OpenAI filed a confidential registration statement on June 8 ahead of a public offering, following a private valuation of $852 billion in March, with a projected $14 billion loss for the year and profitability not expected before 2029 or 2030.
The three figures together describe the bet with unusual clarity. A high valuation, a large loss and distant profit only hold together if the market accepts that the position won now is worth more than the next few years of results.
The loss has a familiar source that's hard to reduce in the short term: the cost of compute to train and serve models, in a period when memory got expensive and cloud capacity grew scarce.
Going public under those conditions changes the nature of the company in ways that reach customers. A public company answers to shareholders quarter by quarter, which tends to pressure pricing, cost cutting and locking customers into longer contracts.
For anyone depending on the supplier, the practical reading isn't about the stock but about predictability: prices from a company that needs to show a path to profit tend to rise rather than fall.