Default 401(k) funds keep buying AI giants whatever their profits are made of

"So whenever a company like Nvidia grows and takes up more of the S&P 500, the fund automatically buys more of it," said Hera Hyeonseo Lee, a doctoral researcher at Binghamton University ^[4]. Fortune reports that this happens no matter how expensive or big the company gets ^[5]. A cap-weighted fund buys every stock in the market and more of the biggest ^[19], so the size of its order depends on market value alone. It does not look at what the earnings under that value are made of ^[19]. Lee's example is Amazon, or rather Amazon's income statement. "In the first quarter of this year, Amazon's net income was almost $30 billion," she said. "But almost $17 billion came from its Anthropic mark-to-market gain." ^[18]

The mark was about 57% of the quarter's profit ^[1]. Take it out and Amazon earned roughly $13 billion ^[2], the same figure as its total investment in Anthropic ^[16]. Anthropic was a new unicorn when Amazon first invested in 2023 and was valued at $965 billion by mid-September ^[16]. The gain counts as income because a 2016 accounting change lets companies with private-market investments book valuation bumps, unrealized gains, as earnings ^[17]. A saver holding Amazon at index weight is holding part of a private company's latest price, reported as profit ^[17]. Fortune does not give the size of Amazon's stake or its cash flow for the quarter, so how much of Amazon's market value rests on the mark cannot be worked out from this record.

Jim Rowley, global head of indexing strategy and solutions at Vanguard, states the first objection. "Market cap weighting isn't a choice, it's not a methodology, it just is," he said ^[6]. "Investors collectively have decided that one stock should be this large, or another stock should be this small." ^[7] In his account, anyone who weights a stock differently from that consensus is investing actively ^[8]. A second objection is that the weights may be earned. If the biggest companies' cash profits grow into their prices, seven stocks at something like 30% of the S&P 500 ^[13] is what cap weighting is supposed to produce. One quarter at one company says little about the rest of the group.

The third possibility cuts the other way. SpaceX, which passed a $2 trillion valuation on its first day of trading in June, joined the Nasdaq 100 in 15 days under the exchange's new fast-entry rule ^[9]. The traditional seasoning period was three months, meant to let a new listing build a trading history and to limit disruption in major indexes ^[10]. Fifteen days is about a sixth of that ^[3]. Fortune expects rules like this to come into play as Anthropic and OpenAI head toward trillion-dollar-plus IPOs ^[11]. The rule governs the Nasdaq 100, and Lee's example of the default fund uses the S&P 500 ^[4].

Of the savers, Lee said: "They didn't choose to buy these stocks, and they don't really know what's going on." ^[1] I think the second half of that sentence is stronger than the first. A saver in a default did choose, in the narrow sense Rowley means, to let prices set the weights ^[6]. The makeup of the profit behind those prices was never part of that choice. Ed Yardeni, president of Yardeni Research, said the largest stocks "are going to be increasingly volatile because they're increasingly controversial" ^[14]. Valentin Haddad of UCLA Anderson said, "It's concentration and size, all packed together, even more than there was during the dotcom bubble." ^[15] The thesis fails if Amazon reports net income near $30 billion in a quarter when Anthropic's valuation does not move.