US IPO proceeds hit $115.6 billion in the first half of 2026 and the market is about to broaden

The concentration trap

Two issuers. Roughly $100 billion. That's the actual headline. SpaceX priced June 12 under SPCX at a $1.77T valuation (Business Insider), instantly the largest IPO on record. SK Hynix followed with ADRs at $149, opened at $170 on Nasdaq, and delivered one of the biggest US share sales by a foreign issuer at roughly $26.5B (MarketWatch). Both rode the AI infrastructure bid — Hynix supplies high-bandwidth memory into Nvidia's supply chain, which is why the deal priced at all. A market that needs two names to clear $100B isn't broad. It's a window propped open by AI capex, and the moment that thesis rolls over, so does the IPO tape.

The pipeline is a wish list, not a calendar

EY's Americas IPO leader Rachel Gerring is constructive on H2 — conditional on filings becoming listings. The names circulating:

• Kraken, Blockchain.com, ConsenSys, Dataiku — crypto infrastructure and AI tooling. Each carries baggage: custody questions, regulatory overhang, or — in ConsenSys's case — a token float that complicates clean equity pricing.
• Anthropic — AP reported a $65B raise at a $965B valuation in May; secondary trades have pushed implied marks even higher. A listing at that level is a public-market referendum on whether AI-native multiples can hold without revenue validation.
• OpenAI — confidential filing chatter persists, but reporting still points more naturally to 2027 than a rushed 2026 debut.
• Anduril (last valued near $14B) and Sierra Space — defense tech as a separate trade from AI, riding NATO budget pressure and Germany's expanded infrastructure commitments.

Advanced manufacturing, energy and automation software round out EY's expected broadening. None of it is filed. None of it is priced. A confidential filing is not a listing. A banker mandate is not an underwriter.

What this actually means for LPs

The "broadening" trade is still a thesis, not a print. Three things to track:

1. Concentration risk is mispriced. Anyone benchmarked to public venture returns is effectively long SpaceX and SK Hynix. If either deal wobbles post-lockup — or if Nvidia's capex cycle cools — the recovery narrative unravels fast.

2. Liquidity is fragmenting. While the Street waits for Kraken and ConsenSys to file, the secondary float is moving elsewhere. Solana rails have already consolidated most of the on-chain tokenized equity volume, giving LPs an exit path that bypasses the traditional IPO tape entirely.

3. Lockup math matters. Two mega-deals priced in Q2 means a wall of supply hits in Q4 and Q1 2027. Watch the post-180-day action on SPCX and Hynix ADRs — that's the real read on whether demand was genuine or just AI-FOMO.

The honest read: 2021 taught public buyers that pricing late-stage private marks at IPO is a one-way transfer of wealth. The 2026 cohort hasn't changed the math. It just found two bigger checks to mask it.