Meta CEO Mark Zuckerberg is weighing plans to launch a commercial cloud computing business to monetize massive artificial intelligence data center capacity, even as soaring infrastructure costs pressure the company’s financial forecasts, according to regulatory filings and earnings statements. Among the four major U.S. hyperscalers, Meta remains unique in lacking a dedicated cloud services division despite capital expenditures that match its competitors.
Evaluating Cloud Infrastructure and Excess Compute Capacity
Meta is aggressively acquiring land and building massive AI data centers, creating a surplus of computing power in a resource-constrained market. According to CEO Mark Zuckerberg, speaking on the company’s second-quarter earnings call, Meta is receiving numerous offers to lease compute capacity at a significant premium over acquisition costs. Anthropic is currently in preliminary talks to lease computing power from Meta, as reported by CNBC.
Launching a cloud infrastructure business would allow Meta to monetize its capacity stockpile. However, Zuckerberg emphasized that the company must carefully balance short-term monetization against the development of future internal assets. “It would be foolish to basically just sell all of the compute and take a short-term profit,” Zuckerberg told investors, noting that Meta requires substantial compute capacity to fuel its own AI initiatives under AI chief Alexandr Wang.
Did you know? Meta is the only major U.S. hyperscaler without a commercial cloud infrastructure business, despite capital expenditure budgets that rival those of Alphabet, Microsoft, and Amazon.
Soaring Capital Expenditures and Financial Forecasts
Meta’s infrastructure expansion has significantly impacted its near-term financial metrics. During the second-quarter earnings report, Meta issued a weaker-than-expected revenue forecast for the third quarter. Free cash flow dwindled by 90% compared to the same period a year earlier, driven by surging capital expenditures.
In response to expanding AI demands, Meta raised the low end of its 2026 capital expenditure guidance by $5 billion, establishing a projected range between $130 billion and $145 billion. This spending mirrors broader industry trends among major tech firms. Alphabet recently hiked the top end of its 2026 guidance to $205 billion, pushing its cash flow negative for the first time, while Microsoft reported projected annual capex of roughly $175 billion.
The heavy spending weighed on investor sentiment. Meta stock sank more than 7% in after-hours trading following the earnings release, extending an annual slump that reached 11% by Wednesday’s close. Jefferies analyst Brent Thill told CNBC’s “Closing Bell Overtime” that investors are seeking clarity on Zuckerberg’s strategy in the compute market.
Building Enterprise Capabilities and Leadership
Entering the enterprise cloud market requires building new operational capabilities. Zuckerberg acknowledged that Meta has historically struggled in the enterprise sector and must develop a new organizational muscle to sell directly to businesses. To support this transition, former Amazon Web Services senior executive Dave Brown is set to join Meta, as CNBC recently confirmed.
Beyond raw compute leasing, Meta’s potential enterprise portfolio includes application programming interfaces (APIs), productivity services, and AI agents. Earlier this month, Meta debuted its Muse Spark 1.1 model. According to AI chief Alexandr Wang, the model represents the company’s strongest system for agentic and coding work yet, priced lower than rival offerings from OpenAI and Anthropic.
Wall Street remains cautious given Zuckerberg’s previous multi-billion-dollar bets, notably Reality Labs and the metaverse initiative launched in 2021. Reality Labs posted a $4.62 billion loss on $431 million of revenue during the latest period, while digital advertising continues to generate 98% of Meta’s total revenue. Despite these legacy costs, Zuckerberg remains committed to the current AI pivot, asserting that long-term investments in compute infrastructure will ultimately reward shareholders.
Frequently Asked Questions
Does Meta currently sell cloud computing services?
No. Among the four major U.S. hyperscalers, Meta is the only one that does not operate a commercial cloud infrastructure and services business, though CEO Mark Zuckerberg has recently floated the possibility of launching one.
What is Meta’s projected capital expenditure for 2026?
Meta bumped up the low end of its 2026 capital expenditure guidance by $5 billion, setting a projected range of $130 billion to $145 billion.
Which AI model did Meta recently release?
Meta debuted the Muse Spark 1.1 model, which AI chief Alexandr Wang described as the company’s strongest model for agentic and coding work yet.
How much did Meta’s Reality Labs lose in the latest financial period?
Reality Labs reported a loss of $4.62 billion on $431 million of revenue during the period.
Explore More: Stay informed on the latest developments in artificial intelligence infrastructure and big tech earnings by subscribing to our newsletter or exploring our continuous coverage of industry financial trends.
Worth a look