Sam Altman's strategy for advanced AI models reveals the real business lesson of this moment. While consumer AI commoditizes—free tier ChatGPT, cheaper rivals, subscription fatigue—OpenAI is moving upmarket. The planned models with PhD-level knowledge at annual salary prices per month target institutional buyers: enterprises solving specific, high-stakes problems where incremental accuracy compounds into genuine value. This isn't price gouging. It's market segmentation made explicit.
The automotive industry's architecture shows how this plays out across sectors. Geely, the Chinese conglomerate that swallowed Volvo in 2010, now controls roughly 65% of Polestar—the Swedish brand that started as a racing tuner and pivoted to electric cars. What matters isn't the ownership chart's complexity, but the business model underneath: Polestar owns no factories, no heavy infrastructure. It's pure product and brand. The company generates margin from design and positioning, not capital. This is the lean model working at scale—and it's precisely the template advanced AI services will follow. Infrastructure stays concentrated; value accrues to whoever coordinates it.
The Oscar voting system, seemingly random trivia, actually demonstrates institutional voting design under constraint. Academy members vote only in their domain—directors vote for directing, actors for acting—except for Best Picture, where everyone votes with ranked preferences. The system works because it solves a coordination problem: avoid tyranny of the majority while respecting expertise. This structure will matter more as AI becomes central to business decisions. Institutions won't pay $20,000 monthly for a model they don't trust their experts to evaluate. Stratification requires asymmetric trust, not just asymmetric pricing.