In 1720, the most brilliant mind in Europe lost a fortune chasing a hot stock. Isaac Newton — the man who could chart the motion of the planets — bought shares in the South Sea Company, watched them climb, sold at a tidy profit, and then, as the mania crested, bought back in near the top. When the bubble collapsed he was ruined badly enough that, the story goes, he forbade anyone to speak the words «South Sea» in his presence. His verdict on the episode has outlasted his loss: he could calculate the motion of heavenly bodies, but not the madness of people.
That madness is about to be rebuilt at industrial scale. SpaceX filed this month to go public. OpenAI is preparing its own offering, and Anthropic is reportedly targeting a listing in the fall. Together, these three private companies could raise close to $200 billion between them — more than every U.S. company that went public from 2022 through 2025 raised combined — at a combined valuation approaching $4 trillion. Wall Street is calling this democratization, a chance for the ordinary investor to finally own the AI revolution. Before the applause, it is worth asking what, exactly, we are being invited to.
The door opens when the room empties
Here is the part the celebration leaves out. For a decade, the companies that defined the artificial-intelligence era were closed to ordinary investors by design. You could not buy OpenAI. You could not buy SpaceX. The explosive years — when a few hundred million dollars of valuation became a few hundred billion — happened entirely behind a private wall, accessible only to venture funds, sovereign-wealth vehicles and a thin circle of insiders.
A company stays private for exactly as long as private money is patient and cheap. It goes public when the people already inside need something specific: liquidity. They need to sell. The industry does not even hide this — the technical term for an initial public offering, in the language of venture capital, is an «exit». An IPO is not the beginning of the wealth. It is the moment private wealth is converted, at a price the seller chooses, into public risk.
So when SpaceX — private since 2002 — and OpenAI — backed by insiders since the last decade — suddenly decide that now is the time to let everyone in, the civic-minded investor should not feel flattered. He should feel handed the check. The hundredfold returns already happened. They happened in rooms he was never allowed to enter. What is on offer now is the part that comes after.
The saver who owns the past
We have been told a comforting story about the public market: that it is where a nation’s growth lives, and that a 401(k) tracking the S&P 500 makes every worker a small owner of the future. That story is fraying. Increasingly, the public market is where mature companies go to be sold, while the genuine wealth creation happens upstream, in private, where the saver cannot follow.
The numbers are blunt. The ten largest companies in the S&P 500 now account for close to 40% of the entire index’s value, and AI-related firms have absorbed nearly 87% of all venture-capital funding. The index the average American already owns is, in other words, a concentrated bet on AI’s past winners. The IPO wave now asks that same American to also fund its uncertain future — at valuations of one, two, even four trillion dollars, for businesses not one of which has yet reported a full year of profit.
The other side of the trade
To be fair, not every public offering is a trap, and the alternative — a permanent private aristocracy that owns every important company and never lets the public in at all — would be worse. Some companies still go public with decades of growth ahead of them, and reward ordinary shareholders handsomely. And it is not as if the public investor has been starved of artificial intelligence: the market is already saturated with it. If you want exposure, the AI stocks Wall Street analysts already favor are one click away, no IPO lottery required.
The objection, then, is not to AI, and not even to IPOs. It is to price and timing. There is a difference between owning a revolution and being sold one at the top.
And the warning signs are not subtle. The Securities and Exchange Commission is already examining whether OpenAI’s earlier investors were given an honest picture of what they were buying. When even the private buyers — the supposedly sophisticated ones — are litigating what they were told, the retail buyer should read the prospectus twice, and then once more.
The oldest rule on Wall Street is also the simplest, and it is roughly what Warren Buffett has spent a career telling his shareholders: be fearful when others are greedy. When the people who understand an asset best become suddenly, urgently eager to sell it to the people who understand it least, the transaction has a name. That name is not «opportunity».
Newton’s mistake in 1720 was not that he bought the South Sea Company. It was that he bought it back — late, high and convinced this time was different. The AI revolution may well be real; that was never the question. The question is whether you want to own it at the price its current owners have decided, after years of keeping the door shut, is finally the right moment to open it. When the smart money is this eager to share, the gift to examine most closely is the one being handed to you.
