Sam Lessin: xAI-X merger is AOL Time Warner with a debt noose, and mid-market AI players are overvalued

Summary

Sam Lessin's read on the xAI-X merger is straightforward: it's AOL Time Warner, except Time Warner comes with a $12 billion debt noose.

The deal is, in his view, vintage Elon. ZIRP made Musk the greatest capital raiser in history — build a long-term vision requiring infinite capital, be the world's best marketer, raise infinite free money. Now that ZIRP is over, AI narratives are the new free money, and Elon has simply moved there. X and xAI were always going to merge: they share a name, one distributes the other, and they're layered on top of each other. The consolidation also lets Elon absorb the Twitter debt into a combined entity, relieving personal liability he'd been carrying.

Who wins from the deal

Lessin sees it as broadly win-win for the original Twitter investors. Those who bought in at $44 billion watched the valuation erode, but now they can exit into a combined entity with an AI premium attached and get out close to where they entered. It's the SolarCity-Tesla playbook: the strategic rationale was thin — "we'll put solar panels on the roofs of Teslas" — but when numbers go up, all sins are forgiven. Elon is, above all, a master of number go up.

The xAI shareholder is in a murkier position. Buying into xAI or X was always just a bet on Elon's sphere of influence, not on discrete assets. The merger makes that bet explicit. Whether that's reassuring depends on how you feel about the sphere.

The mid-market AI valuation problem

Lessin's deeper argument isn't about the merger itself — it's about the category. Mid-market AI players are, in his view, overvalued because people are confusing a technology layer for an economy. Google and Meta aren't technology companies; they're economies with network effects that compound. LLMs are still just a technical layer. The companies claiming LLMs will become compounding economies "actually have nothing right now" — they have fun, largely undifferentiated technology.

The bull case for Grok is that it displaces Google and carves out a slice of a multi-trillion-dollar market. Lessin thinks that framing is wrong. The accountant analogy is his clearest expression of it: if you pay an accountant $120,000 a year and an AI accountant costs $10 a year, the AI accountant isn't worth 120x more — it's worth $10. Value collapses to the marginal price, not the displacement value.

He's also skeptical of the X-as-data-flywheel argument. The vision — X users pour their knowledge into the platform, creator payouts incentivise content, that data trains Grok, Grok gets smarter — is coherent on paper. But search and social both promised information economies and delivered entertainment economies instead. Google became a commerce engine and a farm for SEO spam. Social became a dopamine feed. Lessin sees no evidence the X-xAI combination escapes that gravity.

IPO valuation

Asked where xAI would trade in a public offering tomorrow, Lessin offers two scenarios with a wide spread between them. If it trades like a memecoin — the way GME did — the number could be very large. If it trades on discounted cash flow as a conventional business, it trades poorly. The outcome depends entirely on who the marginal buyer is and what that buyer believes, which is as much a cultural bet as a financial one.