SpaceX 75B IPO|Semiconductor Sell-Off or Floor Test?

The Funding Problem

SpaceX priced its IPO at a fixed $135 per share before the roadshow even began — a deliberate break from every Wall Street convention about price discovery. The offering is oversubscribed, books closed Monday, and first trading on Nasdaq is set for Thursday. At $1.8 trillion, SPCX would debut larger than Amazon on day one.

That oversubscription has a specific structural cost. BNP Paribas quantified it: retail investors are expected to represent roughly 30% of the $75 billion raise, but retail does not hold idle cash. Micron alone absorbed $6.5 billion in net retail inflows over the past month, driving 87% upside. Nvidia, Broadcom, and AMD carry comparable crowding. U.S. leveraged equity ETF assets have reached a record $175 billion, concentrated in Nasdaq and semiconductor products. The reallocation into SpaceX means unwinding those positions — and leveraged ETF redemptions amplify the pressure by forcing derivative unwinds on the underlying names.

The Nasdaq-100 fell 4.8% and the Philadelphia Semiconductor Index dropped 10.3% in a single session last Friday, the week before the IPO. That move was not driven by an earnings miss or demand shock. It was positioning. Retail was already raising cash. The SpaceX offering has not yet settled — which means the pressure that already moved the chip index by double digits is only in its early stage.

The question Friday's drop raised is not whether retail will exit semis. The flow evidence says it already has begun. The question is whether the exit is a crowding reversal that discovers a new floor, or the beginning of a structural rotation out of the AI chip complex entirely.

Who Absorbs the Float

Intel answered part of that question on Monday with a 13% gap higher, reaching $111.90 and extending its year-to-date gain to 182%. The surface trigger was a report that Alphabet's Google placed an order for more than three million tensor processing units from Intel's foundry for 2028 delivery. Morgan Stanley estimates Google's total TPU production for 2027-2028 could reach six million units, putting Intel in position to capture roughly half of Google's chip manufacturing needs for that cycle.

Separately, Nvidia is running technical evaluations on Intel's 18A process for a forthcoming multi-chip package. No order has been placed. But the fact that Nvidia's team is running multi-project wafers at Intel — the standard qualification step before volume commitment — signals that the largest AI chip designer is actively hedging its TSMC dependency. TSMC's own CEO has publicly acknowledged that AI demand will exceed its capacity for years.

The Intel trade is structurally the opposite of the SpaceX-driven retail exit. Retail capital that had crowded into Micron and Nvidia on AI momentum is the same capital now funding SpaceX allocations. The institutional demand entering Intel is coming from a different premise: not AI euphoria, but supply chain diversification. Google and potentially Nvidia are not buying Intel foundry capacity because they believe in the rally — they are buying it because TSMC cannot supply enough. That is a demand signal that does not reverse when retail sentiment turns.

Cadence Design Systems formalized a multi-year DTCO agreement with Intel Foundry on Intel's 14A process this week, adding EDA validation to the customer pipeline. Intel's foundry revenue was $4.5 billion in its most recent quarter, still operating at a loss, but the forward order flow now has named anchors at the hyperscaler level.

What the Intel recovery does not settle is whether the broader semiconductor index has found its floor, or whether Intel's specific foundry story is a single-stock divergence inside a sector that retail is actively abandoning. Micron fell 13.3% Friday and recovered 9.9% Monday — a rebound that BNP Paribas analysts note may stall when the SpaceX allocation window closes and net flows from retail into chips have not yet confirmed reversal. The index move, not the Intel-specific foundry bid, is the live variable.

The Verification Window

The SpaceX IPO sets a concrete timetable. Pricing is expected Wednesday, June 11. Trading opens Thursday, June 12. The 30% retail allocation means roughly $22.5 billion of the $75 billion raise needs to clear through individual investors in a compressed window. That is the cash-raise event BNP Paribas described — and its resolution closes the forced-liquidation pressure from the SpaceX side.

The $175 billion in leveraged ETF assets concentrated in semiconductors and Nasdaq products is the amplifier that makes Thursday's open the most important microstructure moment of the quarter. If leveraged ETF redemptions have front-run the IPO — meaning the unwinding happened last week during the 10.3% semiconductor index drop — then the post-IPO period begins with positioning already cleared. If retail delayed selling until allocation windows confirm, the pressure lands post-Thursday.

Citi raised its S&P 500 target even after Friday's drop, citing the sell-off as positioning-driven rather than fundamental. Morgan Stanley set an 8,000 S&P target for year-end, characterizing the correction as "ultimately healthy." Those institutional reads share a premise: that the AI infrastructure demand cycle remains intact even as retail repositions into SpaceX. The Intel foundry orders from Google are the evidence supporting that premise — hyperscalers are not reducing AI capex, they are diversifying manufacturing.

The scenario where institutional reads prove wrong is not a demand collapse. It is a float problem. SpaceX's entire 555.6 million share offering represents only a 4% float, with Musk's 85% voting control locked for a year and Goldman Sachs leading a syndicate that restricts 100% of founder shares for 366 days. A tiny float with $50 billion in combined retail and passive demand creates a mechanical upside overshoot at open — and a correction once index inclusion (Nasdaq-100 fast-track rules were changed specifically to accommodate SpaceX) forces passive rebalancing. That rebalancing will generate the next semiconductor rotation signal, not this week's IPO day.

The specific number to track is the Philadelphia Semiconductor Index on Thursday's open. If it holds above its Friday low of approximately 4,800 — the intraday level from the 10.3% single-session drop — the institutional absorption thesis holds. If it breaks below that level on IPO-week volume, the leveraged ETF unwind has not yet cleared and the rotation out of chips into space infrastructure is still active.