Brian and Brandon debate whether Anthropic and OpenAI can grow into their enormous compute commitments now that the Anthropic IPO has put real numbers on the table. Brian shares AI-built models showing Anthropic would need revenue growth of 70% in 2027, tapering to 30% by 2030, to reach break-even, which he argues is not unreasonable compared with AWS in its early days, while OpenAI would need 175% growth this year and 100% in 2027. The harder part, he says, is scale: the models project both companies approaching a trillion dollars in annual revenue within a decade, yet no company in the world reaches that today, with Amazon and Walmart leading at just over $700 billion. Brandon says the bet rests on models improving fast enough to bring in users who find today's tools too hard, and argues it has never been easier to start a business. Brian questions how a company built around Claude Code reaches industries like retail, energy, and healthcare, and says he wants Anthropic to break out how much spend comes from a new "Claude economy" of gig work versus massive enterprise workloads. Brandon proposes four buckets for that spend: new businesses, SaaS displacement, job automation, and a "miracle bucket" of scientific discovery. Both then question whether the market can support two trillion-dollar AI labs. Brian notes markets usually split about 70/30 between leader and runner-up, and Brandon suggests the bitter rivalry between the founders could end with one lab far ahead, or with a third player nobody sees coming.
SHOW: 1082
SHOW VIDEO: https://youtu.be/13tLIfpjcRo
SHOW SPONSORS:
SHOW LINKS
GUEST BIO
N/A this week
FEEDBACK?