Finn Brooks: Clara, catch — did you see the Momenta close yesterday?
Clara Bennett: I did. HK$299, give or take.
Finn Brooks: Which is — wait, the offer was HK$295.60. So after 414x oversubscription, after HK$5.89 billion raised, after BlackRock and Fidelity International and Mercedes-Benz all locked in as cornerstone investors — the stock goes up about one percent and calls it a day. I've been trying to figure out if that's a scandal or if I'm just bad at math.
Clara Bennett: You're not bad at math. The math is the problem.
Finn Brooks: And it's not just Momenta — SpaceX gets added to the Nasdaq 100, which is supposed to be the ultimate institutional stamp of legitimacy, and it drops 6.8% to $149.47 on day one of the index. It was at $201.80 in June. That's more than 25% gone while Raymond James is sitting there with an $800 price target on the thing.
Clara Bennett: So the question for today — and I think it's a genuinely uncomfortable one — is whether the machinery we use to signal IPO confidence is still actually connected to price outcomes.
Finn Brooks: Right, like, does 414x oversubscription mean anything? Does cornerstone backing mean anything? Because July 8th, 2026 is a pretty good case study for 'apparently not.'
Clara Bennett: In practice, I think it means something — just not what most investors think it means. That's the distinction I want to get at.
Finn Brooks: Okay but that distinction — what does it actually mean in practice? Because from where I'm sitting, 414 times oversubscribed sounds like the loudest possible signal.
Clara Bennett: Think of it like a restaurant reservation list. A hundred people saying they want a table tells you the place sounds good. It doesn't mean they'll all show up, and it definitely doesn't mean they'll order the expensive bottle. Oversubscription measures book-building demand — people raising their hand during a structured process. That hand goes down the moment secondary trading opens.
Finn Brooks: Wait — so the 414x number is measuring appetite to participate, not conviction to hold.
Clara Bennett: Exactly. And the cornerstone piece is actually more interesting, because — now, people treat cornerstone investors like a launch pad. BlackRock, GIC, Fidelity International locking in at HK$295.60 sounds like 'these serious institutions think this goes up.' But that's not what the mechanism does. They committed to a fixed allocation at the offer price before trading began. They're a floor, not a catalyst. They're not buying more in the open market on day one.
Finn Brooks: So they're sitting there holding their allocation and the secondary market just... prices it independently.
Clara Bennett: Yes. And the Mercedes-Benz piece — that one I find genuinely clarifying. A legacy automaker commits cornerstone capital to Momenta, a Chinese autonomous driving firm. That's not an index fund chasing AI momentum. That's a manufacturer hedging its own autonomy risk. It's insurance. So when people read that list and think 'smart institutional conviction,' they're — mm — they're collapsing two very different motivations into one signal.
Finn Brooks: That is actually a wild reframe. Mercedes-Benz isn't saying 'Momenta will pop,' it's saying 'we cannot afford to be wrong about autonomy.'
Clara Bennett: Right. And there's a structural layer Momenta walked straight into — Hong Kong in mid-2026 had a record wave of lock-up expirations from a strong first half. That's existing holders getting their first legal exit window, all at once, right as a new listing lands. The downward pressure is baked into the calendar, not the company. Momenta allocated roughly 60% of proceeds to R&D and AI computing infrastructure — that's a capital-intensive bet. The market priced that at nearly zero premium on debut day.
Finn Brooks: So the whole system worked perfectly — it just doesn't do what everyone thinks it does.
Clara Bennett: Which brings us to SpaceX — and I think that's where the signal failure becomes, honestly, indefensible.
Finn Brooks: Yes. Okay. Because this is my partial win and I want to claim it properly.
Clara Bennett: Go ahead.
Finn Brooks: So Momenta — you can almost argue the mechanics held. Flat debut, institutional anchors did their floor job, whatever. But SpaceX gets added to the Nasdaq 100, which is supposed to funnel passive inflows automatically, and on that exact day it drops 6.8% to $149.47. It was at $201.80 in June. That's — I mean, picture a Vanguard portfolio manager watching her dashboard on July 7th. She's buying SpaceX because the index forces her to. She's not making a conviction call. And simultaneously Raymond James is publishing an $800 target implying a $10.5 trillion valuation, mostly off AI-compute forecasts. Those two things cannot — wait, actually they literally cannot coexist without one of them being catastrophically wrong.
Clara Bennett: And Jeremy Grantham answered that question pretty directly. July 7th, Morningstar interview — 90% probability SpaceX stock eventually crashes. His specific word was 'speculative infrastructure assumptions.' He's not saying bad company, he's saying the valuation math requires things that aren't built yet.
Finn Brooks: Ninety percent. From Grantham. While Raymond James has eight hundred dollars on the board.
Clara Bennett: And in July 2026 the stock is trading at $149 — closer to its $135 IPO price than to either analyst's ceiling. So the market, at that moment, is voting much closer to Grantham than to Raymond James. That's the kernel of truth in the hot take — index inclusion didn't cushion anything. It coincided with the drop.
Finn Brooks: Which makes the pipeline question so strange to me — because Samsung-backed Rebellions is still pushing toward a Korea Stock Exchange listing this year, NSE just got SEBI approval for a September Rs 30,000-crore IPO. Nobody's blinking. And I think the part that makes this worse — we'll get into it — is that 'AI optimism' as a macro label is covering a lot of company-level doubt about whether AI revenues actually materialize at the scale any of these valuations need.
Clara Bennett: That's the right thread to pull. The deal volume is real. The conviction underneath it — that's a separate question entirely.
Finn Brooks: And that's the thing that keeps snagging me — ION Analytics is literally calling this market 'soaring on AI optimism' on July 7th, the same week Momenta opens at HK$301 and SpaceX is down 25% from peak. Like, those two descriptions cannot both be precisely true at the same time.
Clara Bennett: They can, actually — and that's the precise distinction worth landing. Deal volume is soaring. Price conviction is not. Those are measuring different things.
Finn Brooks: Wait, so the market is open for business but nobody believes in the stock?
Clara Bennett: Company by company, yes. The AI label used to function like a sector lift — attach it to a filing and valuations moved together. Now the market is — I mean, it's underwriting each company's specific path to AI revenue. Momenta's allocating 60% of proceeds to AI computing infrastructure and got zero debut premium. Raymond James' $10.5 trillion SpaceX target is built on AI-compute forecasts that don't exist yet in the revenue line. The market looked at both and said: not yet.
Finn Brooks: And yet — IAI and Elta just got Israeli government approval to merge ahead of a dual IPO. Valued at 100 billion shekels. Two-thousand-twenty-five sales were $7.8 billion. That's a real revenue base. Does that actually land differently?
Clara Bennett: That's the calibrated version of the claim — IAI has $7.8 billion in actual sales. Rebellions has Samsung backing but no public revenue. NSE's Rs 30,000-crore September offering is an exchange, a utility, something tangible. The pipeline isn't homogeneous. The companies with near-term fundamentals are probably fine. The ones pricing on AI futures — that's where the credibility gap lives.
Finn Brooks: So the defensible claim is: AI optimism is still driving deal flow, but the secondary market has started doing company-level due diligence that the primary market — the oversubscription, the cornerstone anchors — doesn't actually do.
Clara Bennett: That's it exactly. And the portfolio manager watching Momenta sit at HK$299 — she got sold a primary-market signal as if it were a secondary-market guarantee. Those are not the same instrument.
Finn Brooks: Okay, that's the one. Primary signal, secondary reckoning. And the pipeline — Rebellions, NSE, IAI — is about to find out which side of that line they're on.
Clara Bennett: NSE in September, Rebellions on the Korea Stock Exchange — those two are the actual verdict. Not a prediction, not a read on sentiment. Post-debut numbers. That's when we find out if Momenta was company-specific or the opening signal of something larger.
Finn Brooks: Yeah. And I'll half-concede the thing I've been resisting — maybe the signals aren't broken, they're just being read by people who forgot what they actually measure. Like, 414x oversubscription was never a price target. I think I knew that and kept arguing past it anyway.
Clara Bennett: That's a fair landing. When Rebellions and NSE report their first few days of trading, we'll know whether July 2026 was a valuation correction on a handful of companies — or the moment AI-labeled IPO appetite fundamentally shifted. One of those is a footnote. The other is a chapter.
Finn Brooks: And we'll know by October. No guessing required.
Clara Bennett: Good conversation. Genuinely.
