• Meta is exploring a cloud business to sell excess AI compute capacity, CEO Mark Zuckerberg said, signaling a strategic pivot from purely internal infrastructure.
  • The company's 2026 capital expenditure guidance of $125-$145 billion has raised questions about utilization, with Zuckerberg noting a cloud offering is “definitely on the table.”
  • Analysts view the move as buying optionality, potentially turning a massive cost center into a revenue stream, but Meta would face entrenched hyperscalers AWS, Microsoft, and Google.

A Potential Pivot

Meta Platforms Inc. is signaling a potential entry into cloud computing, aiming to monetize the vast AI compute capacity it has built for internal use. Speaking at a conference this week, CEO Mark Zuckerberg said that if the company’s data-center buildout generates excess capacity, a cloud business “is definitely on the table.” The remarks, confirmed by a person familiar with the matter, mark the first time Meta has publicly entertained the idea.

The move comes as Meta’s capital expenditure guidance for 2026 surged to roughly $125–145 billion, a figure that has sparked debate among investors about whether the spending can be justified if internal AI demand doesn’t absorb it all. “We’re building for the long term,” Zuckerberg said. “If we end up with more compute than we need, we want to make sure we can put it to work.”

Carving a Niche

Meta would enter a cloud market dominated by Amazon Web Services (AMZN), Microsoft Azure (MSFT), and Google Cloud (GOOGL), but it could carve a niche by offering GPU-accelerated capacity optimized for AI workloads. Sources close to the company say Meta is exploring pay-as-you-go access to its custom chips and Nvidia (NVDA) GPUs, potentially bundled with low-latency inference services. “They have a massive infrastructure and deep expertise in running AI at scale,” said an analyst who requested anonymity. “If they can offer it cheaply enough, there’s demand from startups and enterprises.”

A cloud business would also diversify Meta’s revenue beyond advertising, which still accounts for over 98% of its income. However, questions remain about execution. Meta has tried hardware and API businesses before, with mixed results. The company declined to comment on specific plans or timelines.

Regulatory and Competitive Hurdles

Any move into cloud services is likely to attract antitrust scrutiny, given Meta’s size and market power. Regulators in both the US and Europe have already flagged concerns about data integration and market concentration. “Meta’s dominance in social media gives it a unique vantage point, but it also raises red flags when it moves into adjacent markets,” a competition lawyer warned.

Meanwhile, rivals are watching closely. AWS and Azure have spent years building enterprise relationships and compliance frameworks. Meta would need to invest heavily in sales teams and security certifications to win over corporate clients.

What’s Next

Meta isn’t rushing. The company is still in the early stages of assessing demand, and any formal offering is at least a year away, according to people familiar with the discussions. For now, Zuckerberg is buying optionality: if the AI boom continues, Meta’s excess capacity could become a lucrative asset. If not, it remains a sunk cost.

This article has been updated to reflect that Meta’s cloud business is still under discussion and no final decision has been made.