The REAL Reason AI Companies Want Regulation

Martin argues that AI companies’ calls for regulation may protect businesses without durable moats, not primarily humanity, and that safety-testing requirements could favor U.S. providers while excluding cheaper overseas models. He points to trucking deregulation, which he says immediately cut prices by about 25%, as a warning that rules protecting incumbents can leave consumers paying more; he also highlights the tension between portraying AI as dangerous enough to regulate and promoting its claimed $2 trillion value to investors.

The presenter says roughly 40% of the S&P 500’s value is concentrated in 10 companies, nine tied to AI, and that AI drove about three out of every four dollars of new U.S. economic growth in early 2026. Martin says hyperscaler capital spending is nearing $1 trillion a year, with roughly 80% going to commodities and raw materials, including copper, nickel, silver, lithium, and iron ore.

Martin says AI providers face price pressure because customers can shift spending among models with relative ease, unlike businesses protected by stronger moats such as Apple’s switching costs or Visa’s network effect. He cites Anthropic’s share of spending among companies using both Anthropic and OpenAI falling from about 75% to an even split, alongside a decline in American models’ share of OpenRouter traffic from roughly 70% to 30% as users turned to cheaper Chinese models.