Financial markets are preparing for the IPOs of SpaceX, OpenAI, and Anthropic. These are not conventional companies; they represent some of the big economic narratives of our time: from artificial intelligence to the new space race. Precisely for this reason, the IPO of these private giants raises a recurring question: what will happen when companies of this size go public? And more specifically, can index funds and ETFs become forced buyers that artificially inflate their valuations from day one?
The hypothesis is intuitive, but incomplete. It starts from a simple idea: if more and more money replicates indices, and if these companies are already born with valuations of hundreds of billions of dollars, passive funds will necessarily have to buy them, pushing their prices up. Under this reading, indexing would act as a kind of automatic fuel for a potential tech bubble.
However, market reality is much more nuanced. Index funds do not set the price of a company on its stock market debut. Nor do they buy any company simply because it's large, famous, or trendy. Indexed investing follows very specific rules, and those rules are designed precisely so that indices represent liquid, investable, and relatively consolidated markets, not yet-to-be-proven expectations. Also notable is S&P's decision not to accelerate the entry of SpaceX, OpenAI, or Anthropic into the S&P 500. It will maintain its usual requirements: one year of trading and profitability. This prevents passive investors from automatically buying AI companies at very high valuations, although it also means they could miss out if those companies soar after going public.
In an IPO, the initial price is set by active investors. It is the placement banks, institutional funds, hedge funds, family offices, venture capital, and other professional buyers who determine initial demand; they accept or reject the proposed multiples; and they take on the risk of valuing a company at its debut. If a company debuts with an excessive premium, that premium does not come from indexing, but from the expectations that active investors are willing to pay.
This is especially relevant in the case of OpenAI or Anthropic. Both companies symbolize the explosion of artificial intelligence, but also one of its main unknowns: converting growth into sustainable profitability. The market may be willing to pay very high multiples for their potential, technological position, or ability to capture a substantial portion of corporate spending on AI. But major indices do not automatically include companies based on narrative potential. They need liquidity, trading history, sufficient free float, and in many cases, a clearer demonstration of financial viability.
Artificial intelligence has a particularity that should not be ignored: it is a business of enormous growth, but also of enormous costs. Investment in computing capacity, data centers, chips, energy, talent, and infrastructure can delay the generation of consistent profits for years. Therefore, if OpenAI or Anthropic were to come to market with extraordinary valuations, the origin of those valuations would be in the conviction, or exuberance, of active investors, not in a mechanical purchase by index funds.
The case of SpaceX is different, although it leads to a similar conclusion. Elon Musk's company combines a unique space narrative with a business, Starlink, that has already demonstrated strong commercial traction. Unlike other tech companies that are still loss-making, SpaceX could present itself to the market with more solid revenue and profitability metrics. But here the debate shifts to another area: free float.
If SpaceX were to go public with a very small portion of shares available to the public, some analysts might interpret that the scarcity of paper would cause additional buying pressure from index funds. But modern indices do not weight companies by their total theoretical value, but by their market capitalization adjusted for the capital actually available for trading. That is, if only a small part of the company is freely traded, the potential weight of SpaceX in indices would be adjusted to that reality. Passive demand would not be calculated on the entire company, but on the truly investable part.
This point is fundamental to understanding the role of indexed investing. Indices are not blind machines that buy anything at any price. They are methodologies that seek to represent an orderly part of the market. Companies enter, companies leave, weights are adjusted, free float is reviewed, and liquidity criteria are applied. There can be demand effects around an index inclusion, of course. But it is one thing to recognize those flows and quite another to believe that indexing creates bubbles in IPOs.
In fact, passive investing usually arrives after the active market has done the most challenging work: valuing the company, absorbing initial uncertainty, handling debut volatility, and establishing a price reference. In that sense, indexing does not anticipate market consensus; it replicates it. It does not discover the price; it incorporates it once the company consolidates certain conditions to be part of an investable universe.
This does not mean that the future IPOs of SpaceX, OpenAI, or Anthropic will be free of excesses. On the contrary, the combination of real innovation, abundant liquidity, growth expectations, and public fascination, can generate very demanding valuations. Stock market history shows that big tech narratives tend to go through phases of enthusiasm, correction, and subsequent discrimination between winners and losers. AI will be no exception.
*Giorgio Semenzato, CEO of Finizens
