Meta’s quarterly results put AI growth and investment costs in the same frame

Meta’s latest quarterly results showed a company whose core advertising business is still expanding quickly while its push into artificial intelligence is becoming a larger part of the financial conversation. The company reported second-quarter revenue of $60.8 billion, up 28% from $47.52 billion a year earlier. Net income, however, fell to $15.85 billion from $18.34 billion, with legal expenses and severance costs weighing on the comparison. The contrast makes the quarter more than a simple growth story: investors are being asked to assess both the immediate benefit of AI in Meta’s established products and the cost of building for a longer-term opportunity.

Meta has described AI as a force that is improving its core business today, helping power the next generation of products and opening new enterprise opportunities. In practical terms, that proposition reaches the systems behind the company’s large consumer platforms: recommendations, advertising tools, ranking, creative workflows and the assistants that people encounter across its services. AI at that scale is not a separate line of business sitting next to Facebook or Instagram. It is increasingly part of the machinery that decides what users see, how advertisers reach audiences and how the company tests new formats.

The size of Meta’s audience remains central to that strategy. Daily active users across its family of apps — which includes Facebook, Messenger, Instagram, WhatsApp and Threads — reached 3.6 billion, a 3% increase from a year earlier. That reach gives the company a large operating environment in which changes to recommendation quality, ad relevance or creation tools can have visible commercial effects. It also means that even modest product improvements must work reliably across languages, regions, devices and very different user expectations.

The quarter highlights why AI spending at major consumer technology companies cannot be judged only by a single profit figure. Building models, data centers and supporting infrastructure can raise costs well before a new product has a direct revenue label attached to it. At the same time, AI may already be contributing indirectly through better engagement, more efficient advertising or lower-friction tools for businesses. The result is a financial picture in which investment and return are connected, but do not necessarily arrive in the same reporting period.

For Meta, the question is not simply whether AI can make an existing feed or advertising system more useful. It is whether the company can turn that capability into products that people and businesses choose to use repeatedly, while keeping the cost of serving those products under control. Consumer assistants, business tools and more automated creative systems each raise different questions about quality, safety, privacy and monetisation. A strong advertising quarter can fund those experiments, but it does not settle which of them will become durable businesses.

The lower profit figure also provides important context for the company’s AI narrative. One-time and legal costs can materially affect a period’s reported earnings without changing the direction of product investment, yet they still shape the resources available to management and the expectations of shareholders. Investors will be watching how Meta distinguishes temporary expenses from the recurring cost of its AI infrastructure, and whether revenue growth continues to support that spending as the company broadens its product ambitions.

The immediate takeaway from the results is that Meta’s AI strategy is being tested inside a business that is already large and profitable rather than in an isolated research project. Revenue growth shows the strength of the current platform, while the profit decline is a reminder that expansion carries real costs. Future quarters will show whether the company can demonstrate a clearer link between AI investment, product adoption and financial returns without reducing a complex transition to a single headline number.