Investigation — AI mega-IPO Fast Entry lane desk pass — methodology gate update and pre-event tape classification (2026-06-02)

Question: AI mega-IPO Fast Entry lane desk pass — methodology gate update and pre-event tape classification (2026-06-02) Verdict: s-p-decision-pending-no-change-posted; nasdaq-t15-live; crsp-t5-new-earlier-gate; spacex-offering-blank-lockup-staggered; tape-does-not-confirm-liquidity-drain; early-selective-halo; passive-flow-untested

What we're asking

First substantive desk pass on the parent lane (ai-mega-ipo-fast-entry-liquidity-regime) since it was created 2026-06-02. Three jobs, all producer-only, no child-track mutation:

1. Methodology gates — has the S&P DJI MegaCap decision posted? Is Nasdaq Fast Entry still as modeled? Any new dated gate?
2. Float / lockup mechanics — what does the filed SpaceX S-1/A actually say (vs press)?
3. Pre-event tape — classify the declared baskets at the last completed session against the lane's hypothesis set, ~7 trading days before the reported 2026-06-12 SpaceX listing.

Posture unchanged: this is not a recommendation to buy the IPO. We are testing which existing positions/watchlists become vulnerable or advantaged if passive/index plumbing changes.

What we found

1. Methodology gate update (primary-source)

S&P DJI MegaCap consultation — status: no-final-decision-posted-yet. As of 2026-06-02 no adoption/results announcement is on press.spglobal.com or the mr4292 consultation page. Fortune's 2026-06-02 piece still uses conditional language ("if the changes are adopted, they take effect before the market opens on June 8"), confirming the decision was still pending day-of. The 2026-06-08 effective date is therefore proposed/conditional, not live. Consultation contents re-confirmed: 12mo → 6mo IPO seasoning for S&P Composite 1500; waive 0.10 IWF float + financial-viability/GAAP-profitability for MegaCaps; fast-track MegaCap IPO adds to S&P TMI / Completion / DJ U.S. TSM only — NOT the S&P 500. S&P 500 inclusion remains Index Committee discretionary (verbatim-supported).

Nasdaq-100 Fast Entry — LIVE since 2026-05-01, mechanics unchanged. Primary FAQ (extracted verbatim): top-40 rank + eligibility → added after 15 trading days, subject to the 3× float cap (cap applies up to 33.3% float; above that, full listed cap), 10%-float minimum removed, $5M ADTV-from-listing still required, cohort "typically >$100B full market cap as of March 2026." First quarterly rebalance under new rules: 2026-06-22.

NEW gate — CRSP 5-trading-day fast-track (was not in the thread). The secondary coverage (Moneywise/Yahoo, Motley Fool) surfaces that CRSP — the index behind Vanguard's VTI (~$607B AUM) — fast-tracks large IPOs after ~5 trading days, materially earlier than Nasdaq's T+15. This is a distinct, earlier passive-flow gate the lane should model separately. Reconcile the T+5 (CRSP/VTI) vs T+15 (Nasdaq-100) conventions explicitly; one source also cited a Nasdaq "seventh-trading-day" figure (likely conflation) — treat the primary FAQ's T+15 as authoritative for the Nasdaq-100 leg.

The only fast mega-IPO index path that is actually LIVE today is the Nasdaq-100 Fast Entry (T+15) plus the CRSP/VTI total-market fast-track (~T+5). The S&P side is (a) not yet decided and (b) even if adopted, S&P 500 inclusion is gated by 6-month seasoning + committee discretion.

2. Float / lockup mechanics (primary — SpaceX S-1/A No. 1, filed 2026-06-01)

The filed registration statement corrects several press "facts":

• Offering size, price range, and float % are BLANK (placeholder fields). The ~$1.75T valuation / ~$75B raise / 3–4% float are press numbers, not filed facts. Every forced-flow dollar estimate downstream is provisional until the deal prices.
• Lockup is genuinely STAGGERED — base 180 days, with earnings-gated + price-trigger early releases for non-founder holders: ~20% after Q2-2026 earnings; +10% if SPCX ≥30% above IPO on ≥5 of 10 days into that print; +7% laddered at T+70/90/105/120/135; +28% after Q3-2026 earnings; remainder at 180d. Musk: 366-day lockup, no early release (~85.1% voting; controlled company). This means meaningful supply can hit as early as the Q2 print (~late July/Aug), well before the nominal 180-day expiry — the crux of the T+30/lockup "does supply overwhelm passive demand?" gate.
• "30% retail carve-out" is press characterization, not in the S-1 (only retail selling-group brokers named). 5-for-1 split effective 2026-05-04; xAI merger closed 2026-02-02 (common-control recast).

3. Allocation-crowding evidence (two falsifiable poles now in the wild)

• Forced-buying / liquidity-drain pole: Dunn/AlgorithmicFIRE quant model (~$49B SpaceX-alone, ~$97B all-three, 5× inelastic-markets multiplier → ~$485B displacement); SpotGamma per-index plumbing ($8–12B→$200B S&P, $7B+ QQQ); moomoo "Macro Moover" relocates the drain to the book-build/funding- pocket window (NVDA/AAPL/MSFT/AMZN sold to raise allocation cash) and notes the macro buffer is gone (reserves −$560B, RRP drained).
• Nothingburger / float-math pole: Friedman (InvestmentNews) — 3–4% float → float-adjusted S&P weight only ~0.08–0.12%; Tesla is ~17× SpaceX's expected weight; ~$80–120 forced buying per $100k in S&P funds.
• Number hygiene: the "$950B JPMorgan" figure is fragile (cited as $95B vs $950B across outlets; Dunn's independent model lands ~$49B/$97B). Do not publish a forced-selling number without the primary JPM note.

4. IPO clustering (regime evidence, child cases stay separate)

Three filings now stack inside ~2 weeks: SpaceX public S-1/A (2026-06-01), OpenAI confidential draft (~2026-05-22, secondary), Anthropic confidential draft S-1 (2026-06-01, primary company statement). Confidential drafts are not public S-1s, so the parent lane's graduation criterion #3 has not fired — but the clustering is direct evidence toward "the market treats the IPOs as one allocation regime." Child tracks remain separate; this is cross-link only.

5. Pre-event tape classification (2026-06-01 close — last completed session)

Today's 2026-06-02 intraday is not fetched (no broad scan in a background pass); the 2026-06-01 close is the last completed session, ~T-7 to T-9 before the reported June 12 listing. Required basket read (precompute, code-computed):

Cross-check: 24/7 Wall St (Moadel, 2026-05-28) independently reported the selective halo — LUNR +9%, RKLB −3%, PL −2% — "positioning, not a fundamental shift," consistent with the 7d divergence above.

Verdict + reasoning

Scan signature for this pass: unrelated-or-insufficient on the broad liquidity-drain leg, with early confirms-halo on the 30d/3m space horizon (selective, cooling at the 7d edge). Passive-flow and comp-repricing legs are untestable pre-listing.

Reasoning:

• Liquidity drain — not confirmed (leans contradict). The vulnerable basket (NVDA/Mag7, neoclouds) is among the strongest baskets at the 2026-06-01 close; SPY/QQQ sit at 52wk highs with hot RSI. The "crowded winners sold to fund allocations" signature is absent at the index level. The moomoo book- build thesis predicts this drain would show up in the T-7→T-1 window — so the cleanest test is still ahead, not behind.
• Halo — early confirm, selective. Space basket broadly up 50–140% over 30d/3m into the IPO, but the 7d shows the leaders (RKLB/ASTS/FLY) cooling while niche names (RDW/LUNR/PL) hold — consistent with capital rotating within the sector and with SpaceX's S-1 naming RKLB a competitor.
• Passive-flow bid — untested. No index-entry date has arrived; earliest is CRSP/VTI ~T+5 (~June 18) then Nasdaq-100 T+15 (~July 3). Provisional anyway until the deal prices (offering blank).
• Comp repricing — untestable pre-listing.
• Nothingburger — still alive. Friedman's float-math base case (0.08–0.12% S&P weight) is the bar the drain thesis must clear; nothing in the pre-event tape has cleared it.

Lane graduation: none of the four criteria have fired. S&P decision pending; no observable forced-flow effect yet (no entry date reached); confidential drafts ≠ public S-1; this is desk-pass #1. Keep as thread + scan lens.

Mutations done in this pass

• research/threads/2026-06-02-ai-mega-ipo-fast-entry-liquidity-regime.md — Dated Gates updated: S&P 06-08 marked conditional/pending; CRSP T+5 gate added; lockup-ladder dates added; methodology corrections folded in.
• research/workflow/SCANS.md — S&P methodology row annotated no-final-decision-posted-yet as of 2026-06-02; SpaceX tape-classifier row given this pass's signature.

Follow-ups filed (no child-track mutation)

• 2026-06-05 (SpaceX T-7) fresh tape read — narrow intraday fetch of the 22 basket names; test the moomoo book-build/funding-pocket drain in the pre-listing window. (TASKS-RESEARCH / SCANS.)
• 2026-06-08 S&P decision re-check — already filed; confirm against press.spglobal.com / mr4292; classify s-p-fast-track-live / s-p-partial-or-delayed / s-p-no-change.
• CRSP T+5 (~June 18) and Nasdaq-100 T+15 (~July 3) inclusion-date watch — the first real passive-flow tests once SpaceX lists.
• Reconcile the JPMorgan "$950B vs $95B" forced-selling figure against the primary JPM note before any number is published.
• SpaceX priced-deal re-read — when the S-1/A offering blanks fill, recompute float % and refresh all forced-flow magnitudes.

Update — 2026-06-02 (Economist + HN giga-IPO synthesis)

User dropped two primary-ish sources into the lane: The Economist's "Giga-IPOs" piece (full archive.today text) + companion Buttonwood column, and the HN thread (item 48364055). Both captured as inputs and cited above. Net effect: confirms the desk-pass methodology corrections, adds hard index-weight numbers, corrects one gate label, and hardens the "why this lane exists" structural case — without changing the tape classification or firing graduation.

Gate correction — the T+5 fast-track is FTSE Russell, now double-sourced. The Economist states plainly: "FTSE Russell has slashed its waiting time to five days"; HN/augstein independently confirms "FTSE Russell: cut to 5 days." Desk pass #1 attributed the ~T+5 gate to CRSP/VTI (from a single Moneywise/Yahoo secondary). Correction: name FTSE Russell (Russell 1000/3000) as the firmly-sourced 5-trading-day early gate (a very large passive pool); the CRSP/VTI ~5-day attribution was weaker single-source and may also be ~5 days but should not be stated as primary. Both Nasdaq T+15 and S&P-not-yet-decided re-confirmed by both sources.

New hard numbers — the float-math / nothingburger leg is now well-quantified (primary-ish):

• SpaceX initial S&P 500 weight ≈ 0.1% (free-float weighting; only ~$75bn floated) — Economist, corroborated by HN math (0.08–0.12%).
• NASDAQ-100 ≈ 0.5% even with the 3× free-float cap (NDX "changed its rule to weight up to 3× free float, in an apparent effort to woo Mr Musk").
• Russell 1000 ≈ 0.1%. Context: Russell 3000 ≈ $79trn, S&P 500 ≈ $69trn, NASDAQ-100 ≈ $40trn; ~$200bn raised across the trio is "a rounding error."
• Lockup ladder independently confirms our S-1/A read (366d Musk no-early- release; 20% after Q2 report; +10% if ≥30% above IPO; further dated tranches).

Mechanism refinement (HN): "forced buying" is not uniform — ETF index trackers have no discretion (must hold), but pension operators are typically NOT passive and large managers won't rebalance instantly (they match returns, not holdings tick-for-tick). So the mechanically forced pool is narrower than the headline "$30trn 401k" framing. Treat the recurring "$30trn / 19% float in 6mo / 24% Russell+NDX" figures as unattributed forum/secondary estimates — not for publication.

Structural thesis (strengthens the lane, does NOT fire graduation): the Economist + Haghani (Elm Wealth) frame the real story as a multi-year "capital diet" — tech giants slowing/halting buybacks, reinvesting in AI, tapping the bond market, while the newest members tap equity; "shares increasingly scarce" reverses. Plus concentration risk: the 10 biggest AI-related firms ≈ 40% of the S&P 500, so a giga-trio stumble read as bad-news-for-AI could precipitate a correction (equal-weight funds offer partial protection). Ritter's base rate (avg IPO −20pp vs market over 3y; >40× revenue −58pp; SpaceX ≈ >90× revenue) is a cross-link to the SpaceX child case, not parent-owned.

Revised synthesis of the Dunn-vs-Friedman tension: the Economist effectively adjudicates it — near-term sides with float-math (initial index weights tiny because of the ~4% float, so no mechanical broad drain at listing), long-term agrees with a slow drain as lockups expire and buybacks slow. That is fully consistent with desk pass #1's tape read (unrelated-or-insufficient on broad drain + early selective halo) and with the moomoo book-build thesis as the remaining near-term drain channel to test at T-7. Classification unchanged. Graduation criterion #2 still not fired (no observable forced-flow effect yet — no index-entry date reached). The lane's regime framing is materially hardened.

Mutation: thread Dated Gates updated to name FTSE Russell (T+5, double- sourced) and carry the 0.1% S&P / 0.5% NDX weight numbers; Desk-Pass Log appended.

Sources

• Tape data (price/RSI/trend figures): desk pre-computed watchlist summaries as of the artifact date — research/market-engine/data/summaries/*.json.