Available data strongly suggests the All‑In Podcast has a large audience, but it doesn’t let us cleanly verify Chamath’s specific 2022 comparison to CNBC and CNN.

What we know about CNBC and CNN in 2022

Nielsen-based rankings compiled by Adweek for calendar year 2022 show:

• Most CNBC shows averaged only ~0.1–0.25 million total viewers: for example, Fast Money Halftime (236k), Squawk on the Street (234k), The News with Shepard Smith (220k), Closing Bell (206k), The Exchange and TechCheck (197k), Power Lunch (195k), Fast Money (174k), Mad Money (148k), Squawk Box (117k), Worldwide Exchange (55k). (adweek.com)
• CNN’s main shows were generally higher: Anderson Cooper 360 averaged 868k total viewers in 2022; other CNN shows like The Lead, Erin Burnett OutFront, Situation Room, Inside Politics, CNN Newsroom and others ranged roughly from ~700k down to ~400k viewers. (adweek.com)
• Network-wide weekly averages for late 2022 put CNN at ~0.43M total-day and ~0.47M primetime viewers, and CNBC around 0.12–0.17M, underscoring that most CNBC programs are in the low-hundreds-of-thousands range, while CNN’s are several hundred thousand to nearly 1M. (barrettmedia.com)

So numerically, “most CNBC shows” are well below ~250k average viewers, while “some CNN shows” cluster around 400–700k+.

What we know about All‑In’s audience

• A February 2023 blog post about All‑In notes that on YouTube alone the show then had 300–400k views per weekly episode and ~23M cumulative views. (thelongerweekend.com) That is YouTube-only and already exceeds the typical 2022 audience of most CNBC programs listed above.
• A 2025 Wall Street Journal profile, citing Podscribe, reports that each episode now averages about 750k downloads plus YouTube views and that All‑In is among the top 100 podcasts in the U.S. (wsj.com) This confirms that by 2025 their per‑episode reach clearly exceeds the 2022 average audience of essentially all CNBC shows and overlaps with or surpasses many CNN programs.

However, none of these sources provide a precise, independently-audited figure for All‑In’s weekly cross‑platform audience specifically by the end of 2022. We can only infer backwards from:

• 300–400k weekly YouTube views as of early 2023, plus
• the general pattern that audio downloads usually add substantially to a podcast’s YouTube audience.

Those inferences make it plausible that, by late 2022, All‑In’s total weekly audience was already larger than the average audience of most CNBC shows, and possibly comparable to lower‑rated CNN shows. But:

1. Metrics aren’t directly comparable. Chamath talks about “more than a million people a week that listen,” i.e., weekly unique listeners across platforms. TV ratings use average concurrent viewers per minute per airing. A show can have 200k average Nielsen viewers but millions of unique viewers over a week; without uniform methodology we cannot rigorously say one “passed” the other.
2. We lack time‑specific podcast numbers. The only concrete third‑party figure near the prediction window is the Feb 2023 blog’s YouTube‑only estimate; the Podscribe 750k‑per‑episode number is from 2025 and doesn’t specify historical 2022 levels. (thelongerweekend.com) Rephonic and similar services gate their exact listener estimates, so we don’t have a verifiable 2022 series.
3. His base comparison to MSNBC/CNN already seems numerically dubious. Public data show MSNBC’s average 2022 primetime audience was around 1.2M viewers, substantially larger than typical podcast‑scale audiences. (en.wikipedia.org) That undercuts the reliability of the original boast and makes it harder to treat his million‑per‑week number as a trusted baseline.

Because of (a) the mismatch in measurement standards, and (b) the absence of audited 2022 cross‑platform audience figures for All‑In, we cannot definitively determine whether by the end of 2022 its weekly audience had actually surpassed the average viewership of most CNBC shows and some CNN shows, even though it likely does so by 2025.

Therefore, the fairest classification is:

• Result: "ambiguous" – enough time has passed, but the available data and incompatible metrics do not allow a clear right/wrong judgment.