Sources: Elon Musk claimed his new startup xAI was worth $20B while he was trying to recruit top AI talent, including by offering equity in xAI
what could possibly go wrong? Hisan Kidwai / Android Headlines : Elon Musk launches new AI company xAI to understand the universe Bluesky: Emil Protalinski / @epro.social : who said you need crypto to create money out of thin air [embedded post] E.W. Niedermeyer / @niedermeyer.io : If xAI has a valuation that means it has projected cashflows, which in turn implies that there are specific ways of making money by “solving math,” “uncovering the true nature of the universe,” and “creating godlike superintelligences” in the company's plans. — So... probably merch, right? [embedded post] Romit Mehta / @romitmehta.com : It's a question of when, not if, he will lose his luster. After he has been exposed multiple times as the owner of Twitter (the technology and the business), I wonder who will continue imagining he is worth investing in. [embedded post] Twitter: Louise Matsakis / @lmatsakis : Two of Musk's new AI hires, Yuhuai Wu and Christian Szegedy, just confirmed my reporting with @ReedAlbergotti that xAI is working on a model that can do math https://www.semafor.com/... @mikeisaac : i guess to be fair he is not wrong in assuming there will be a legion of investors willing to pay insane premiums based on little more than the fact he's running it @mikeisaac : if all it takes to recruit people is making up imaginary numbers to tell people they could maybe be rich then i am on the wrong career path [image] @rmac18 : “claimed” + “if” + “came up with” https://www.semafor.com/... [image]
Context & Ripple Effects
When Elon Musk formed xAI in July 2023 with researchers drawn from Google, Microsoft, Tesla, and OpenAI, he was selling recruits equity in a company with no product — and, per Semafor, telling them it was already worth $20B. The hires working on math-capable models, including Christian Szegedy, were being asked to price that claim themselves.
The subsequent record gives the claim an unusual afterlife: within five months xAI filed to raise up to $1B (SEC filing) after taking in roughly $135M, then secured $500M in commitments targeting a $15B–$20B valuation (January 2024 round) — landing almost exactly on the number Musk quoted recruits. By September 2025 it had raised $10B at a $200B post-money valuation (latest round).
First-order effects
- Recruits weighing xAI equity offers had to underwrite a valuation asserted by the founder before any model shipped, making Musk's own credibility the collateral for the compensation package.
- X investors being invited into the January 2024 round gave the raise a built-in buyer base tied to Musk's platform rather than to xAI's standalone financials.
Second-order effects
- Competing labs recruiting the same Google, DeepMind, and OpenAI alumni now bid against equity priced off a founder-declared number, pushing rival offers toward narrative-based valuations rather than revenue multiples.
- Each successful round at or above the claimed figure — $15–20B in early 2024, then $200B by late 2025 — retroactively validates the recruiting pitch, compounding Musk's ability to attract both talent and capital ahead of labs with comparable headcount.
Third-order effects
- If the pattern holds, top-tier AI lab valuations get set primarily by founder narrative and distribution reach rather than product milestones, structurally advantaging operators who control large audiences and capital networks.
- It also normalizes a widening gap between a lab's stated mission ('understand the true nature of the universe') and its financial framing, which regulators and limited partners may eventually have to price as a distinct risk.
The trend: Frontier AI valuations are increasingly anchored to founder-declared numbers and platform distribution rather than revenue, with each closed round validating the next recruiting pitch.