In the dot-com boom, the public shared the gains. In the AI boom, it does not.
He blames the high cost of going public and the fact that few Wall Street analysts, who research stocks for investors, follow smaller listings.
BIG IDEAS BY NEW ECONOMIES · Why Startup Valuations No Longer Make Sense | Eric HippeauEric Hippeau, a venture capitalist at the New York firm Lerer Hippeau and a former chief executive of HuffPost, was explaining why promising young companies stay private so much longer now. He has invested through several technology cycles, and he says the shift has changed who gets rich when a company succeeds.
“All the upside goes to the private investors, people like us,” Hippeau said.
What he meant was a reversal from the late 1990s. In the dot-com era, companies sold shares to the public early, after raising tens of millions of dollars rather than billions. Private backers could still do well by holding their shares, but most of the gains went to ordinary shareholders who bought in the initial public offering, the first sale of stock to the public.
Today it is the inverse. Artificial intelligence companies can raise billions while private, so by the time they might go public, the early growth has already happened behind closed doors. The public, he said, has little chance to share in the extraordinary wealth being built.
The reason, in his view, is partly cost. In the United States, being a public company now costs $12 million to $15 million a year in accounting, legal work, and reporting. For a business with $200 million in revenue and $40 million in earnings before interest, taxes, depreciation and amortization (EBITDA), a standard measure of operating profit, that bill alone cuts profit by a third and hurts the valuation.
The other problem is attention. A company that goes public worth $3 billion will struggle to get any Wall Street analyst to write about it. He wants banks to rebuild that research, and to revive a culture in which the public follows new offerings.
“You can buy $10,000 worth of shares,” he said. “You'll be able to participate, as you mentioned, in the incredible growth that we think is going to happen as a result of AI.”
Questions people ask
Why do AI startups need to raise billions before going public?
Training advanced AI models requires huge spending on specialized chips, data centers, and engineers, often long before revenue arrives. Private venture funds, sovereign funds, and large tech partners can supply that scale more quickly than a public offering. That lets companies stay private while they are still losing money.
Can an ordinary person invest in private startups?
Generally no, because U.S. rules limit most private startup shares to accredited investors, meaning people or institutions that meet high income or wealth thresholds. Some crowdfunding sites and secondary funds offer limited access, but stakes are small, fees can be high, and shares are hard to sell. For most people, the practical entry point is still the public market.
Why would a company want to go public at all?
A public listing lets early investors and employees sell shares, helps a company raise large sums for expansion, and can make it easier to hire and build trust with customers. The tradeoff is quarterly reporting, public scrutiny, and the $12 million to $15 million annual cost Hippeau cites. Many founders now delay that tradeoff as long as private money is available.