Meta Earnings: Strong End to 2025 as AI Monetization Begins to Show

Key Morningstar Metrics for Meta

• Fair Value Estimate: $850
• Morningstar Rating: ★★★★
• Morningstar Economic Moat Rating: Wide
• Morningstar Uncertainty Rating: High

What We Thought of Meta’s Earnings

Meta META reported strong fourth-quarter earnings, with sales growing 24% to $60 billion. Operating margins contracted 700 basis points to 41% as AI-related costs continue to mount. The firm’s 2026 guidance calls for $125 billion in capital expenditures and $162 billion in operating expenses.

Why it matters: Meta’s ad business continues to perform at full capacity, with the firm’s investments in artificial intelligence continuing to drive metrics such as engagement, ad efficacy, and content recommendation.

• Engagement metrics, including time spent on Instagram and Facebook, remain strong, with video engagement particularly strong. As users spend more time on its platform, Meta can show them more ads, driving ad impressions up 18% for the quarter.
• We were similarly impressed by Meta leveraging its AI investments to drive growth in ad clicks (up 3.5% for Facebook) and conversions (1% increase for Instagram). While marginal, these improvements can unlock substantial topline dollars considering Meta’s scale.

The bottom line: We maintain our $850 fair value estimate for wide-moat Meta, with the firm’s outperformance on sales offset by its 2026 capital and operating expenses guidance coming in ahead of our estimates. Despite shares being up, we continue to view them as undervalued.

• We think that as the year continues, investors will align with our bullish view on Meta as more datapoints regarding the impact of the firm’s AI investments on its core ads business come to the fore.

Coming up: Meta expects sales for the first quarter to grow a whopping 30%, driven primarily by a healthy demand for its ad products. We expect this strength to continue into the remainder of the year, and model 2026 sales growth at 25%.

• We expect the launch of Meta’s latest large language model in the coming months, with the LLM likely to be competitive with frontier labs such as Google, OpenAI, and Anthropic.