ALL INPredictions
Last updated Nov 29, 2025
Prediction
Chamath
← Back to hostEpisode
Chamath @ 00:23:58Inconclusive
aiventuretech
Over the next 4–5 years (by roughly 2028), large tech companies’ AI platforms and models (e.g., Llama 2, Tesla Dojo/FSD, other foundational models) will be made available on an open‑source or quasi‑free basis, and this will enable the creation of hundreds of new startups, including at least one major breakout company built on top of these freely available AI tools.
all of that stuff will be given away essentially, I think open source, quasi free to the ecosystem over the next few years. And I think that will be a really important moment, which will create hundreds of new companies doing really clever and cool things. So we haven't yet seen the big breakout company yet. And so I think that right now most of this CapEx is going to the big guys. But the dividends of all the work that these big guys are doing will be seen over the next few years in the startups that get started in the next 4 or 5 years.View on YouTube
Explanation

As of November 30, 2025, we are only about two years into the 4–5 year window Chamath specified (extending roughly to 2027–2028), so the prediction cannot yet be definitively judged.

Substantively, several pieces of the prediction are trending in the direction he described:

  • Major tech firms have released “open-weight” or quasi-free large models: Meta’s Llama 2–4 families (source-available with broad commercial rights, though not fully OSI-open-source) and related ecosystem; Google’s Gemma series of open-weight models; and Databricks’ DBRX under an open model license. (arstechnica.com)
  • Additional open models built on these platforms (e.g., TinyLlama, TigerBot, OpenVLA) illustrate the “open / quasi-free platform” dynamic he described. (arxiv.org)
  • Industry leaders like OpenAI have publicly committed to releasing at least one open-weight model, reinforcing the broader move toward more freely usable foundational models. (wired.com)

There has also been a surge in AI startups, some of which build directly on these open or low-cost models, and the tech press explicitly highlights open models like LLaMA and Gemma as key enablers for developers and smaller companies. (techradar.com) However, verifying that “hundreds” of new companies and at least one major breakout startup are specifically attributable to these freely available big-tech models—and judging the final scale of this effect—requires seeing the full 4–5 year period play out.

Because the forecast horizon has not yet elapsed and the ultimate startup outcomes are still unfolding, the correct status for this prediction today is inconclusive (too early to tell).