CharlieKey takeaways from your favorite podcasts
← 30 October 2026TBPN

Google Earnings, Jensen Targeted in Chinese Spy Plot, The Rise of Fake Wall St. Traders | Diet TBPN

Google's earnings miss masks a booming cloud business and a hard AI-search monetization problem, while a Chinese spy plot targets Nvidia's chip supply and fake simulated-trading firms profit from aspiring day traders.

Watch on YouTube ↗

Charlie's summary

  1. Google Cloud has become a monster business that was nearly shut down, and its backlog should lift valuation

    The hosts note that Larry Page once considered pulling the plug on Google Cloud, per 2018 reporting. Now revenue doubled to $30.7 billion, operating profit tripled to $11.1 billion, and margin is 36.2%. The backlog is $561 billion, and 55% should be recognized within two years, roughly $275 billion. Standalone, they say, it could be worth hundreds of billions, maybe a trillion dollars.

    ▶ 3:33
  2. AI search is growing usage but is harder to monetize than ad-packed traditional search

    Search ad growth slowed from 17% to 12% as AI Mode takes share. The host argues Google wants to monetize every pixel, and a decade of search has become mostly ads. AI Mode doesn't yet have “every single sentence” as a sponsored ad. Figuring out how to auction links, offers and brand messages inside a full AI response is the coming challenge.

    ▶ 5:35
  3. Owning the frontier model makes Google vertically integrated, which could be strategic even without showing up in API data

    Google guides to about $200 billion of 2026 capex, with a significant increase expected in 2027, and Gemini passed a billion monthly users. The host argues that having its own frontier model internally drives YouTube, AI search and internal workloads without outside costs. It can fine-tune models for its infrastructure, such as fast responses for search. Both Google and Microsoft undershot capex estimates because of delays, not lack of will.

    ▶ 7:36
  4. Games and simulated environments may be the path to better reasoning, and Google stayed the course with Genie

    A rumor cited on air says an internal model's reasoning jump comes from more aggressive RL, including training in games and simulated environments, with strategies transferring to math and research. The hosts note Google kept investing in Genie while Anthropic focused on coding and OpenAI chased side quests. They point to computer use, once the janky Operator, as a side bet that paid off. Whether Genie follows is the open debate.

    ▶ 12:46

Give your email to unlock the rest of the key points — for every episode.

Do more with Charlie

Free — sign in with your email, no password.