One month after the biggest IPO in history, SpaceX is fading — and George Noble just put a 900% number on the reason why. If you're just joining: SK Hynix came to the Nasdaq as a Korean memory giant, listing ADRs under SKHY to fund AI-chip capacity. The open question was whether around $149 a share and a $26.5 billion raise would pull in real U.S. demand against Micron and the rest of the AI-hardware book. Today, we have an answer. This is AI IPO Watch. Today — a float bomb, sixty lawyers on a plane, and one deal surging while another sputters. Let's start with SpaceX, one month on. Here's AOL:
Although the company had decided to price its shares at $135 each, the price immediately shot up to $150 that first day, climbing to $176, before closing at $160.95. It solidified SpaceX as the largest initial public offering (IPO) of all time.
One month later, AOL's asking if it lost momentum. The answer's on the tape: priced at $135, closed day one at $160.95, ran to $225 intraday, and now the mood's flipped from frenzy to concern. And listen to what analysts say actually carried it. Keith Snyder at CFRA — quote, first time people felt they could invest in something marketed as an AI play. Marketed. That's the word. Right. Willy Lee says everyone saw it as an AI story. The rocket company. The AI story rests on xAI — now SpaceXAI — and Grok. Retail didn't buy propulsion; they bought a chatbot narrative bolted onto a launch business. The Nasdaq-100 passive buying was supposed to cushion this — JPMorgan estimated billions would flow in on index inclusion. One month in, that cushion clearly hasn't held. The stock's giving it back. Benzinga has the details on this one. George Noble put a number on what everyone's been dancing around — a 900% float explosion. That's a supply claim, not a price call, so my first move is simple: where does it come from? Either you can back it out of the lock-up schedule and the S-1/A number two float math, or you can't. That's the peg I've wanted all week. With the float going up nine-fold, retail suddenly learns what 'float' even means. The shares you could actually buy at debut were a sliver. Right, and one month on — we heard it in the AOL piece — the momentum's already fading. Noble's warning has moved past someday. The first real unlock window is a live event now. And here's the ugly geometry: on the other side of a 900% float increase is the retail buyer who came in at $152. Sophisticated backers up 1,500 percent are handing the supply to exactly the people least likely to price it in. So I want the mechanism named on air. If the number's right, insider shares that were locked at debut convert into free-trading supply — and that supply's been sitting in the filing the whole time, just not on the cover page anyone read. 900% is a scary headline. But strip out the shock value and it's just the unlock calendar doing exactly what unlock calendars do. The only surprise is that anyone's surprised. Here's Bloomberg Law:
Running the legal side of the largest IPO in history meant 8 a.m. huddle calls seven days a week for a Gibson Dunn & Crutcher team spanning six regions globally and more than 60 lawyers. Hillary Holmes, who led the team, said in an interview that the synchronized, multi-jurisdiction retail restructure for the $75 billion SpaceX IPO might be of use to other companies seeking to go public.
Sixty lawyers, six regions, 8 a.m. calls seven days a week. Everybody keeps calling the roadshow the hard part — this is the hard part. And listen to what Hillary Holmes says it was for: a synchronized, multi-jurisdiction retail restructure. The machine was built to let foreign retail investors buy in on day one, instead of waiting for the secondary market. Right, and that's the structural first Bloomberg Law is flagging. Historically, an IPO marketed abroad leaned on registration exemptions, so it only reached institutions and high-net-worth buyers. Nasdaq and Deutsche Börse trading in one unified transaction has basically never been done. So here's what nags at me. That legal architecture tells you counsel expected regulatory friction in multiple markets. The single U.S. filing I've been reading doesn't show me the jurisdiction-specific disclosure — so the document I trust is only one slice of a much bigger structure. And look who they engineered all of this for. Sixty lawyers, regulator by regulator, using the UK's brand-new Public Offer Platform — institutions could already get in. This was built to get retail through the door worldwide, on day one. That lands differently a month later, when the momentum piece we just hit says the thing's fading. You opened the front door to retail globally, right at the top. From Tom Lauricella at Morningstar:
US shares of SK Hynix SKHYV jumped after the stock’s IPO on Friday, as the South Korean memory semiconductor producer raised $26.5 billion, despite its home-listed stock having fallen 25% over the past three weeks. SK Hynix priced its American depositary shares at $149. The newly minted stock ended the day Friday trading at $168 per share, a roughly 12.8% increase from the IPO price.
So the SK Hynix ADR story lands cleanly — priced at $149, finished at $168. That's your 12.8% pop, against a home-listed stock that had dropped 25% in three weeks. And here's the number that actually reassures me — Einhorn at Renaissance says they priced at a 3% premium to the as-converted Korea close. It's a premium you can tie back to a real reference price, not some fantasy multiple. And that's why the deal worked, right? They funded their capex, gave U.S. buyers a modest return, and didn't blow up the Korea listing. That's a book that behaved. $26.5 billion — second-largest U.S. IPO on record — and it clears with a 13% pop that nobody's writing a bombshell headline about. Compare that with the raise a month ago that everyone's still autopsying. A 12.8% pop off a $149 price is discipline. My question is still whether that book is long-only, or whether it's fast money that surges and then leaves in September. A 13% pop is exactly the range where sticky demand shows up without leaving money on the table. From Prakhar Agarwal at Seeking Alpha:
SpaceX trades at a market capitalization close to ~$2 trillion and at over ~100× sales, a valuation that remains outside any analytical scope. Only the Starlink segment currently remains profitable, while the other business lines that underpin much of the valuation and TAM narrative remain loss-making. SPCX is rated Strong Sell due to the extreme disconnect between its ~$2 trillion valuation and its underlying fundamentals and business economics.
Prakhar Agarwal puts a hard number on what I've been circling all week — roughly 100 times sales, and his phrase is 'outside any analytical scope.' That's stronger than calling the stock expensive. From a CFA charter holder, it means the model basically doesn't exist. And here's the part that connects to the Starshield problem: only Starlink is profitable. Everything else propping up that TAM slide is loss-making. The invisible segments I flagged in the S-1 are now invisible in the analyst coverage too. He also names the mechanism: less than 5% of shares are publicly tradable. So the roughly $2 trillion market cap is being set by a sliver of float — exactly the supply story the Benzinga piece we just hit was screaming about. Right, and stack that against SK Hynix today — $26.5 billion raised, actual memory revenue, a real book. One of these deals is being read on fundamentals. The other is being read on scarcity. That jump is what gets me — about $350 billion private to roughly $1.77 trillion at IPO in under two years, with a roughly $1.25 trillion merger dropped in as the anchor. Those steps look like negotiating references, not valuation work. And the person on the other side of that anchor is whoever bought at $152. Agarwal rates it Strong Sell. Those staggered lock-up expiries — what he calls a 'material overhang' — are the float explosion with a calendar attached. If AI IPO Watch helps you keep up with the market, take a second to subscribe or leave a review wherever you’re listening. It’s a quick way to support the show and help other people find it.
Links to every story we covered today are in the show notes, if you want to dig further into any of the filings, funding moves, or market signals that caught your ear.
That’s AI IPO Watch for today. This is a Lantern Podcast.