When SpaceX goes public at an expected $1.75 trillion valuation with approximately 5-15% of shares floating, index funds tracking the Nasdaq-100, S&P 500, MSCI, and Russell will be required to purchase shares to match their benchmarks. At low float, the Gabaix-Koijen Inelastic Markets Hypothesis suggests this forced buying could inflate the stock price 14-32% above its IPO valuation, with a modal estimate of 19.5%. If insiders sell 10-20% of shares at the inflated peak, the resulting wealth transfer may reach $17-55B (modal $30B, MC median $33B).
These estimates depend on uncertain parameters — particularly the actual IPO float, post-lockup supply, the price-impact multiplier, and the degree to which NDX capping mechanisms attenuate post-lockup forced buying. The model was reviewed by two independent research councils, updated for Nasdaq's final May 2026 low-float rule, and checked against Tesla's December 2020 S&P 500 inclusion as a single-event calibration reference. A detailed practitioner critique identifies additional attenuating factors we address in Section XI.
Reading these numbers: The modal scenario uses the central value of each parameter distribution—a single deterministic point estimate. The Monte Carlo median (10,000 simulations sampling all parameter ranges) produces slightly higher estimates: $371B overvaluation, $33B realized transfer, $198 per $100K. All figures in this report use the modal scenario unless otherwise noted.
| Fix | Issue | Effect |
|---|---|---|
| Smooth convexity | Hard 5% step-threshold replaced with logistic sigmoid | Eliminates discontinuity; ~40% of impact from convexity |
| Nasdaq final rule | 5x / 20% proposal replaced by final 3x low-float cap | Initial NDX flow materially lower; no 20% cliff |
| Three-tier wealth transfer | Algebraic identity replaced with modeled staggered sales | Realized ($30B) < extraction cap ($51B) < peak overval ($341B) |
| Temporal decay | No decay between cascade epochs | Premium mean-reverts at 8%/mo between forced-buying events |
| Tesla calibration check | Full-cascade validation claim revised | Single-event order-of-magnitude sanity check |
| L1/L2 swap | Passive before front-running | Front-runners deplete supply before passive funds arrive |
| S&P eligibility gate | Always included | S&P 500 gated on float ≥ 10%; S&P TMI fast entry remains consultation-only |
| Supply depletion | Fixed supply | Each level faces tighter supply (10% floor) |
| Dynamic NDX denom | Fixed $27T | Denominator updated as SpaceX mcap grows |
Gabaix and Koijen (2021/2024) establish that the aggregate stock market has a price-impact multiplier of approximately 5: each $1 of net flow raises aggregate market cap by ~$5. Most equity is held by investors who cannot respond flexibly to price changes—mandated pension funds, index trackers, households with inertial allocations. Hedge funds hold only ~5% of the equity market. When new money arrives, prices must move substantially to induce the marginal seller.
At the individual stock level, the picture is more extreme. Davis, Kargar & Li (2024) estimate demand elasticities as low as 0.3, implying a micro multiplier of ~3.3. Kyle (1985) shows that price impact per unit flow rises as liquidity thins. Amihud (2002) captures the empirical regularity that impact per dollar increases with illiquidity. Wurgler & Zhuravskaya (2002) find that stocks without close substitutes experience larger price jumps upon inclusion. SpaceX has no close substitute—it is a monopoly orbital launch provider.
When forced buying approaches available supply, impact per dollar accelerates. The model uses a smooth logistic transition based on Kyle/Amihud supply-exhaustion dynamics rather than a hard threshold. At modal parameters, convexity contributes approximately 41% of total price impact. This is the single largest contributor to overvaluation after the base multiplier itself.
Greenwood & Sammon (2025) find the S&P 500 inclusion effect has declined from +7.6% in the 1990s to +0.8% in the 2010s. The primary mechanism: firms absorb passive demand via equity issuance and stock-based compensation. This channel would be largely unavailable to SpaceX at 5% float during the 15-day Fast Entry window, though secondary issuance after IPO remains possible.
Passive growth makes markets more inelastic. Haddad, Huebner & Loualiche (2025, AER) find passive investing has made demand 11% more inelastic over 20 years. Chinco & Sammon (2024) show true passive ownership is ~33.5%, roughly double the headline 16% figure. The model applies a 1.11x passive boost to both multipliers.
The model runs a sequential cascade where each level's output becomes the next level's input. If the Gabaix-Koijen multiplier applies at each stage (as the IMH predicts), the compound effect across different pools of capital could produce substantial price inflation. The following are modal-scenario estimates at 5% float:
The cross-index reflexive cascade is the model's most original contribution, independently identified by both research councils as a genuine analytical advance not previously modeled in the academic literature. Each downstream index treats the upstream-distorted price as exogenous, causing the Gabaix-Koijen multiplier to compound across index families rather than applying once.
v4 computed wealth transfer as a simple algebraic identity. Both research councils flagged this as the ceiling, not the modeled outcome. v5.1 decomposes wealth transfer into three tiers:
The realized transfer uses a 6-tranche staggered selling model where each tranche's sale reduces market cap by Msell × tranche_flow / mcap. The selling multiplier is discounted by 0.2x relative to the buying multiplier to reflect post-lockup conditions: 4x more float, gradual sales over 6 months, higher arbitrageur absorption, and anticipated selling partially priced in.
| Sell % | Extraction Cap | Realized P10 | Realized Median | Realized P90 |
|---|---|---|---|---|
| 10% | $37B | $18B | $27B | $39B |
| 15% | $56B | $20B | $33B | $50B |
| 20% | $74B | $18B | $35B | $57B |
The realized transfer would be borne by passive index fund investors through diluted returns, distributed across ~$16.65 trillion of passive capital tracking the affected indices. At the Monte Carlo median:
Tesla's December 2020 S&P 500 inclusion remains the closest historical analogue, but v5.1 treats it as a single-event calibration check rather than a full validation of the SpaceX cascade.
| Parameter | Tesla (Dec 2020) | SpaceX (proposed) |
|---|---|---|
| Float at inclusion | ~80% | ~5% |
| Seasoning | Standard (multiple quarters) | 15 trading days |
| Float cap / multiplier | None | 3x cap (NDX) |
| Forced buying / supply | 15-20% of float | 10-35% of float |
| Lock-up | N/A (already public) | S-1 lock-up schedule |
| Cross-index cascade | Already in NDX | NDX → S&P → MSCI → Russell |
| Close substitutes | Multiple EV peers | None (monopoly) |
| Firm issuance channel | Active | Disabled (5% float) |
| Metric | Observed | Model Median | Model P25-P75 |
|---|---|---|---|
| Pre-inclusion appreciation | 57% | 78% | 64% – 93% |
| Forced buying | $86B | $86B | — |
The observed 57% appreciation is below the single-event median but in the same order of magnitude. This supports the rough flow-multiplier scale, but it should not be cited as a formal validation of the SpaceX cascade.
Research Affiliates documented that in the six months following inclusion, Tesla underperformed the deleted stock (Apartment Investment & Management) by 78.6 percentage points, consistent with the pattern that index additions systematically buy high.
| Metric | P5 | P25 | Median | P75 | P95 |
|---|---|---|---|---|---|
| Total Forced Flows ($B) | 59 | 65 | 70 | 74 | 81 |
| Overvaluation ($B) | 244 | 314 | 371 | 438 | 551 |
| Overvaluation (%) | 14% | 18% | 21% | 25% | 32% |
| Realized Transfer ($B) | 17 | 26 | 33 | 41 | 55 |
| Cost per $100K | $101 | $155 | $198 | $247 | $332 |
| Convexity contribution (%) | 29% | 36% | 40% | 43% | 47% |
| Scenario | Float | Flows | Overval Med | Transfer Med | Peak MCap | $/100K |
|---|---|---|---|---|---|---|
| Extreme | 2% | $63B | $344B | $29B | $2.09T | $177 |
| Bear | 3% | $65B | $354B | $31B | $2.10T | $184 |
| Base | 5% | $70B | $372B | $33B | $2.12T | $199 |
| Moderate | 8% | $77B | $398B | $37B | $2.15T | $220 |
| Moderate | 10% | $96B | $420B | $39B | $2.17T | $237 |
| Bull | 15% | $116B | $479B | $48B | $2.23T | $289 |
| High | 20% | $137B | $562B | $60B | $2.31T | $359 |
Higher float reduces the initial squeeze but can increase downstream float-adjusted weights. Under Nasdaq's final 3x cap, the NDX lock-up effect rises smoothly until the cap stops binding around 33.3% float.
He, Kondor & Li (2025) demonstrate that standard demand elasticity estimation understates true inelasticity by a factor of 2-7x. Both research councils recommend this remain a sensitivity analysis, not the baseline, as the correction is theoretically motivated but empirically unvalidated at this scale.
| Metric | P5 | Median | P95 |
|---|---|---|---|
| Overvaluation ($B) | 633 | 1,207 | 2,372 |
| Realized Transfer ($B) | 43 | 95 | 193 |
| Peak Market Cap ($T) | $2.4T | $3.0T | $4.1T |
SpaceX is not an isolated event. Multiple mega-IPOs are expected in 2026-2027, each subject to the same forced-buying mechanics.
| Company | Valuation | Float | Forced Flows | Realized Transfer | Cumulative |
|---|---|---|---|---|---|
| SpaceX | $1.75T | 5% | $70B | $33B | $33B |
| OpenAI | $1.00T | 5% | $40B | $19B | $52B |
| Anthropic | $0.45T | 8% | $20B | $10B | $62B |
| Stripe | $0.15T | 10% | $8B | $4B | $66B |
| Databricks | $0.08T | 10% | $4B | $2B | $68B |
Caveat: Each IPO is modeled independently. Cumulative figures are additive estimates, not a state-aware sequential simulation. Actual cumulative effects may be smaller (overlapping AUM, active capital depletion) or larger (crowding, narrative momentum). S&P itself has calculated that if the 10 largest venture-backed US companies all joined the S&P 500, their combined weight would be approximately 4.5%—more than the entire energy sector.
As of May 2026, all three major US index providers are responding to the coming wave of mega-IPOs, but the rule status now differs by provider. Nasdaq's change is final; S&P and FTSE Russell remain consultation/scenario inputs in the sources reviewed.
Final methodology effective May 1, 2026: Fast Entry for top-40 market-cap companies after the 7th/10th/15th trading-day process, with weighting based on Modified Market Capitalization and a 3x free-float cap for low-float securities. This replaces the earlier 5x / 20% proposal.
February 18, 2026 consultation: proposed its own fast-entry mechanism, explicitly naming SpaceX, OpenAI, and Anthropic. Current March 2026 ground rules still retain quarterly/semiannual timing, 5% minimum free float, 5% public voting rights, and lock-up shares excluded until expiry unless final changes are published.
April 30, 2026: S&P opened a mega-cap consultation. The S&P 500 proposal does not create automatic fast entry; it contemplates narrower MegaCap exceptions while preserving committee discretion and eligibility constraints. The dashboard therefore treats S&P early entry as a scenario, not a base-case fact.
The structural parallel to 2006-2007 credit rating agencies is worth examining. Both are private companies whose revenue depends on the entities whose products they assess. Both control infrastructure with public-good characteristics. Both face competitive dynamics that penalize unilateral maintenance of standards. Whether this analogy is precise or overstated depends on how much weight one gives to the attenuating factors (Section XI) — including capping mechanisms, committee discretion, and the fact that index providers may exercise Expert Judgment to phase inclusions.
Most unmodeled channels operate in the direction of amplifying forced flows, though some (e.g., IPO allocatee flipping, OTC hedging) could provide additional supply. On balance, the omissions suggest the model may understate the total effect, but the magnitude of this bias is uncertain.
The public discussion of the NDX consultation has produced important analysis on multiple sides. Keubiko's "Nasdaq's Shame" and George Noble's analysis raised the initial alarm (endorsed by Michael Burry). A detailed practitioner rebuttal by bau postings challenges several of those claims with index methodology expertise. We draw on all three and address their arguments here.
As bau postings argued during the consultation, the early "5x multiplier dramatically increases forced buying" framing was misleading. Nasdaq's final rule is more conservative still: 10% float now receives 30% of listed market-cap weight, not 50%, and 5% float receives 15%, not 25%. The accurate statement is narrower: NDX's modified market-cap methodology still gives low-float mega-caps materially more weight than a purely float-adjusted index, but the final rule attenuates the original scenario.
As bau postings catalogs extensively, MSCI includes stocks at day 10, FTSE at day 5, TOPIX at 4-9 weeks, KOSPI at 15 days, S&P/ASX at 5 days. Fast entry alone is not unusual. What is unusual is the combination of fast entry with market-cap (not float-adjusted) weighting at very low float. Neither fast entry nor market-cap weighting is novel individually; the conjunction at $1.75T scale is what produces the effects this model quantifies. As Keubiko notes, the simultaneous relaxation by three providers is the structural concern.
The final Nasdaq methodology uses Modified Market Capitalization, weight interpolation for Fast Entry / quarterly additions, and concentration constraints. The model now reflects the 3x low-float cap but still simplifies interpolation and full capping mechanics. This remains a legitimate attenuation factor for L7/L8 and should be modeled explicitly in v6.
The preliminary S-1 leaves share count and price blank, so the 5% IPO float remains a scenario input. It does provide materially better lock-up information: a 180-day company lock-up, a 366-day founder/significant-investor lock-up, and staged early releases for some shares. The model's float parameter is adjustable via the interactive simulator, but v6 should ingest amended S-1 share counts directly.
The model treats post-lockup supply through the arbitrageur absorption parameter, but bau postings makes a stronger point: many locked-up investors will actively want to sell at $1.75T (70-100x TTM revenues). They are not captive — they are eager sellers. OTC forwards, margin lending (as with the Alibaba IPO precedent bau postings cites), and phased lockup expiry all provide earlier supply channels. This suggests our arb_absorption_lockup parameter (0.30) may be too low, and our post-lockup selling multiplier discount (0.2x) may be too generous to the distortion estimate.
None of the above critiques — including bau postings' detailed rebuttal — engage with the Gabaix-Koijen multiplier: the empirical finding that $1 of flow moves market cap by ~$5. Bau postings treats $14-18B of forced buying as having $14-18B of market cap impact. The multiplier is what transforms moderate forced flows into substantial overvaluation, and it is the most empirically well-grounded parameter in our model (Gabaix & Koijen 2024, JPE; range 3-8 across specifications). Similarly, none address the cross-index reflexive cascade, the convexity of price impact at extreme flow-to-supply ratios, or the structural inability to short a 5% float stock (Miller 1977; Duffie, Garleanu & Pedersen 2002).
The critiques are strongest on framing and on specific attenuation factors (capping, float uncertainty, post-lockup supply). They are weakest on the core mechanism: what happens when $50-100B of forced flows hits an inelastic, un-shortable, low-supply stock. The direction of our analysis is robust. The magnitude has genuine uncertainty — perhaps a factor of 2x in either direction — which is why we provide an interactive simulator where every parameter can be adjusted.
The model is implemented in Python (spacex_inclusion_model.py) with:
The model was reviewed by two independent multi-model research councils (Claude, GPT, Gemini), which identified 10 corrections and converged on the cross-index reflexive cascade as the central analytical contribution.