"While many companies are looking for ways to profit from artificial intelligence (AI), Meta Platforms has flipped the narrative on its head. The company has long used sophisticated algorithms to surface relevant content and direct targeted advertising to users of its social media platforms. The advent of generative AI has taken that approach to the next level. When Meta reported its quarterly results after the market close on Wednesday, one thing was crystal clear."
"For the fourth quarter, Meta generated revenue that grew 24% year over year to $59.9 billion. This fueled diluted earnings per share (EPS) of $8.88, which increased 11%. For context, analysts' consensus estimates were calling for revenue of $58.47 billion and EPS of $8.22, so Meta sailed past Wall Street's expectations. The results were fueled by Meta's social media audience, with daily active users of 3.58 billion, up 7% year over year."
"Zuckerberg said the company plans to lean into its AI-driven success. In 2026, Meta is planning to spend between $115 billion and $135 billion on capital expenditures (capex), primarily on AI infrastructure. Meta's ability to scale down its Llama large language models (LLMs) to smaller AI systems for targeted advertising has been tremendously successful, boosting user engagement and making its adtech business even more profitable."
Meta generated $59.9 billion in fourth-quarter revenue, a 24% year-over-year increase, and delivered diluted EPS of $8.88, up 11%, surpassing analyst estimates. Daily active users reached 3.58 billion, up 7% year over year, expanding the audience for digital advertising. Ad impressions rose 18% year over year and average price per ad climbed 6%, driven by AI enhancements. The company plans $115–$135 billion in 2026 capital expenditures, primarily for AI infrastructure. Meta scaled down its Llama large language models into smaller systems for targeted advertising, improving engagement and adtech profitability. The approach generated strong ROI on AI spending.
Read at The Motley Fool
Unable to calculate read time
Collection
[
|
...
]