According to Similarweb data cited by TechCrunch on January 18, Meta’s text-based app Threads averaged 141.5 million daily active users on iOS and Android as of January 7, compared with 125 million daily mobile users on Elon Musk’s X. While X still commands roughly 150 million daily visits via web browsers, Threads has shown steady month-over-month growth on mobile since its 2023 launch, signaling that Meta’s user engagement initiatives—such as integrated cross-app notifications from Instagram and Facebook—are gaining traction among smartphone audiences.

In late 2025, CEO Mark Zuckerberg unveiled Meta Compute, an organization co-led by Santosh Janardhan and Daniel Gross (with presidential oversight by Dina Powell McCormick) tasked with centralizing data center design and strategic capacity planning. The initiative aims to build tens of gigawatts of compute capacity this decade and scale toward hundreds, levels comparable to those of small nations. Meta allocated $39 billion in capital expenditures in 2024 and guided that 2025 capex would reach approximately $71 billion, with CFO Susan Li projecting an even larger increase in 2026 versus 2025. These investments are designed to support in-house training and inference for models like Llama 4 and future generations, closing the gap with peers such as Google and Microsoft.

Meta reported $37.7 billion in net income on $141.1 billion of revenue in the first nine months of 2025, bolstered by an advertising business that generated $50 billion in Q3 alone—98% of total revenue—through tools like Advantage+ AI, which automates ad placement and optimization. Across Facebook, Instagram, WhatsApp, Messenger and Threads, the company reached 3.54 billion daily active users as of September 30, 2025, cementing its lead in global engagement. Management cited AI-driven features as key to increasing time spent per user and lowering costs for advertisers, suggesting that ongoing capex will yield improved ROI through higher ad load and elevated pricing power.

Investors will monitor whether Meta’s unprecedented compute build-out translates into differentiated AI products and sustainable monetization. Last year’s Llama 4 model received muted market response, prompting Meta to emphasize vertical integration of hardware, software and renewable energy procurement under Meta Compute. Leadership has set a goal of outpacing AI rivals by owning more of the end-to-end stack, from chip design to data center operations. Success will depend on accelerating model performance improvements, expanding API adoption among enterprise developers and translating advanced capabilities into higher ad impressions or subscription revenues over the next three to five years.