Eleven days ago I published The Fifteen Days, about SpaceX joining the Nasdaq-100 fifteen trading days after its IPO. That entry documented the mechanics: a fast-track rule written in May, up to $27 billion in potential passive buying across index-linked products, a public float of about 5 percent. It closed on the lockup schedule. It should have closed on a question. Who sells to a buyer that has to buy?
The eleven days answered it. SpaceX entered the index before the open on July 7 and traded around $158.77 that day, up about 1 percent. On Wednesday, July 15, the stock slipped below its $135 IPO price for the first time. Friday it closed at $123.99. The stock is down about 20 percent in July, and roughly $1 trillion in market value is gone from the June 16 peak of $225.64. An index fund that bought at the inclusion price is down about 22 percent in nine trading days. The rule worked exactly as written. That's the problem.
The Other Side of a Forced Trade
The other side of the trade now has a shape. About 185 million SpaceX shares are sold short, roughly 29 percent of the publicly tradable float and about $25 billion in bearish positions, according to S3 Partners. Three weeks ago the short position was an estimated 40 million shares, between 5 and 7 percent of the float. Draw the window. Three weeks back from mid-July lands in late June, when Nasdaq's June 26 announcement told every trader in the market that a price-insensitive buyer was arriving, with a published date and a size you could calculate from the index math.
"We are seeing continuous demand from short sellers building speculative positions since the IPO," Matthew Unterman, head of research at S3 Partners, told CNBC. Continuous demand. The phrase usually describes buyers.
Ortex puts the shorts' paper profit at $8.7 billion since the IPO. Its co-founder Peter Hillerberg told Reuters the stock "has been a rollercoaster for the short sellers, and it has ended up firmly in their favor. Rather than take profits, the bears kept adding the whole way down." By Ortex's count, close to half the free float is now out on loan, and each $1 move in the stock is worth more than $300 million to the short side.
What I Got Wrong
The Fifteen Days treated forced passive buying as a distortion of price discovery, something the market would need time to digest. Too gentle. A buyer who can't say no, whose arrival date is announced three weeks in advance, and whose order size can be estimated from public index rules is the best counterparty a short seller will ever meet. I also described the float-adjusted weight cap as a protection: the index refusing to assign a weight the available shares couldn't support. It protects the weighting arithmetic. It does nothing for the people who hold the weight. The mechanism was designed so the index could absorb SpaceX. Nobody designed anything to protect the index's investors from the price they would absorb it at.
One rule did work. S&P Global rejected its own fast-track proposal on June 4, keeping the 12-month seasoning period and the requirement of four consecutive quarters of positive GAAP earnings, writing that exceptions "should not be granted solely based on market capitalization." Five weeks ago that read like the old guard defending a slow process. Today it reads like the only risk control in the index business that fired. S&P 500 investors own none of this drawdown. The seasoning requirement everyone called outdated turned out to be the seatbelt.
The Supply Ahead
The short position is built for what comes next. SpaceX's initial float was about 5 percent of roughly 13 billion shares outstanding, and KeyBanc estimates the first major unlock arrives around second-quarter earnings, when about 11 percent of outstanding shares become eligible for sale. Further tranches of roughly 4 percent each begin around day 70 after the IPO, with more tied to performance milestones and third-quarter earnings. Elon Musk's stake, about 42 percent of shares outstanding, stays locked until June 2027. Every unlock adds supply on top of the shares the forced buyers already paid for.
Reuters attributes part of the weakness to investor concern over debt-funded AI spending. The Fifteen Days noted that index buyers were purchasing a company that had already changed what it is: $27.8 billion in annual AI compute contracts disclosed in the S-1, none of it producing revenue yet. In June that read as ambition at a blended price. At $124 it reads as exposure.
The Abort
On Thursday the catalyst that was supposed to steady sentiment aborted on the pad. Starship Flight 13 shut down at T-minus zero after at least four Raptor engines failed to meet their starting parameters. The flight software looked at the inputs, decided the risk was unacceptable, and refused to launch. Musk said two engines need replacing. The next attempt targets July 20. It was the right call, and it's the exact capability the index lacks. Inclusion has no startup check. The rule fired on schedule, the funds bought on schedule, and the only participants free to abort the trade were the ones on the other side of it.
The chill is spreading to the rest of the 2026 listing class. Anthropic filed its confidential S-1 on June 1, OpenAI is expected behind it, and none of the three AI-era giants can enter the S&P 500 this year under the rules S&P just reaffirmed. Whoever prices the next $1 trillion AI listing will price it for a market that watched the year's flagship IPO trade below issue inside five weeks, with the index bid spent and the shorts still adding. The Fifteen Days asked what happens when an index must buy a company the market hasn't priced. Now we know. The market prices it afterward, and it sends the bill to the funds.