An IPO Is a New Car
It depreciates the second it leaves the lot, and three tells say the sticker is a premium
Think about the last time you bought a new car.
The moment you drove off the lot, the car was worth less than what you just paid, sometimes thousands less, and nothing about the car had really changed. Same engine, same paint, same new car smell. What changed was that the price stopped being a number the dealer set and started being a number the broader market would actually pay. The sticker was a premium.
An IPO works the same way.
The people selling you shares in a company going public set the price themselves, the same way a dealer sets a sticker price. An initial public offering, the first time a company sells stock to the public, is in the usual case priced by the company and its bankers before a single share has ever traded in the open. There is no public price yet to anchor to. So the number on the windshield is the seller’s number, built to clear the lot at a premium, and the people who built the company know more about what is really under the hood than you do. That last part is one reason the premium exists, but it is not the whole story. The bigger story is what happens after you drive off.
This is worth thinking about right now because the lot is filling up. SpaceX has filed to go public and is reportedly days from its debut,[1] while OpenAI and Anthropic are earlier but moving in the same direction. The investors who missed the private rounds are being told this is finally their chance to get in at the “start”. So ask a plain question. When you buy an IPO on day one, what are you actually buying?
You are buying the sticker, not the discovered price.
The depreciation is the whole point
The price of a newly public company gets discovered over the first few years it trades, not in the first hour. The mechanism is slow and it is grinding. Lockups expire, meaning the stretch after listing when insiders are barred from selling ends and more shares hit the market. Earnings reports arrive and test the story the price was built on. The float, the slice of shares that actually trades freely, grows as more holders are allowed to sell. By year three or four, the market has driven the car long enough to know what it is.
And on average, that discovery process has not been kind. Across decades of data, newly public companies have lagged the broad market over the three to five years after they list, measured from the first day’s close, the price you can actually get.[2] Tim Loughran and Jay Ritter called it the new issues puzzle. It is not a law, and it does not hold for every cohort. The drag falls hardest on the young, unprofitable companies that go public in hot years, which is a fair description of most of what is about to come public. The new car depreciates. That is not a metaphor I am stretching. It is what the long-run numbers describe.
So the smart move is not to buy the car the day it rolls off the line at a price the seller wrote. The smart move is to wait. Let the company trade in public for a few years. Let a couple of earnings reports land. Let a lockup or two expire so you can watch whether the insiders keep selling. Then buy the seasoned stock at the price the market discovered, not the price the seller set. You give up a first-day pop you probably were not going to be allocated anyway, and in exchange you get the one thing the lot never hands you up front. A price somebody other than the seller agreed to.
Three tells the sticker is a premium
You cannot test-drive a private company. But you can read three tells, and each one says the same thing. The sticker is a premium, and the depreciation is coming.
First, watch who is selling. If the founders and the early funds are handing you shares, they have made a quiet judgment. They think those shares are worth more to you than to them. That does not make them crooks. They are diversifying, paying a tax bill, buying a house. But the direction of the trade is information. They built the car. They are cashing the premium. You are paying it.
Second, watch who is being invited to buy. When an offering steers a bigger slice toward retail investors, the premium is being distributed to the least informed buyers in the room, so it is fair to ask why the sophisticated money passed. Institutions see the deal before you do. If they fill their boots, retail gets the scraps. If retail is getting a generous helping, sometimes that is a company choosing to be democratic, and sometimes it is the tell that the people who price these things for a living already had their look and walked. You usually cannot know which from the outside. You can know that a deal sold hard to the crowd is a different animal from one the crowd had to fight its way into.
Third, watch the duration. A company’s duration, in the loose sense, is how far into the future its cash flows sit, and the term is borrowed loosely here, not the technical bond measure. Some businesses going public earn real profits today. Others are priced almost entirely on what they might earn a decade out. The further out the money is, the more of today’s price is an unverified forecast, and the more premium there is to come off when the forecast meets reality. OpenAI is the clean example. In the first half of 2025 it reportedly burned around two and a half billion dollars and was on track to burn far more for the year, so almost all of what you would buy is a story about cash that has not arrived yet.[3] SpaceX is the other kind. It already earns real money, most of it now from Starlink, so the question there is not whether the cash exists but whether it can grow into the price being floated.[1:1] Either way, the wider the gap between today’s price and today’s cash, the bigger the part of the sticker that is just a promise.
The pop you were promised is not the pop you get
The popular story is that IPOs pop. You get in, the stock jumps on the first day, you are up before lunch. And new listings do often jump on day one. The trouble is who captures it. That first-day move goes to whoever was allocated shares at the offer price, which is mostly institutions and the underwriter’s favored clients, not the person buying once it opens to everyone. By the time you can click buy, the pop has usually already happened to someone else. What you get is the marked-up street price the morning after.
Of course, none of this means every IPO is a trap. Some of the best companies of the last thirty years were buyable near their first public price, and waiting cost real money. But that is survivorship. You remember the winners and forget the hundreds that quietly sank, and the median experience is built from both. If a business is genuinely early in a long run of compounding, the short term depreciation after IPO is a rounding error against where it ends up. So the honest version of the rule is not “never buy an IPO.” It is that the day-one price is the seller’s price, the depreciation tends to come, and the seasoned stock a few years later is the one that has been driven long enough to reveal what it actually is. Buy that one.
So when SpaceX or OpenAI or Anthropic finally rings the bell, and the headlines tell you the future just went on sale, picture the lot. The car is brand new, the number on the windshield is the seller’s, and the road has not touched it yet. Let it drive a few years. Then go look at what it is worth.
As always, none of this is financial advice.
SpaceX filed its IPO prospectus on May 20, 2026, planning to list on the Nasdaq under the ticker SPCX, and the filing disclosed about $18.7 billion in 2025 revenue, with Starlink contributing roughly $11.4 billion. See “SpaceX files S-1: IPO could make Elon Musk a trillionaire”, NBC News, May 20, 2026, https://www.nbcnews.com/tech/elon-musk/spacex-files-s-1-ipo-make-elon-musk-trillionaire-rcna346157; and “6 Charts on SpaceX’s Pre-IPO Financials”, Morningstar, May 2026, https://www.morningstar.com/stocks/6-charts-spacexs-s-1-financials.
Jay R. Ritter, “The Long-Run Performance of Initial Public Offerings,” The Journal of Finance 46, no. 1 (1991), 3 to 27; Tim Loughran and Jay R. Ritter, “The New Issues Puzzle,” The Journal of Finance 50, no. 1 (1995), 23 to 52; and Ritter’s regularly updated figures, “Initial Public Offerings: Updated Long-run Statistics,” University of Florida (Warrington College of Business), updated Mar. 23, 2026, https://site.warrington.ufl.edu/ritter/files/IPOs-long-run-returns-on-IPOs.pdf.
OpenAI reportedly generated about $4.3 billion in revenue and burned about $2.5 billion in cash in the first half of 2025, with a full-year cash-burn target near 8.5billion,perTheInformation.See “OpenAI’s First Half Results: $4.3 Billion in Sales, $2.5 Billion Cash Burn”, The Information, Sept. 2025, https://www.theinformation.com/articles/openais-first-half-results-4-3-billion-sales-2-5-billion-cash-burn; reported by Reuters, Sept. 29, 2025, https://finance.yahoo.com/news/openais-first-half-revenue-rises-043012507.html.