How Meta makes its money
Meta Platforms runs Facebook, Instagram, WhatsApp and Messenger. Almost all of its revenue comes from advertising in these apps. In the second quarter of 2026, advertising revenue was $59.4 billion out of total revenue of $60.8 billion. That is around 98%.
The group reports in two segments. Family of Apps bundles the apps and the advertising business. Reality Labs develops smart glasses, VR headsets and the related software. This second segment is small and deep in the red. In the second quarter of 2026, it posted $431 million in revenue against a loss of $4.62 billion.
The foundation of the advertising business is reach. In June 2026, an average of 3.6 billion people a day used at least one of the group’s apps. That means more than four in ten people worldwide are part of the potential audience every day. Recently, both price levers grew at the same time. The number of ads delivered rose by 14%, and the average price per ad by 12%.
Why I know Meta from two sides
I run ads on Facebook and Instagram for my own company all the time. So for me, Meta is more than a stock. It is also a tool in my everyday work. This dual perspective helps me put the figures into context.
As an advertiser, I see how strongly Meta is automating its ad system. The platform selects target audiences, manages bids and increasingly creates images and text as well. By the end of 2026, Meta wants to have campaigns created almost entirely by AI. Advertisers will then mainly set the budget and the goal. If this improves results, customers become willing to pay more. That is exactly what the rising ad prices show.
As an investor, I still stay sober. A strong product alone does not make a good stock. For the valuation, what counts is profit development, capital requirements and risks.
How the business has performed recently
In the 2025 financial year, Meta generated revenue of $200.97 billion, an increase of 22%. Already in that year, $72.22 billion went into capital spending, mainly on data centers.
In the second quarter of 2026, revenue grew by 28% to $60.8 billion. Net profit nevertheless fell by 14% to $15.85 billion. The explanation lies on the cost side. Expenses rose by 55% to $42 billion. This includes $2.4 billion for legal proceedings and $1.18 billion in severance payments after job cuts in May 2026.
The stock market reacted to this mix with a significant drop in the share price (in German), because earnings per share missed expectations. So the revenue engine is running, but costs are currently determining the result. For the third quarter of 2026, Meta expects revenue of $61 to $64 billion.
AI determines the investments
Meta uses artificial intelligence in two places. The first is the advertising system. Models decide who sees which ad, raising hit rates and prices. The second is its own model family, Llama. Meta releases many of these models openly. They also power the assistant Meta AI in WhatsApp and Instagram.
Huge data centers are being built for this. CEO Mark Zuckerberg announced plans to invest several hundred billion dollars in computing power. A dedicated unit called Meta Superintelligence Labs is meant to develop models that match or exceed human capabilities in many areas. The first large cluster, Prometheus, is due to come online in 2026. The Hyperion site in Louisiana is set to grow to up to 5 gigawatts of capacity over several years.
This costs a great deal of real money. For 2026, Meta plans capital spending of $130 to $145 billion. In the second quarter alone, $31.08 billion went into property and equipment. As a result, free cash flow shrank to $784 million. That is a decline of more than 90% within one year.
The dividend is a side issue
Meta has paid a dividend since 2024 (in German). It currently stands at $0.525 per share per quarter, or $2.10 a year. Measured against the share price, that gives a yield well below one percent (in German). Meta did not buy back any of its own shares in the second quarter of 2026. This also shows you the priority: for now, the money goes into data centers first.
Three questions before a possible purchase
The first question concerns the business. Can advertising revenue carry the AI spending in the long run? To answer this, I look at ad volume, ad prices and the profit margin over several quarters. A single record quarter says little. It gets interesting when costs grow faster than revenue, as they did recently.
The second question concerns the trend. Is the price trading above its long-term moving averages, and are these lines rising? This is not a signal to act. But it shows you whether you would be acting against the prevailing market direction.
The third question concerns momentum. After strong moves, the Relative Strength Index helps with the assessment. An extreme value alone decides nothing. Together with the business figures and the trend, however, you can see whether the situation is currently overheated or balanced.
The main risks
Advertising depends on the economy. Around 98% of revenue comes from ads. In a downturn, many companies cut their advertising budgets first. Such a cycle hits Meta more directly than more broadly diversified groups.
The EU is intervening in the business model. Because of its “pay or consent” model, the European Commission imposed a fine of €200 million under the Digital Markets Act in April 2025. Since January 2026, Meta has offered an option with less personalized advertising in the EU. Less use of data could push down ad prices in Europe.
Reality Labs costs billions. The glasses and VR segment lost $4.62 billion in the second quarter of 2026 alone. Whether and when this bet pays off remains open.
Youth protection is becoming a legal risk. In the US, consolidated lawsuits from families and school districts accuse the apps of an addictive design. The $2.4 billion charge for legal proceedings shows the scale. Since December 2025, Australia has banned users under 16 from social networks (in German). Other countries are discussing similar rules.
The AI bet needs returns. Capital spending of $130 to $145 billion in a single year is only sustainable if it generates additional profit. If it does not, depreciation and operating costs will weigh on earnings for many years.
My conclusion on Meta stock
Meta combines a very profitable advertising business with one of the largest investment programs in corporate history. The advertising side is growing at a double-digit rate and funds the transformation. At the same time, legal costs, Reality Labs and the AI build-out are weighing on profit and free cash flow.
Anyone assessing the stock should therefore read both levels together. In the short term, the question is how long the advertising market can carry the spending. In the long term, it is whether Llama, Meta AI and the superclusters can become a second pillar.
I work through such questions in a fixed order: first business development, then the long-term trend, then short-term momentum. This order does not make an investment decision for you, and you will deliberately not find a forecast here. But it gives you a framework for checking that works the same way for every stock.