TIM COHEN: The IPOs that could rewire capitalism

Before the end of the year, or possibly by early next year, three rather large IPOs are going to hit planet Earth: SpaceX, OpenAI and Anthropic. Journalists are renowned for being breathless, but this is all kind of mind-boggling. And it’s not just the size — though that is gobsmacking enough, probably around $3-trillion in collective market cap. But the broader significance is also shape-shifting.

Just on the scale issue, one indicator is the gaudy question of Elon Musk’s net worth. At the beginning of 2024, the Forbes billionaires list tallied his total wealth at $251.3bn. There were actually other people who were close to his wealth at the time. After the SpaceX IPO, Musk will probably be worth about $800bn … and there will be nobody even close to him. Jeff Bezos, you know, impressive … but sorry. It’s Musk’s world; we just live in it.

Here is another indicator: there are other exciting IPOs scheduled for this year in China and India. In India, telecoms company Reliance Jio is preparing for a Mumbai listing that could be India’s biggest-ever IPO, and early indications suggest a market cap of about $180bn. The largest scheduled Chinese IPO is the Singapore-based logistics and digital-infrastructure group GLP, which is targeting a valuation of about $20bn. These are significant IPOs, but even collectively they will not be worth as much as the third-largest prospective US IPO, that of Anthropic, the owner of AI product Claude.

There is a third and less heartening statistic: If SpaceX, OpenAI and Anthropic were to IPO in 2026, one effect would be a dramatic spike in the Buffett Indicator, the ratio of total market capitalisation to GDP. Currently, the US stock market is already “strongly overvalued” by this measure, with a Buffett Indicator of about 230% ($72.1-trillion market cap/$31.3-trillion GDP). Adding these three “super-IPOs” would push this ratio into uncharted territory.

It’s worth noting that in 2000, the market was considered wildly speculative when the ratio hit roughly 140%! These IPOs will take the index to its highest in history, at around 240% by the highest measure. But there is a counterargument that the index doesn’t work anymore because these are international companies, and comparing them to just the US economy might have been relevant once but is no longer. Still.

Amazingly, for all the anger and bile in the world, the IPO market is already booming this year. Reuters reports equity capital markets issuance rose 40% year on year to $211bn in the first quarter, while IPO proceeds rose 47% to $44bn. But Iran war volatility is still hanging around like an underdressed guest at a wedding. That is why these deals matter so much: they are not just listings; they are stress tests for market depth, appetite and, as always, that funny thing called collective delusion.

SpaceX looks the most likely to list this year. Reuters reports it has confidentially filed for an IPO and is discussing a valuation first around $1.75-trillion and then above $2-trillion, with a possible fundraise of as much as $75bn. At those levels it could become the largest IPO ever.

The bigger shift is that public investors are going to be asked to finance businesses that no longer fit old categories

The reason SpaceX is the easiest of the three to take seriously is that underneath Musk’s usual fireworks factory there is a very real business. Reuters has reported about $8bn in profit on $15bn-$16bn of revenue last year, with Starlink providing roughly 50%-80% of revenue and serving more than 9-million users. In plain English: unlike many glamorous tech stories, this one already has cash flow, scale and market dominance. It is still expensive, obviously. But at least it is expensive in the old-fashioned way, with money attached.

OpenAI is the most culturally significant listing but financially the most conceptually awkward. Reuters reports it has just raised $122bn at an $852bn valuation. OpenAI also reportedly hit more than $25bn in annualised revenue by the end of February, after generating $13bn of revenue in 2025 while spending $8bn. That sounds marvellous until one gets to the part where Reuters also reports OpenAI is targeting roughly $600bn of compute spend through 2030, driving adjusted gross margin down to 33% from 40% in 2024. This is what makes OpenAI such a strange public-market proposition: it is both a software company and a very expensive argument with physics.

Anthropic is the interesting third child in this family portrait: quieter, more enterprise-orientated, more safety-branded and, therefore, in some ways more plausible to sober fund managers. Reuters reports it raised $30bn in February at a $380bn valuation and said its run-rate revenue was $14bn at that point, with Claude Code alone above $2.5bn. Anthropic is a public benefit corporation too, with governance shaped partly by a long-term benefit trust, which is either admirable institutional foresight or an early warning that public shareholders may not be the only people in the room.

So how do these deals change the IPO game? First, they accelerate the migration of giant private companies into public benchmarks almost immediately. Nasdaq has changed its rules so newly listed mega-caps can enter the Nasdaq-100 much faster, explicitly because companies now stay private longer and arrive at the public market already enormous. Second, they force public investors to accept governance structures that look more like constitutional experiments than normal corporations. Third, they risk crowding out other issuers.

But the bigger shift is that public investors are going to be asked to finance businesses that no longer fit old categories. They are part software, part utility, part geopolitical instrument. And part — I don’t know how to describe it — a mirrored tool of human thought? A statistical simulator of intelligence?

How much is that worth? We are about to find out.