The prediction is explicitly about the next 3–4 years from April 2023, i.e., approximately April 2026–April 2027. As of today (November 30, 2025), only about 2.6 years of that window have elapsed, so the full period Chamath was talking about has not yet played out.

Available evidence is mixed and not decisive either way:

• Venture markets have clearly shifted away from the 2020–2021 mega‑bubble: reporting in 2023–2024 shows harder fundraising, smaller checks, and underperformance/strain at large multi‑stage and late‑stage funds, with commentary that big VC funds have been underperforming smaller ones and are struggling to raise and deploy capital efficiently. (institutionalinvestor.com) This is directionally consistent with the idea that very large funds can be oversized in a tougher environment.
• At the same time, large funds still dominate total capital raised across private markets: one fundraising review finds that funds larger than $1B accounted for about 81% of total capital raised in 2024, indicating that LPs continue to back very large vehicles despite headwinds. (aranca.com) And within AI specifically, major AI‑focused funds have been raised in the high hundreds of millions (e.g., Radical Ventures’ nearly $800M AI fund in 2024), which is close to the $1B scale Chamath argued would be oversized. (ft.com)
• The average U.S. venture fund size fell to around $105M in early 2024, with fewer billion‑dollar-plus funds closing, and press coverage explicitly framing this as a “moment” for smaller and mid‑sized funds—even while giants like a16z still raise multi‑billion vehicles. (wsj.com) That suggests partial support for the idea that smaller funds may be better matched to the current opportunity set, but it does not validate a specific “right” number like $50M, nor prove that $1B is categorically too large.
• Chamath himself has since emphasized pulling back from traditional VC fundraising and largely investing his own sizable balance sheet, criticizing the large‑fund VC model as misaligned—behavior that is consistent with his skepticism about big funds, but not by itself proof that his numerical sizing call was correct. (institutionalinvestor.com)

Because (1) the time horizon he specified has not yet finished, and (2) there is no clear, agreed‑upon metric that would already let us declare that $1B vehicles are definitively “oversized” or that ~$50M over four years is empirically “right” for new-company formation in a generative‑AI era, this prediction cannot yet be judged as right or wrong. It is therefore best classified as inconclusive (too early).