allowed economics that Intel's monolithic chips couldn't match. This is how challengers defeat incumbents: not by copying their playbook, but by rewriting the rules of competition entirely.

The biological lesson is punctuated equilibrium: long stasis broken by radical change when survival demands it. Most dying companies optimize incrementally - 5% cost cuts, modest product improvements, 'strategic reviews' - and die slowly. Su made an irreversible bet on completely new architecture with 3-year development time and zero revenue. Existential pressure drove transformative innovation because there was no fallback plan. You can't pivot to radical change if you're hedging.