NEW YORK — Investors are preparing to judge one of the clearest financial tests yet of the artificial-intelligence infrastructure boom as SB Energy moves toward a U.S. initial public offering.
The SoftBank-backed data-center developer filed paperwork for a Nasdaq listing as demand for AI computing infrastructure continues to attract enormous amounts of capital.
A New Kind of AI Stock
SB Energy is not primarily a software company.
Its business is building the physical infrastructure required to operate large computing facilities.
That makes the company a useful indicator of how Wall Street values the less visible side of the AI economy.
The company reported first-half revenue of $138.7 million, up sharply from a year earlier.
But it also reported a multibillion-dollar net loss.
That contrast is central to the IPO story.
Investors are being asked to finance a company with an enormous development pipeline while accepting substantial current losses.
SB Energy says it has billions of dollars worth of projects under development and is building multiple gigawatts of capacity.
The company's customers and partners include major technology companies.
That provides credibility.
But long-term leases and project commitments do not automatically guarantee profitability.
Construction costs, financing expenses, electricity availability and customer concentration can all affect returns.
The IPO also arrives after other AI infrastructure companies have received strong investor attention.
CoreWeave's market performance has demonstrated that Wall Street is willing to assign large valuations to companies serving AI computing demand.
But the market is becoming more selective.
Investors want evidence that infrastructure companies can generate cash rather than simply accumulate projects.
SB Energy's offering will therefore be watched beyond the company itself.
A successful IPO could encourage more capital to flow into data centers, power infrastructure and specialized computing facilities.
A weak reception could signal that investors are beginning to question whether AI infrastructure spending has moved too far ahead of actual earnings.
