Meta Raises 2026 AI Spending Outlook to $130-145 Billion as Q2 Profit Falls 14%
Key Takeaways
- •Meta's second-quarter revenue rose 28% year-on-year to $60.8 billion, driven by strong digital advertising demand and AI-powered improvements to its ad platform.
- •Net income fell 14% compared with the prior-year quarter due to higher legal costs, severance charges, and substantial artificial intelligence investments.
- •Meta raised its 2026 capital expenditure guidance to a range of $130-$145 billion, positioning it among the largest AI infrastructure spenders alongside Microsoft, Alphabet, and Amazon.
- •The company's family of apps reached approximately 3.6 billion daily active users, with Instagram surpassing 2 billion daily users and Threads growing to around 500 million monthly active users.
- •Investors sent Meta shares lower in after-hours trading, seeking clearer evidence that the company's outsized AI spending will translate into sustainable earnings growth.

Meta Raises 2026 AI Spending Outlook to $130-145 Billion as Q2 Profit Falls 14%
Meta Platforms reported a robust second quarter with revenue climbing 28% year-on-year to $60.8 billion, propelled by strong digital advertising demand and AI-driven enhancements to its ad platform. Despite the revenue surge, the results failed to reassure investors. Shares of the Nasdaq-listed company, trading under the ticker META, dropped sharply in after-hours trading as net income declined and spending on artificial intelligence infrastructure continued to escalate. (Sources: BBC News, CNBC)
Profit Declines Amid Rising Costs
Net income fell 14% compared with the same period a year earlier, as Meta absorbed higher legal costs, severance expenses, and substantial investments in artificial intelligence. The company also revised its 2026 capital expenditure guidance upward to a range of $130-$145 billion, reinforcing CEO Mark Zuckerberg's determination to expand AI infrastructure capacity even as near-term earnings and cash flow take a hit. The revised figure represents a significant step-up from prior guidance and positions Meta alongside Microsoft, Alphabet, and Amazon as the largest spenders on AI infrastructure among U.S. technology giants, with much of the outlay directed toward data centers, specialized semiconductors, and computing capacity needed to train and deploy large-scale AI models.
Free Cash Flow Under Pressure
Heavy AI expenditures markedly reduced Meta's free cash flow, underscoring the mounting financial burden of competing in the global artificial intelligence race. Digital advertising remains Meta's dominant revenue generator, but investors are increasingly focused on the timeline for when the company's AI investments will begin producing meaningful financial returns. Meta's spending encompasses its open-source Llama family of large language models, which the company distributes freely as part of a strategy to build ecosystem adoption and differentiate its approach from closed-model competitors such as OpenAI's GPT and Google's Gemini.
User Base Continues to Expand
Meta's family of apps sustained its upward user trajectory, reaching approximately 3.6 billion daily active people. Instagram surpassed 2 billion daily users, while Threads grew to roughly 500 million monthly active users. This expanding user base provides Meta with a larger audience across its platforms for monetization through advertising and AI-powered services.
Investors Await AI Monetization
Although Meta's advertising business has shown resilience, investors are calling for clearer evidence that the company's outsized AI investments will translate into sustainable earnings growth. With capital expenditure expected to remain elevated, subsequent quarterly results are likely to be evaluated on AI monetization progress as much as on core advertising performance. The intensified spending cycle also draws attention to the broader semiconductor supply chain, particularly demand for high-performance GPUs from suppliers such as Nvidia, which has become a critical input for Meta and its peers as they scale model training and inference workloads.
Source: Economic Times Markets