AI mega-IPO lane desk-pass #2 — T-8 pre-position tape: halo unwinding (sell-the-news), no drain, deal priced
AI mega-IPO lane desk-pass #2 — T-8 pre-position tape: halo unwinding (sell-the-news), no drain, deal priced
Question: AI mega-IPO lane desk-pass #2 — T-8 pre-position tape: halo unwinding (sell-the-news), no drain, deal priced Verdict: open
Signature (this pass): primary confirms-sell-the-news (Behavioral/regime);
secondaries confirms-nothingburger (Flow/drain leg) + confirms-dispersion
(space cools while neocloud holds). No graduation criterion fired — keep as thread + scan lens.
What we're asking
Desk pass #2 on the ai-mega-ipo-fast-entry-liquidity-regime thread, run at T-8 into the
~2026-06-12 SpaceX print. Two jobs: (1) classify the pre-position tape — which lane
hypothesis does the required basket actually support 4 trading days before pricing? (2) resolve
the two near-term gates — the 2026-06-08 S&P DJI decision and the ~2026-06-05 (T-7) pre-position
window. Folds in the now-priced deal
([[2026-06-03-spacex-s-1a-priced-135-per-share-75b-offering]]).
What we found
Required basket — code-computed tape (summary CLI, 2026-06-04; precompute run on fresh OHLC)
| Sleeve | Ticker | RSI | 7D | 30D | 3M | %52wkHi | Read |
|---|---|---|---|---|---|---|---|
| Broad oxygen | SPY | 68 | +0% | +5% | +10% | −1% | at ATH |
| QQQ | 77↑ | +2% | +11% | +22% | −1% | at ATH, overbought | |
| IWM | 57 | −1% | +4% | +10% | −2% | small-caps participating | |
| TLT | 53 | −0% | +1% | −3% | −7% | bonds flat/soft (no duration bid) | |
| Mag7 | NVDA | 51 | +1% | +8% | +17% | −9% | healthy, leads |
| MSFT | 53 | +1% | +4% | +6% | −23% | constructive, lags off high | |
| GOOGL | 39 | −8% | −6% | +19% | −12% | 7d pullback after strong 3m | |
| AMZN | 39 | −7% | −8% | +15% | −10% | 7d pullback after strong 3m | |
| META | 51 | −2% | +2% | −7% | −22% | laggard | |
| Neocloud | NBIS | 68 | +14% | +43% | +157% | −10% | ripping, near high |
| CRWV | 51 | +6% | −12% | +40% | −41% | volatile, off high | |
| CORZ | 73↑ | +6% | +37% | +83% | −5% | ripping, overbought, near high | |
| IREN | 61 | +1% | +32% | +49% | −15% | strong | |
| APLD | 54 | −6% | +25% | +56% | −12% | strong | |
| Space halo | RKLB | 49 | −20% | +43% | +60% | −24% | huge 3m run, sharp 7d unwind |
| FLY | 47 | −24% | +19% | +86% | −46% | huge 3m run, sharp 7d unwind | |
| ASTS | 55 | −13% | +57% | +3% | −20% | 30d strong, 7d pullback | |
| LUNR | 50 | −22% | +33% | +82% | −28% | huge 3m run, sharp 7d unwind | |
| RDW | 57 | −23% | +116% | +101% | −30% | explosive run, sharp 7d unwind | |
| PL | 51 | −11% | +12% | +66% | −17% | run + 7d pullback | |
| IRDM | 63 | −4% | +23% | +101% | −9% | strong, mild pullback | |
| SATS | 43 | −5% | +0% | +6% | −18% | flat |
(Price-truth from precompute only; table is code-computed, not prose math.)
Reading the four sleeves
- Broad oxygen is risk-ON at all-time highs (SPY RSI 68 / QQQ RSI 77↑ overbought, both −1% from 52wk high; IWM +10% 3m). There is no liquidity drain in the broad tape — the opposite. Expected: the forced-passive-demand drain mechanism fires at index inclusion (Russell T+5 ~06-18, Nasdaq T+15 ~07-03), not 8 days before pricing — so the broad-drain leg remains structurally untestable pre-listing, and what is visible (risk-on melt-up) cuts against a pre-emptive drain.
- The space halo ran huge and is now actively unwinding into the print. Every core halo name is up +60% to +116% over 3m (the halo is real and already paid) but −11% to −24% over the last 7 days (RKLB −20, FLY −24, LUNR −22, RDW −23). Desk pass #1 (06-02) flagged the halo "cooling at the 7d edge"; at T-8 that cooling has deepened into a sell-the-rumor unwind. This is the cleanest, most active signal in the basket.
- Neocloud stays hot (NBIS +157% 3m near high, CORZ RSI 73↑, IREN/APLD +49/56% 3m). The AI-IPO-supercycle risk appetite is intact — so the space cooling is space-specific, not a broad AI risk-off. The basket does not move as one bloc → dispersion.
- The moomoo "funding-pocket drain" test fails. Thesis: Mag7/NVDA get sold to raise the cash
pocket for SpaceX allocations. Tape: NVDA RSI 51, +1% 7d (fine); MSFT +1% 7d; the GOOGL/AMZN
−7/−8% 7d dips are idiosyncratic single-name pullbacks after strong 3m runs, not a uniform
Mag7 drawdown. No coordinated mega-cap drain →
confirms-nothingburgeron this leg.
Gate resolution
- 2026-06-08 S&P DJI MegaCap decision — STILL PENDING as of 2026-06-04. Comments closed 5/28;
no adoption notice posted; roadshow underway (~6/4). The methodology change (MegaCap exceptions +
six-month IPO-seasoning) that would govern the non-500 fast-track legs (TMI / Completion /
DJ-TSM) is not yet live. S&P 500 itself stays Index-Committee-discretionary regardless.
Classifier remains unresolved → carry the
06-08gate. - ~2026-06-05 (T-7) pre-position window — resolving now as a HALO UNWIND, not accumulation. The pre-position tape shows the halo being sold into the event and no funding-pocket drain. Positioning ahead of T-7 is de-risking the crowded halo trade, consistent with the net-flow model's read that the anticipatory bid was the trade and the IPO itself sets up as overhang.
Priced-deal fold-in (event-truth)
The deal is no longer blank: $135.00 fixed / 555,555,555 Class A shares = $75.0B base + 83,333,333 greenshoe ($11.25B), prices ~06-11, trades ~06-12 ([[2026-06-03-spacex-s-1a-priced-135-per-share-75b-offering]]). The realized terms land on the net-flow model's parametric base case (V≈$1.75T, R=$75B, float₀≈4.3%), graduating it parametric → deal-confirmed; verdict unchanged (early passive bid ~10× too small vs staggered lockup supply; net flips to overhang ≈ Q2 print). Greenshoe (+~0.6pp float) is immaterial to the asymmetry.
Verdict + reasoning
- Primary signature:
confirms-sell-the-news. The dominant, active signal at T-8 is the space halo unwinding into the print after a 3m run — buy-the-rumor exhausting before the event. This is the natural successor to desk pass #1's "earlyconfirms-halo, cooling at the 7d edge." - Secondary:
confirms-nothingburgeron the Flow/drain dimension. No broad-market drain (tape at ATH/overbought), no funding-pocket Mag7 drain. The drain mechanism is structurally a post-inclusion event; pre-position, there's nothing to confirm and a soft counter-signal. - Secondary:
confirms-dispersion. Space halo de-risks while neocloud holds — the basket is not one correlated bloc. This is itself evidence against graduation criterion #1 ($SPCXexplains cross-market behavior beyond the space basket). - No graduation. Criterion #1 (cross-market) actively contradicted by dispersion; #2 (S&P/Nasdaq forced-flow observable) blocked by the pending 06-08 gate; #3 (OpenAI/Anthropic public S-1s treated as one regime) not yet — both remain confidential drafts. This is desk pass #2; criterion #4 needs ≥3 passes leaning on the classifier. Keep as thread + scan lens.
What would change the read
- At the print (~06-12): if
$SPCXopens weak and the broad tape also sells → re-testconfirms-sell-the-newsas a cross-market event (toward graduation #1). If$SPCXrips and drags the halo basket back up → the unwind was just pre-event de-risking, halo intact. - 06-08: an actual S&P adoption notice flips the
06-08gate tos-p-fast-track-liveand makes the forced-flow leg testable (graduation #2). - Lockup/earnings (≈ Q2 print, late Jul/Aug): the net-flow model's bid→overhang flip — the first real supply test.
Desk pass #2 on research/threads/2026-06-02-ai-mega-ipo-fast-entry-liquidity-regime.md.
Tape: code-computed summary CLI, 2026-06-04, precompute on fresh OHLC. Priced-deal input:
[[2026-06-03-spacex-s-1a-priced-135-per-share-75b-offering]]. Net-flow model:
research/investigations/2026-06-03-spacex-net-flow-model-passive-demand-vs-lockup-supply.md.
Predecessor: research/investigations/2026-06-02-ai-mega-ipo-lane-desk-pass-methodology-and-pre-event-tape.md.
Sources
- Tape data (price/RSI/trend figures): desk pre-computed watchlist summaries as of the artifact date —
research/market-engine/data/summaries/*.json.
Cited
1 eventRelated
5 eventsSource links
Tweets, videos, filings, articles, and source pages tied to this read.