The IPO Market That Only Opens for Trillion-Dollar Companies

Primary Markets Group counts only initial public offerings. SpaceX raised $75 billion in June — selling 555.6 million shares at $135 each in the largest IPO in history. The rest of June's 19 deals brought in roughly $23 billion. In July, seven smaller IPOs raised $3.0 billion.

That's a 97% month-over-month drop in proceeds. The official headline from Primary Markets Group — S&P Global's tracker for new equity issuance — is that the market "marked a significant slowdown."

But here's the thing: the slowdown isn't the story. The concentration is. The July number is barely interesting on its own. What's worth examining is what sort of market raises nearly $100 billion in one month and $3 billion the next, and calls both of them signs of recovery.

This is basically a bond-issuance calendar wearing IPO clothing.

When the bond market is active, you don't say "rates are recovering" because a $30 billion Treasury auction landed on a Tuesday. You say the calendar was lumpy and the big maturities clustered. The IPO market works the same way, but nobody talks about it that way because IPOs carry narrative weight that bond auctions don't. A mega-IPO isn't just capital formation — it's proof of market health, a barometer of investor confidence, a thing banks write about in their quarterly marketing materials. So when SpaceX prints in June and nothing else of that size follows in July, the data looks like a crash instead of what it is: a month that happened to fall between fireworks.

Let's check the plumbing. The first half of 2026 saw roughly $114 billion raised in traditional U.S. IPOs, compared to $14.8 billion in the same period last year. That's a sevenfold increase. PwC, which tracks this independently, notes that even excluding SpaceX entirely, H1 2026 capital raised was nearly three times H1 2025. So the window is open. It's just that the window is currently shaped like SpaceX.

The aftermarket mechanics reinforce the concentration story. Nearly 97% of H1 2026 IPOs opened above their offer price on the first day. Almost half priced at or above the top of their marketed range. That's not a market functioning normally — that's a market where institutional demand is chasing a very short list of acceptable names, and the underwriting system has the leverage to set take-it-or-leave-it prices because the queue is long enough. SpaceX didn't negotiate its $135 offer price the way issuers traditionally do. It set it. The roadshow was a presentation, not a negotiation. That only works when the buyers are the ones who feel lucky to be in the book.

Then there's the classification boundary, which makes the July data even more misleading than the raw numbers suggest. SK Hynix, the South Korean chipmaker, raised $26.5 billion in July by listing American depositary shares on the Nasdaq. That's the second-largest U.S. equity offering in history, behind only SpaceX. But it's not an IPO. SK Hynix has been publicly traded on the Korea Exchange for decades. What happened in July was an uplisting — a seasoned issuer creating a U.S. trading venue and raising fresh capital from American investors. Primary Markets Group excluded it from the IPO count. The WSJ called it a "share sale". The market treated it like a mega-IPO, with the same kind of institutional demand and price discipline.

The distinction matters because it reveals how much capital is moving through the primary market in July that isn't counted as an IPO. If you include SK Hynix, July's capital-raise number jumps from $3.0 billion to roughly $29.5 billion — still less than June, but not the collapse the headline implies. The real question is whether the pipeline for actual new issuers is open or whether it's a door that only swings for companies that have been building toward this moment for five years.

The evidence points to the former, with caveats. Jersey Mike's Subs, Bending Spoons, Reformation, Scribe Therapeutics, IMC Rare Earths — these July issuers aren't trillion-dollar plays, but they priced in a market where the afterlife of going public hasn't been punished. Cerebras Systems, the AI infrastructure company, priced more than 50% above its initial filing range in earlier months and opened 89% above offer. Innio and Madison Air, in industrial tech and data-center cooling respectively, both appreciated roughly 45% since pricing. The aftermarket is working, which means the feedback loop that kills IPO windows — weak debits discouraging the next queue — isn't operating right now.

But the market that's working is a specific market. Venture capital in Q2 2026 funneled roughly 95% of its funding into AI-focused companies. Eight financings exceeded $1 billion, all but one in AI. The IPO pipeline mirrors this concentration: the companies going public right now are either directly in AI infrastructure, adjacent to it (space, data centers, chip manufacturing), or are consumer and biotech plays that have survived long enough in private markets to arrive with credible scale. Traditional software-as-a-service, healthcare services, and consumer discretionary IPOs are thin. The VIX is hovering around 16, well below the 20 that usually signals a "choppy" window, but the calm market isn't making everything tradeable — it's making the already-hot sectors more concentrated.

Here's the incentive structure that makes this arrangement stable for now. Underwriters want mega-deals because the fees are enormous and the reputational payoff is asymmetric — one successful SpaceX-size listing is worth more marketing than a dozen mid-cap debuts. Institutional investors want mega-deals because they have the mandate flexibility to buy them and the index-providers have already created fast-track mechanisms (Russell, Nasdaq, MSCI, S&P Total Market) to pull large-cap IPOs into benchmarks within days rather than months. The companies that can go public right now are the ones that could have stayed private — they have venture backing, sovereign wealth interest, or operational cash flow sufficient to not need the IPO as a lifeline. They're using it as a liquidity event for insiders and a valuation reference point, not as a capital necessity.

That means the "recovery" is real but narrow. It's a recovery for companies that don't need recovering. The mid-cap window — General Atlantic, the private equity firm, expects deals in the $750 million to $1 billion+ range to lead the next phase of broadening — is still waiting. The primary risk to that expansion is a high-profile aftermarket disappointment. One major IPO that tanks after pricing could shorten the window fast.

SpaceX itself is now a case study in why short-term price action doesn't tell you whether the structural story holds. SPCX is trading at $133.11, up 15.8% on the day with $30.6 billion in turnover, but down 8.4% over the past 20 days and flat year-to-date. It peaked at $225.64 earlier in its trading life and hit a 52-week low of $104.83. The stock has been volatile, which is expected for a freshly public company with a restricted float and dual-class governance (Musk retains 85.1% of voting power). The aftermarket hasn't confirmed the valuation, but it hasn't destroyed it either. For the IPO machinery, that's acceptable. The market doesn't need every IPO to moon; it needs them not to crater, so the next issuer doesn't get cold feet.

The pipeline for H2 2026 includes OpenAI and Anthropic, both of which confidentially filed for public offerings during Q2. Those, if they land, would extend the mega-deal pattern into the second half. But they would also reinforce the same structural point: the IPO market in 2026 is not broad, it's tall. A few enormous companies are going public, the machinery is working for them, and the monthly data looks wildly volatile depending on which of them happens to price in any given window.

The simplest model is this: if you're counting IPO market health by total proceeds, the number that matters is whether the mega-deal calendar has a hole in it, not whether investor appetite has changed. If you're counting by whether a $300 million software company in Boston can actually price its offering without begging the market for attention, the answer is probably yes — but only if the mega-deals keep performing well enough to keep the aftermarket optimistic. The smaller IPOs are riding the coattails of the bigger ones. That's not a broken market. It's just a market where the plumbing currently runs through a very narrow pipe, and everyone in the system knows it.