In 2017 a photo-sharing app called Snap went public selling shares that came with no votes whatsoever. You could buy in, own your sliver of the company, and have precisely no say in how it was run.
The index providers took one look and changed the rules. S&P barred firms with multiple share classes from the S&P 500. FTSE Russell did much the same.
The reasoning was straightforward. An index is meant to be a mirror held up to the market. Let in companies that give outside shareholders no real control, and the mirror starts to flatter.
Nine years on, the same argument is back. Only this time it’s about the three biggest stock-market debuts the world has ever seen.
The class of 2026
Three private giants have filed to go public in the same few months, which has more or less never happened.
SpaceX leads. It’s selling 555.6 million shares at $135 each, raising around $75 billion at a valuation of roughly $1.77 trillion. That’s the largest IPO in history, comfortably past the $29 billion Saudi Aramco raised in 2019, and would make SpaceX, on day one, about the seventh-largest listed company in America. It lists on the Nasdaq as SPCX, with UK trading expected mid-June.
Anthropic, maker of the Claude AI models, filed confidentially on 1 June. Its last private round in May valued it at $965 billion, enough to overtake its rival OpenAI for the first time. Analysts expect it to cross a trillion when it lists, probably in the autumn.
OpenAI, maker of ChatGPT, filed a few days earlier and is targeting the final quarter of the year somewhere between $852 billion and a trillion.
Three companies. Comfortably over three trillion dollars between them. All arriving at once.
What you’d actually be buying
It’s worth pausing on what these businesses do, because the three could hardly be more different and only one of them currently makes money.
SpaceX is really three companies stapled together. There’s the rocket business, which launches more than eight in every ten kilograms humanity puts into orbit. There’s Starlink, the satellite-internet arm, which sells a dish and a monthly subscription and now has 10.3 million customers across 164 countries. And there’s an AI arm, bolted on in February when SpaceX absorbed Elon Musk’s xAI.
Of those, Starlink is the engine. It brought in $11.4 billion last year — about 61% of group revenue — and it’s the only segment turning a profit. The rockets roughly wash their face; the AI arm loses money prodigiously. Group revenue was $18.7 billion in 2025, up a third on the year. The group as a whole still lost around $4.9 billion, most of it down to the AI side. The jury is out on whether Musk is using the SpaceX listing to re-seize control of Tesla.
Anthropic and OpenAI sell something more abstract, access to a model, by subscription and through enterprise contracts that bill by usage. The numbers are extraordinary and the losses, in OpenAI’s case, are too.
OpenAI is taking in roughly $2 billion a month, running at something like a $25 billion annual rate, with around 50 million consumer subscribers and nine million business users. It also reportedly lost about $1.22 for every dollar of revenue in the first quarter. Growth and gravity, in the same set of accounts.
Anthropic has grown faster still on paper, a revenue run-rate of around $47 billion by May, up from roughly $10 billion a year earlier and has told investors it expects its first profitable quarter. It’s also structured as a public benefit corporation, with a trust sitting above the shareholders that’s meant to keep the mission ahead of the share price. Whether public markets reward that or merely tolerate it is an open question.
A word of the obvious, since it needs saying: a famous name is not the same thing as a safe investment, and the value of shares can fall as well as rise. Two of these three are spending more than they earn, by design, on the bet that scale arrives before the cash runs out.
How a British investor gets in
For most of these names, the answer is: not yet, or only sideways. You can’t buy OpenAI or Anthropic directly until they list.
SpaceX is the exception. In an unusual move, it’s reserving around 30% of its float for retail investors, roughly triple the usual slice and that includes a window for the UK. A long list of platforms have signed up to take orders, among them Hargreaves Lansdown, AJ Bell and Interactive Investor, alongside several others.
There’s also the indirect route, which some UK investors have quietly been on for years. Scottish Mortgage, the Baillie Gifford investment trust, backed SpaceX as a private company back in 2018; the stake now accounts for around a fifth of the trust. Getting exposure to a company before it’s public, through something you can buy on an ordinary platform, is one of the few genuine advantages the trust structure still offers.
None of which is a nudge in any direction. It’s simply where the doors are.
The argument worth having
Which brings us back to the club.
When a company this size arrives, index funds that track the big benchmarks are obliged to buy it and because so much money now sits in those funds, that buying is close to automatic. For a company floating less than 5% of its shares, that’s a lot of forced demand chasing very little stock.
So the index providers had a decision to make, and they didn’t all make the same one.
Nasdaq and FTSE Russell flexed. They adjusted their rules to let mega-cap newcomers in sooner, which means SpaceX could enter the Russell indices as early as this autumn’s reshuffle.
S&P Dow Jones held the line. On 4 June it announced it wouldn’t budge: companies still need a twelve-month trading history, a record of actual profit, and a meaningful free float before they join the S&P 500. SpaceX, loss-making and thinly floated, fails on two counts. Its earliest entry is now mid-2027.
Both positions are defensible, which is what makes it interesting.
The case for holding firm is that the rules exist for a reason. They were written to stop ordinary savers being funnelled, through their pensions and tracker funds, into untested, unprofitable, founder-controlled companies they never chose. “It speaks highly of the credibility of S&P Dow Jones to be rules-based,” as one strategist put it. Bend the rules for the biggest names and the rules stop meaning much.
The case for flexing is just as clean. An index is supposed to represent the market. Leave out the seventh-largest company in the country and you’re no longer holding up a mirror, you’re holding up an edited version. If the rules now exclude the very companies most investors would assume they own, perhaps it’s the rules that have aged, not the companies.
For now, an index built to hold America’s largest companies has decided it would rather not hold one of America’s largest companies. Make of that what you will.
We help clients across Hook and the surrounding counties think through exactly this sort of thing, not whether a headline is exciting, but whether it has any business being anywhere near their money. Usually the two questions have different answers.