Two clocks hit zero this week — SpaceX's quiet period lifts, and HK$50 billion in Chinese AI lock-ups expire. One lets analysts start printing numbers. The other lets insiders start printing exits. If you're just tuning in: SpaceX's first weeks as a public company haven't really been about a first-day pop. They've been about repricing mechanics. The record June listing raised $85.7 billion, low float and index inclusion have shaped the trading, and the debate's moved from valuation into operating quality — Starlink adding subscribers while average revenue per user slips, and the broader AI-driven losses are still very much in the frame. This is AI IPO Watch. Today, the first bank-stamped price target lands, insiders in two markets get their exit windows, and we finally get to argue over real numbers instead of whispers. Cassidy, that $401 — let's start there. Here's Space & Defense Intelligence Brief:
The mandatory quiet period for the 23 underwriters ends on Tuesday, July 7, 2026, freeing them to publish ratings, research, and price targets. Arete Research analyst Andrew Beale, not constrained by the quiet period, issued a buy rating and a Street-high price target of $401, implying 154% upside from Wednesday’s close.
Here's where the SpaceX repricing story stands — the underwriter quiet period expires tomorrow, July 7, and the first bank-stamped targets are about to land. But the number already in print isn't from the syndicate. It's from Arete Research's Andrew Beale, who was never under the gag order in the first place. $401. Street-high, 154% upside. So the loudest number came from the one analyst who didn't need permission to say it. That tells you something about who's positioning ahead of the crowd. And that's the tell I care about. $401 is a real, attributed, name-on-it number — finally. But it's from outside the 23-bank syndicate. When the paid banks publish tomorrow, watch whether anyone matches Beale, or whether they all quietly cluster lower. The part nobody's saying out loud: the stock closed around $161 on day one, printed at $135, popped 50%, and now trades below that first-day close. Most of the aftermarket is underwater right now — and that's the tape a $401 target is walking into. This one's from The Standard:
Shares of Chinese AI startups Z.ai (2513) and MiniMax (0100) face near-term pressure as cornerstone investor lockups worth over HK$50 billion expire this week, which could prompt a rotation back into traditional tech, analysts say. Z.ai, also known as Zhipu AI, will unlock 25.68 million shares for cornerstone investors on Wednesday.
Here's the mirror image of the SpaceX story we just hit. Instead of index funds forced to buy, you've got Z.ai's cornerstone holders — 25.68 million shares, HK$46 billion — finally able to sell on Wednesday. Same clock, opposite direction. And this is the scarcity premium getting tested in real time. Z.ai hit a HK$1 trillion valuation last month, and it's already down more than twenty percent this week — before the lock-up even lifts. The market's front-running the exit. Up 14.4 times from a January IPO. Six months, and the earliest, most committed holders get their window. If you were a cornerstone investor sitting on that, you're not holding for the vision. What I like about this one — it's Hong Kong, not New York. There's no Nasdaq-100 passive bid catching the falling knife. When HK$50 billion of supply hits, it hits a tape with no forced buyer on the other side. And MiniMax unlocks HK$5.7 billion a day later, Thursday. Two names, back to back, into a market analysts are already saying will rotate back into old-school tech. That sets up more like a hard landing. The cornerstone allocation was the whole scarcity story on the way in — locked-up, committed, believers. Turns out belief has an expiration date, and it's this Wednesday. From Sean Williams at The Motley Fool:
SpaceX's historically low float(i.e., tradable shares), coupled with forced buying by index funds -- SpaceX was or will be added to the Russell 1000, Russell 3000, and Nasdaq-100 -- can artificially boost its share price. But this tailwind for SpaceX stock has a rapidly approaching end date.
So here's the number that actually matters in the Fool piece — SpaceX sold 555 million shares, which sounds huge, but it's under five percent of the company. Companies usually float ten to twenty-five percent. Which is the whole trick. You raise a record $85.7 billion — nearly triple the old high, per the Fool — off a sliver of the cap table, and the rest is sitting behind a lockup with a release date attached. And that date is disclosed. The headline says insiders “dump on retail” like it's some ambush — the schedule's in the S-1. Retail just doesn't open the S-1. No one hid the calendar. The people forced to hold this through Nasdaq-100 inclusion just can't read the release calendar the insiders wrote for themselves. That's the pairing that gets me. Index funds get forced in during the same window early holders get their exit. Forced buyers meeting willing sellers — guess who set the timing. Here's Angela Tan at The Straits Times:
The rocket company has all the hallmarks of a market phenomenon: a charismatic founder, a compelling vision of humanity’s next frontier, and a valuation that defies both gravity and conventional financial modelling. As rumour mills spun private market valuations ever higher, and implied valuations climbed to the stratosphere, the pressure to participate grew harder to ignore.
This line gets me: “The problem was not that SpaceX could not become dominant, but that the price already assumed it would.” That's the whole thing. That's every AI IPO I've watched founders prep for. And Angela Tan at the Straits Times passed. She felt foolish for about a week — then the stock dropped 20% off the peak, roughly $600 billion of market value gone. Discipline looked dumb right up until it didn't. What I like is she named the actual mechanism, not the vibe. She framed the bet around price: does the number already assume flawless execution? At $135 to a $401 target, you're pricing zero regulatory friction, zero competition, zero capital calls. It's the same math the quiet-period note we just hit is quietly leaning on. Somebody has to eat that $4.9 billion loss trajectory, and “assume dominance” is how you make it disappear. Right, and Tan's writing a regret column in public. The institutions who passed don't publish. So retail reads the $401 and reads her feeling foolish — and never sees the part where the 20% drop proved her right. The honest version of the piece: she didn't lose money by not owning it. Eric, an investment committee may hate the optics — but it can defend the arithmetic. If AI IPO Watch helps you stay ahead of the market, take a moment to subscribe and leave a review wherever you listen. It really helps other people find the show.
Looking ahead: SpaceX's IPO quiet period ends Tuesday, July 7, so its 23 underwriters can start publishing ratings, research, and price targets. On Wednesday, Z.ai's cornerstone investors unlock 25.68 million shares; MiniMax follows Thursday with a 16.5 million-share unlock. And SpaceX's first quarterly operating results are currently estimated for August 6.
You'll find links to every story we covered in the show notes — worth opening the ones you want to dig into. That's AI IPO Watch for today. This is a Lantern Podcast.