Meta Platforms shares jumped more than 6 percent on Wall Street on Wednesday after Bloomberg News reported the company is building a new cloud computing business. The plan: sell the excess artificial intelligence computing power it has been stockpiling to outside customers.

The move would turn a cost center into a revenue line. Meta has poured hundreds of billions of dollars into data centers and AI chips while chasing what CEO Mark Zuckerberg calls “superintelligence.” Wall Street has spent months asking how any of that spending comes back as profit. This is Meta’s answer.

A Direct Challenge to AWS, Azure and Google Cloud

The new business would put Meta in competition with Amazon Web Services, Microsoft Azure and Google Cloud, the three companies that currently dominate cloud infrastructure. Meta is reportedly also weighing whether to sell its own internally built AI models to business customers, not just raw computing capacity.

It’s a shift for a company that has spent two decades selling ads, not server time. But the infrastructure is already built. Renting out what it doesn’t need for its own AI work is a faster path to revenue than waiting for ad growth to catch up with the spending.

A Stock That Needed Good News

Meta shares had fallen nearly 15 percent this year as of Tuesday, underperforming the S&P 500 by a wide margin. The company told investors in April it plans to spend as much as $145 billion this year on capital expenditures, mostly data centers and the graphics processing units needed to train and run large AI models.

That level of spending has fed investor anxiety about return on investment. A cloud business that can generate outside revenue from the same infrastructure gives Meta a clearer story to tell about where the money goes.

Ripple Effects on Neocloud Rivals

Not everyone cheered. Shares of CoreWeave and Nebius, two “neocloud” companies that rent AI computing capacity to large customers including Meta, fell 10.8 percent and 12.4 percent respectively. Investors are weighing two risks at once: Meta becoming a competitor, and Meta needing less of their capacity now that it plans to monetize its own.

Neither risk is confirmed yet. Meta hasn’t detailed pricing, launch timing, or how much capacity it intends to sell.

Meta hasn’t said when the cloud business launches or how much of its AI capacity it plans to sell to outside customers.