Meta's stock rose nearly 9% on a report that it may have overbuilt
Note: This image contains AI-generated elements.
Meta's stock rose nearly 9% on a report that it may have overbuilt.
On July 1, Bloomberg reported that Meta is planning a cloud business, internally called Meta Compute. It would sell AI computing power and access to hosted models. Shares closed up about 9%. Two specialist AI-cloud providers, CoreWeave and Nebius, fell 10.8% and 12.4% the same day.
Nine weeks earlier, the same spending plan looked reckless. On April 29, Meta raised its 2026 capital-spending guidance to $125 billion to $145 billion, up from $115 billion to $135 billion. The stock fell roughly 8.6% the next day.
Nothing about the spending changed. What changed is that Meta named a use for capacity it may not need.
Zuckerberg was careful about this in May. Asked at the shareholder meeting about selling spare capacity, he called it "definitely on the table" if Meta reached a point where it had overbuilt. He said companies approach Meta almost every week. Meta had not sold access, he added, because it still expected to use the compute itself.
That is a contingency. The market is pricing it as a strategy.
At Eagle Talon, we ask whether the capital was committed against a plan, or the plan arrived after the capital. Meta has not said which market it is entering, and that choice decides the answer.
Selling access to hosted models puts Meta against Amazon's cloud unit and Google Cloud. Those two generated $128.7 billion and roughly $58.7 billion in 2025, on a decade of enterprise relationships Meta does not have. Selling raw computing cycles puts it against CoreWeave and Nebius, where price is most of the pitch.
Either way, Meta Compute would grow at cloud margins rather than advertising margins.
If Meta's first customer buys raw cycles rather than hosted models, is that a business Zuckerberg built, or a buyer he needed?
🔗 Source: Meta Pops 9% as Company Makes Cloud Push to Sell Excess AI Compute Power Capacity
