On January 12, Meta’s CEO announced the Meta Compute Initiative, committing to build tens of gigawatts of data-center capacity this decade, with a longer-term goal of hundreds of gigawatts. One gigawatt of data-center capacity can power roughly 750,000 homes, underscoring the scale of the plan. This infrastructure push is designed to secure Meta’s independence in AI development, data sovereignty and operational control. The company expects to spend in excess of $100 billion on capital expenditures in 2026, more than four times the $23 billion planned by its closest social-media competitor, strengthening Meta’s position in content recommendation and targeting as generative AI becomes embedded in its advertising platform.

Meta’s rapid expansion of data centers has coincided with a sharp rise in research and development and capital spending. In the third quarter of 2025, sales rose 26% year-over-year to $51.2 billion, while adjusted earnings per share increased 20%, reflecting strong operating leverage. Management reports that every dollar invested in AI and infrastructure translates into higher average revenue per user thanks to more relevant ads and improved algorithmic efficiency. Despite a recent one-time charge, the company maintains operating margins above 40%, illustrating its ability to scale fixed-cost investments across its 3.5 billion daily active users.

Meta’s share price has climbed over 500% in the past decade, yet it remains the only major U.S. technology firm in its peer group to have never conducted a stock split. With shares trading above $600, the company appears poised for a split that could enhance liquidity and broaden retail participation. Research from Bank of America indicates that firms announcing splits tend to outperform the market by 13 percentage points on average in the subsequent 12 months. Meta currently trades at less than 28 times forward earnings, below both its historic multiple and the broader market, suggesting attractive valuation support ahead of any corporate action.

Investors should be aware of three material risks as Meta accelerates its AI investments. First, the lack of a direct cloud-service revenue stream means returns on infrastructure spending will be realized indirectly through advertising metrics, potentially extending the payback period. Second, free cash flow declined from $54 billion in late 2025 to approximately $45 billion most recently, driven by rising capital expenditures that are forecast to increase further in 2026. Finally, investor sentiment remains sensitive to guidance around future spending; a 11% share-price drop followed the Q3 earnings announcement when management signaled a notable increase in next year’s capital requirements.