01The setup
The largest IPO ever — on an extraordinarily thin float
SpaceX listed on Nasdaq as SPCX on June 12, 2026, at a $135 offer price — roughly 556 million new shares raising about $75 billion, the largest IPO on record. Yet of its ~13 billion total shares, under 5% actually floated; almost everything else sat locked up.
That razor-thin float was the root of the early volatility. With demand chasing scarce supply, the stock spiked intraday to $225.64 within four sessions, then slid back to the $110–120 range by late July as the coming supply came into focus.
02Korean exposure
How much Korean ETFs actually hold
Because Korean retail investors could not readily join a U.S. IPO, indirect exposure through space- and AI-themed ETFs surged. Six domestic ETFs that bought in on day one (Jun 12) together held about ₩334.5bn of SpaceX as of Jun 15.
| ETF (manager) | Weight | Amount | Est. shares |
|---|---|---|---|
| KODEX US AerospaceSamsung · 미국우주항공 | 25.08% | ₩188.6bn | ~870k |
| ACE US Space Tech ActiveKorea Inv. · 미국우주테크액티브 | 23.26% | ₩75.9bn | ~350k |
| TIME US Nasdaq-100 ActiveTimefolio · 미국나스닥100액티브 | 1.01% | ₩25.0bn | ~115k |
| TIGER Global AI ActiveMirae Asset · 글로벌AI액티브 | 2.90% | ₩17.5bn | ~80k |
| TIME Global AI ActiveTimefolio · 글로벌AI인공지능액티브 | 0.67% | ₩17.0bn | ~79k |
| TIME Global Space Tech & DefenseTimefolio · 글로벌우주테크&방산액티브 | 3.50% | ₩11.8bn | ~55k |
| Total (6 funds) | — | ~₩334.5bn | ~1.55M |
Funds that entered later at T+1/T+2 followed. Notably TIGER U.S. Space Tech — once a >₩2tn fund — added SpaceX at roughly 25.78%, making it the single largest domestic holder, alongside SOL, KIWOOM and Hana's 1Q (~13%). Adding these, total Korean-ETF holdings are estimated at ~3.5–4.5 million shares.
Korean-ETF share counts are derived from reported won amounts via price and FX. Exact figures appear daily in each manager's PDF (constituent) file. Shares that NPS, KIC and Mirae Asset took directly via IPO private-fund LP allocations are undisclosed and excluded.
03The core mechanic
When wanted shares exceed buyable shares, scarcity shows up as price
A thin float can face simultaneous demand from forced index buying, ETFs and retail. The key is that markets clear on price, not quantity. When demand exceeds supply, there is no "sold out" — the price rises until holders willing to sell appear. Scarcity surfaces as spikes, high volatility and premiums, not as zero shares.
The ~3.5–4.5m shares (~$0.4–0.5bn) Korean ETFs needed is small next to even the thin float (500m+ shares, tens of billions of dollars). The bottleneck was not Korean ETFs — it was the wall of global passive money from Nasdaq-100 inclusion plus retail FOMO.
04The index's self-limit
Index buying is float-adjusted — and therefore self-limiting
Major indices like the Nasdaq-100 weight by float-adjusted market cap — weight = price × floated shares, not total shares. So although SpaceX's headline cap is ~$1.7tn, only ~5% floats, so the index's "float cap" is far smaller — and the required passive purchase is capped accordingly.
This has an elegant property: passive buying = (total passive AUM) × (weight), and because weight tracks float, funds end up buying the same fraction of every stock's float. A thin float does not force passive money to buy a disproportionate slice of that thin supply.
The discipline applies to passive money only. Active/thematic ETFs and retail aren't bound by index weight and can buy aggressively versus the float — which is where squeezes come from. Korea's KODEX and TIGER holding ~25% each reflects product design, not float rules.
05Lock-up expiry
When the lock-up lifts, can demand absorb the supply?
There is a decisive asymmetry versus the IPO. At listing and inclusion, a one-time forced demand hit a fixed thin supply, pushing price up. A lock-up expiry is the opposite — an expansion of supply.
- Good news: a bigger float lifts the float cap and index weight, triggering additional passive buying in tranches at each rebalance. Demand isn't a one-and-done event.
- Scale: after August earnings ~911m shares become eligible (float → ~12%). If only 10% actually sells, that's ~91m shares (≈$10bn) — roughly the same order of magnitude as the incremental passive demand from the float step-up.
Two caveats, though. First, the IPO's forced buy was one-time — the big single-session purchase at July 7 inclusion doesn't repeat, and later rebalance tranches are smaller and spread out. Second, timing is mismatched — selling can hit immediately, while float updates and rebalances cluster on scheduled review dates. Supply now, demand later and lumpy, means volatility in between.
"Eligible ≠ sold" is the crux. Musk's ~46% is locked until June 2027, and employees and VCs often borrow against stock rather than sell, for tax reasons. And the fear is already in the price — the slide from $225 to the $110–120s pre-reflects the coming supply. Studies and market experience show a pattern of pre-expiry weakness, then absorption at lower prices if actual selling underwhelms, then recovery — the anticipation effect is frequently larger than the actual impact.
Bottom line: if real selling stays light (~10%) and is already priced in, index/active/retail demand can clear it near current levels — and lighter-than-feared selling can even produce a relief bounce. If selling runs heavier, clusters, or earnings disappoint, gradual passive demand won't keep up and price falls further. And this is a repeating gauntlet — August through December, then Musk's stake in June 2027.
06Takeaways
What to hold onto
- A ~4.9% float is the structural cause of the early spike ($135→$225). Scarcity clears through price, not quantity.
- Passive index buying is float-adjusted and self-limiting; real squeezes come from active and retail demand outside that rule.
- Korean-ETF exposure is an estimated 3.5–4.5m shares — a participant, not the bottleneck.
- Lock-up expiries grow both demand and supply, but asymmetrically. Whether demand "absorbs" it hinges on how much actually sells and how much is already priced in.