Available audience data show that the All-In podcast has grown substantially since 2021 but has not reached anything close to ~10 million weekly listeners, even by late 2025.

• A July 2025 Wall Street Journal profile, citing Podscribe data, reports that All-In averages about 750,000 downloads and YouTube views per episode across platforms. (wsj.com) With roughly weekly episodes, that implies on the order of hundreds of thousands—not tens of millions—of listeners per week.
• A detailed reference entry summarizing Podscan analytics estimates “more than 5 million monthly listeners” for the show, with frequent top rankings in the Technology category on major platforms. (leviathanencyclopedia.com) Even if all of those monthly listeners are distinct, that is still well below 10 million weekly listeners.
• A feature in The Information notes that the hosts have “multiplied their weekly audience on YouTube by an order of magnitude” and that the show “regularly attracts hundreds of thousands of listeners per episode across platforms,” but it does not claim anything near 10 million per week. (theinformation.com)

On the influence / distribution part of the prediction, mainstream coverage supports the idea that they now have direct distribution of their ideas to millions of people:

• The Washington Post describes candidates using All-In to reach “millions of listeners” and portrays the show as a key vehicle for tech and political messaging. (washingtonpost.com)
• A 2025 Vanity Fair profile calls the podcast a “kingmaker pulpit” for Silicon Valley elites and notes that their summit and show have “millions of their fans,” underscoring their ability to inject ideas directly into a large audience without traditional media gatekeepers. (vanityfair.com)

So, Chamath was directionally right that the show would become a powerful direct channel to millions of people. But the specific quantitative claim—growing from about 1 million weekly listeners to roughly 10 million weekly listeners—has clearly not materialized based on the best third‑party estimates available as of late 2025. For that reason, the prediction as stated is rated wrong.