Meta Is Reportedly Building a Cloud Business to Rival Amazon and Microsoft -- and a Way to Make Its Massive AI Bet Pay Off
Meta is reportedly developing a cloud business, 'Meta Compute,' to rival Amazon and Microsoft by selling AI computing power. This move aims to monetize massive capital expenditures on data centers. Following the Bloomberg report, Meta's stock jumped 8.8%. While the plan remains unconfirmed and faces execution challenges, analysts view it as a potential revenue stream alongside Meta's accelerating core advertising growth.
For months, the knock on Meta Platforms (META 4.80%) hasn't been its business. It has been the bill. When management raised its 2026 capital expenditures guidance in April to a range of $125 billion to $145 billion, shares sank on the news. And heading into Wednesday, the stock was down nearly 15% for the year, sitting well below its 52-week high of $796.25 -- even as the company reported accelerating growth.
Then investors got a look at what could become the other side of that spending story. Bloomberg reported Wednesday that Meta is developing plans for a cloud business that would sell access to artificial intelligence (AI) computing power and models, putting the social media company in competition with the cloud units of Amazon and Microsoft. Shares jumped 8.8% to $612.91.
The reaction makes sense. But can renting out computing capacity actually change the return math on one of the biggest capital spending programs in corporate history?
Image source: Getty Images.
What Meta is reportedly planning
According to Bloomberg, the effort is internally called Meta Compute, and the company is debating two approaches: giving developers access to AI models hosted on Meta's infrastructure, or selling raw computing power. The report said the plans are still in development and could change. And it's worth emphasizing that Meta hasn't announced anything.
Still, the idea isn't coming out of nowhere. CEO Mark Zuckerberg said in May that selling excess computing capacity was "definitely on the table" if Meta ends up building more data center capacity than it needs, according to the report.
Overbuilding is precisely the worry that has weighed on the stock. Meta's spending plan for 2026 -- raised in April from a prior range of $115 billion to $135 billion -- compares to $72.2 billion in capital expenditures in 2025. In other words, spending could double this year.
And unlike Amazon, Microsoft, and Alphabet, Meta has no cloud computing business renting its infrastructure to outside customers. Every dollar of return on those data centers has to come from Meta's own products, mainly advertising. If the company builds more capacity than its apps and AI ambitions need, the excess earns nothing. A cloud business would change that equation, turning idle capacity into revenue -- and giving Meta a source of sales beyond the ad market.
The spending may already be paying off
What's easy to miss in the debate over Meta's spending is that the core business is accelerating -- a sign its AI investments are already generating returns. First-quarter revenue rose 33% year over year to $56.3 billion -- a step up from 24% growth in the fourth quarter of 2025 and 22% growth for the full year.
In addition, the company is making progress on its efforts to build a superintelligence.
"We're on track to deliver personal superintelligence to billions of people," said Zuckerberg in the company's first-quarter earnings release.
None of this makes the reported cloud plan a sure thing -- or even a near-term one. Building an enterprise cloud business requires sales teams, support operations, and reliability commitments that Meta would be starting mostly from scratch, while Amazon and Microsoft have spent nearly two decades building exactly that. And selling raw computing capacity tends to carry lower margins than Meta's advertising business. So even if the reported plans turn into a product, it could take years for the revenue to matter.
NASDAQ: META
Key Data Points
In the meantime, the stock's valuation looks undemanding. After Wednesday's jump, Meta trades at about 21 times earnings -- a multiple that arguably still reflects the market's skepticism about all that spending rather than the company's growth.
To me, the stock is a buy -- just not because of Wednesday's report. An unconfirmed plan shouldn't be anyone's investment thesis. The better reason is the combination of an accelerating core business and a modest valuation. The reported cloud business is best viewed as a free option: If it launches, Meta gains a second way for its AI infrastructure to pay off. If it never does, buyers today still own one of the market's fastest-growing big tech companies at a reasonable price.
There are risks. Spending could climb even higher, and advertising demand can turn quickly in a weak economy. But among big tech's AI spenders, Meta now offers a rare combination: accelerating growth, a modest multiple, and if the report proves right, a new way to get paid for all those data centers.
