ALL INPredictions
Last updated Nov 29, 2025
Episodes / E160: 2024 Predictions! Markets, tech, politics, and more
Back
Jason @ 00:58:04Wrong
ventureaimarkets
By the end of 2024, private‑market valuations for LLM‑focused startups will have fallen roughly 50–80% from their 2023 peak implied valuations, making them one of the worst‑performing asset categories, even if these markdowns are not fully reflected on company cap tables.
My worst performing asset in 2024 is LLM startups... I think they're all going to come down by, you know, 50, 60, 70, 80% in terms of their valuations that won't get marked in their books, but that will be the reality of where their stocks will trade on the private markets.View on YouTube
Explanation

Evidence from 2023–2024 private funding rounds and sector-level data shows the opposite of what Jason predicted.

  1. Flagship LLM startups saw big up-rounds, not 50–80% markdowns, between 2023 peaks and end‑2024.

    • Cohere was valued around $2.1–2.2B in June 2023, then raised a new round in July 2024 at $5.5B, more than doubling its valuation instead of falling 50–80%. (en.wikipedia.org)
    • Mistral AI was valued at “over $2B” in Dec 2023 and then raised a June 2024 round at €5.8B (~$6.2B), roughly tripling its valuation in a year. (en.wikipedia.org)
    • OpenAI completed an employee tender in late 2023 valuing it at $86B, and a follow‑on deal reported in Feb 2024 again valued it around $80–86B, i.e., holding or rising rather than dropping by half or more. (cnbc.com)
      Across the best‑known LLM startups, primary and secondary transactions in 2024 point to higher or at least sustained valuations vs. 2023, not a broad 50–80% reset.

  2. Sector‑wide VC data shows GenAI/LLM companies were among the hottest areas of investment in 2024, not one of the worst‑performing asset categories.

    • An EY generative‑AI deal study (Dec 2024) reports that GenAI VC investment continued to surge in 2024, with total 2024 deal value expected to exceed $45B, up sharply from 2023, and explicitly describes GenAI as leading the way within AI funding. (ey.com)
    • Barron’s notes that overall VC funding in 2024 was still well below 2021, but AI investments grew over 80% year‑on‑year and accounted for close to one‑third of all VC funding, underscoring that AI/LLM startups were a central bright spot, not a laggard. (barrons.com)
    • A 2025 ecosystem report summarizing 2024 shows GenAI grabbing nearly half of all global VC money by end‑2024, characterizing GenAI as “the hottest investment trend,” again inconsistent with it being among the worst‑performing asset classes. (startupwired.com)

  3. Other sectors, not LLMs, are singled out as 2024 underperformers.
    A global startup‑ecosystem report notes that sectors like Cleantech, Edtech, AdTech, and Digital Media saw funding and exits fall sharply in 2024, and describes their outlook as “rather gloomy,” while predicting AI will “remain the major growth sector” and continue to draw increasing VC capital. (startupgenome.com) That’s directly at odds with LLM startups being among the worst asset categories.

  4. No credible data supports a broad, hidden 50–80% private‑market markdown for LLM startups by end‑2024.
    There were certainly many overhyped or weaker AI startups and some quiet down‑rounds or shutdowns, but available funding data and secondary‑market valuations for leading LLM companies show strong or rising prices, not a sector‑wide crash. If LLM startups had truly been marked down 50–80% across the board and become one of the worst‑performing asset classes in 2024, we would expect to see widespread reports from PitchBook, Carta, secondary‑market platforms, and major financial media. Instead, contemporaneous reporting frames GenAI as the standout growth and capital‑magnet sector of 2024.

Given the observable data by December 31, 2024, Jason’s prediction that LLM startups would broadly fall 50–80% in valuation and become his worst‑performing asset/one of the worst asset categories did not materialize.