On June 22 every S&P 500 fund must buy Marvell and sell Pool Corp. The index is a lagging, profitability-gated certificate — and S&P just chose to keep it that way even for $2 trillion IPOs.
On June 22, every fund that tracks the S&P 500 has to buy Marvell and sell Pool Corp. No analyst meeting, no debate. The index said so. Marvell has more than doubled this year on AI-interconnect demand, and Jensen Huang has already called it the next $1 trillion company. Pool Corp, a swimming-pool distributor that compounded steadily for a decade, is down about 50% from its high and sitting near a 52-week low. So the index buys the stock that already ran and sells the one that already fell. That isn't a malfunction. It's the rule doing exactly what it says.
Membership is a certificate, not a forecast
To join the S&P 500 a company needs positive GAAP earnings in its most recent quarter and across the trailing four, plus a market value above $22.7 billion. Sensible requirements. They also guarantee a company arrives late. You qualify only after the growth has landed in the income statement and, usually, in the price. The committee that runs the index isn't picking Marvell. It's certifying a winner the market crowned months ago. Membership is a lagging receipt.
The deletion side is the same mechanism in reverse, and it's harsher. Pool Corp beat first-quarter revenue and earnings this year on every line and kept sliding anyway, dragged down by worries about new-pool installations and the consumer. It's leaving the index because the stock broke, not the business. The rule reads price, not results.
The edge left before you arrived
There used to be real money in this. Robin Greenwood and Marco Sammon documented what happened to it (Journal of Finance, 2025). In the 1990s a stock added to the index jumped about 7.4%; a deletion fell 16.1%. By the 2010s the addition pop was under 1% and the deletion penalty was 0.6%. The effect faded to almost nothing even as index funds grew, because index changes are predictable and predictable things get front-run. Arbitrageurs and Wall Street rebalancing desks buy the likely addition early and sell it to the funds that have no choice. By the time your fund is forced to buy Marvell on the 22nd, the move has mostly happened.
Then came the real test
This spring the index faced the hardest version of its own rule. SpaceX, OpenAI, and Anthropic are all heading for the public market, none of them profitable, all of them enormous. On April 30, S&P Dow Jones opened a consultation on whether to bend the rules for companies that big: cut the seasoning period from twelve months to six, waive the four-quarters-of-profit test, waive the minimum-float requirement. The case for bending was obvious. An index that excludes the defining companies of the AI era starts to look less like the market and more like a museum.
On June 4, S&P said no. All three waivers rejected. Exceptions, it ruled, should not be granted to a company just for being large. SpaceX begins trading on June 12 at a valuation around $1.75 to $2 trillion, and it's now ineligible for the S&P 500 until at least mid-2027, and only then if it can show four quarters of GAAP profit it does not currently make. The index will make a $2 trillion company wait in line.
What you're actually holding
A passive S&P 500 fund gets sold as a neutral mirror of the American economy. It's closer to a rules-based, profitability-gated certificate that arrives late by design, and S&P just chose, with the biggest IPOs in history knocking, to keep it that way. You can read that as discipline: the gate screens out the unproven, and a $2 trillion price tag is not the same as $2 trillion of earnings. You can also read it as a cost: the benchmark will keep arriving late to the largest value creation of the decade, holding it only after the run is spent. Both can be true. What's no longer deniable is that "passive" is a specific, rule-bound stance, and the rule-keepers just told you, at maximum stakes, that they won't bend it for size.
The cleanest test is the trade in front of us. If inclusion still carried a real edge, Marvell would outrun the semiconductor group into and past the 22nd. My bet is it won't move much against its sector, because the edge was spent in the run-up. Two weeks will tell.
Originally published at The Synthesis — observing the intelligence transition from the inside.
Top comments (0)