The editorial argues the committee's decision is a statement about what the index is for — a rules-based product, not a discretionary one. Waiving the GAAP profitability rule for SpaceX would have set a precedent that eroded the index's claim to mechanical neutrality, so the line had to hold.
The editorial reframes the story by arguing most coverage gets the causality backwards: the S&P 500 isn't snubbing AI, it's refusing to treat profitability as a formality the way private markets have for three years. The committee is effectively running an actuarial check on companies like OpenAI ($500B) and Anthropic ($300B+) that are spending more on compute and talent than they earn.
By surfacing the Ars Technica reporting, the submitter highlights that roughly $13 trillion in assets track or benchmark against the S&P 500. Exclusion means losing a guaranteed daily bid from passive index funds — a material loss for SpaceX, OpenAI, and Anthropic that goes well beyond a vanity label.
The S&P 500 Index Committee declined to waive its profitability requirement to fast-track SpaceX, and in doing so quietly slammed the door on OpenAI and Anthropic for as long as they keep burning cash. The rule is mechanical: a company needs positive GAAP earnings in the most recent quarter and positive cumulative earnings across the trailing four quarters. SpaceX, despite a private valuation that briefly touched $400B, doesn't clear it. Neither do the two AI labs — Anthropic at a reported $300B+ secondary mark, OpenAI at $500B — both of which are spending more on compute and talent than they earn in revenue.
The committee's refusal to grant an exception isn't a political statement; it's a statement about what the index is for. S&P Dow Jones Indices runs the methodology as a rules-based product, not a discretionary one. Waiving the profit rule for a politically connected rocket company would have set a precedent that every unprofitable trillion-dollar private would then try to extract, and the index's entire claim to mechanical neutrality would erode. Per the Ars Technica reporting, internal guidance reiterated that GAAP profitability is the line, and the line is the line.
The downstream effect is what makes this a real story rather than index trivia. Roughly $13 trillion in assets either passively track or are benchmarked against the S&P 500. Inclusion isn't just a vanity label — it's a guaranteed daily bid from index funds rebalancing into the name. Exclusion means that bid never materializes.
Most coverage of this story is framing it as the index "snubbing" AI. That gets the causality backwards. The S&P 500 is doing what fundamental investors have stopped doing for the last three years: refusing to mark unprofitable companies as if profitability were a formality. The committee is, in effect, running an actuarial check that the private markets have abandoned.
Consider the math the committee is implicitly doing. To clear the trailing-four-quarter GAAP test, OpenAI would need to generate something on the order of $1–2B in net income across a year. Public estimates put OpenAI's 2025 losses at $5B+, against ~$10B in revenue, with compute and training costs scaling roughly with revenue. Anthropic's loss profile is reportedly steeper relative to revenue. Neither has a credible path to GAAP positivity inside 18 months without a step-change in unit economics — which would require either a collapse in inference costs or pricing power the market hasn't yet validated. The committee looked at that and said: not our problem to subsidize.
The second-order effect is on employee comp. Index inclusion is the mechanism by which RSUs become liquid wealth at a reliable clearing price; without it, the bid-ask spread on tender offers widens and discounts deepen. A senior engineer at OpenAI holding $5M in paper equity is now looking at a longer, lumpier path to that becoming spendable money. Stripe figured this out years ago and built a tender-offer cadence to compensate; OpenAI's tenders have been irregular, and Anthropic's are gated. Expect both to professionalize this in the next 12 months, or watch attrition tick up.
The HN reaction caught the right thread: the committee isn't wrong that these companies aren't earning money, and the rule was written for a reason. The reason was 2000, when including unprofitable dot-coms at peak valuations turned the index into a pro-cyclical loss machine for millions of passive investors. The S&P 500 institutional memory is longer than the venture capital cycle, which is the whole point.
There's a counter-view worth taking seriously: index methodology is increasingly out of step with where economic value actually lives. Tesla waited years for inclusion. Meta and Alphabet's early-stage analogs would have flunked the same test. Holding the line on GAAP profitability arguably mis-prices the index relative to the actual investable universe. But the alternative — discretionary inclusion of money-losing megacaps — is how you get the Nasdaq-100 in 2021, which then took a 35% drawdown when rates moved.
For practitioners, three things change. First, if you're holding pre-IPO equity at one of these companies, the secondary-market discount you should mentally apply just got wider. A reasonable haircut on OpenAI or Anthropic paper relative to last-round mark is now in the 30–45% range, not the 15–20% range that was implicit when index inclusion felt inevitable. Recalibrate your personal balance sheet accordingly.
Second, if you're evaluating offers, the comp comparison between an Anthropic SWE package and a Meta or Google package is no longer apples-to-apples on the equity line. The Meta RSU is liquid at a known price every vesting day. The Anthropic equity is liquid at an unknown price on an irregular tender schedule, with a now-uncertain path to public-market liquidity. Adjust the expected value of the equity component down by something like the implied tender discount plus a duration risk premium — call it 25–35%. The cash and signing bonus matter more than they used to.
Third, if you're building on these companies' APIs, this is a small but real input into supplier-risk planning. Companies that can't tap the cheapest pool of public-market capital are more sensitive to compute-cost shocks, more pressured on pricing, and more likely to introduce usage tiers or price hikes to chase GAAP positivity. The 2026 pricing changes you've been seeing are not unrelated.
The most likely consequence is acceleration. OpenAI's IPO timeline, which was vaguely "2027-ish," probably compresses — because the alternative is watching senior engineers leave for shops where the equity actually liquidates. Anthropic, structurally further from public-readiness, will lean harder into the structured secondary tender market and possibly a direct listing. SpaceX will keep doing what it's doing because Musk doesn't need the S&P bid. The interesting tell over the next two quarters will be whether any of the three start publishing real GAAP financials voluntarily — that's the signal that they've decided the index door is one they actually need to walk through.
Big relief for me. As a passive investor, I want the indices to follow the same passive strategy they always have, and specifically not make exceptions for specific companies like SpaceX wanted.Plenty of ways to get exposure to that stock without it going into the indices it is not qualified for.
Yep!! Respect to them. I was planning to move to an equal weight index but this gives me a little more time to evaluate options.
Is there any global indexes that include only companies which have been profitable in past? Say with same criteria of 1 quarter with 4 consecutive positive quarters overall? Might lose some gains, but probably also cut away lot of more risky bets.
This is very smart of these folks because for just three companies, they can't ruin the trust and impeccable reputation they have built over the years.This decision alone is worth several trillion dollars.
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.
Letting new stocks marinate in the market and get 4 quarters of SEC filings along with following all the GAAP accounting practices will definitely help evaluate them before inclusion. The last large boom/bust cycle had a couple of companies, at least, that were doing illegal things. I'm no