SpaceX’s IPO Triumph, Anthropic’s Fable Fumble, OpenAI’s Price War

TL;DR

• 1

  SpaceX may have pulled off the first true AI IPO: A year after barely having an AI angle, SpaceX used a reported $15 billion Anthropic deal and $30 billion Google deal to pitch AI as a $26.5 trillion TAM and debut above $2 trillion, making Elon Musk the world's first trillionaire.

• 2

  The IPO was engineered for maximum pop: Ranjan points to the tiny 4.3% float, the rapid narrative shift from rockets and Starlink to data centers, and $550 million in banking fees as signs this was a tightly orchestrated launch, not a normal public offering.

• 3

  Anthropic's Fable rollout backfired because the restrictions looked absurdly broad: Users reported refusals on topics like mitochondria, math, chemistry, and even Fable's own system card, turning what Anthropic framed as safety into a public argument about censorship, product design, and possibly marketing.

• 4

  There may be a two-tier AI strategy hiding under the Fable mess: Alex floats the idea that a nerfed public model could make premium, less restricted access to Mythos more attractive for big-paying customers inside Anthropic's Glasswing-style enterprise circle.

• 5

  OpenAI appears ready to cut prices hard, but that could damage its premium position: With executives suddenly focused on token costs and OpenAI reportedly weighing major cuts, Ranjan argues a race to the bottom would be risky for companies that still lose billions and do not have software-like margins.

• 6

  The biggest value may sit above the model layer now: Both hosts land on orchestration, routing, and agent harnesses like Claude Code and co-work as the sticky layer, with Alex saying his biggest recent 'holy crap' moments came from medium-power models doing spreadsheet, login, and event-admin work.