According to Woofun AI, Meta (META.US) showed a sharp divergence between high revenue growth and negative cash flow in the third quarter of 2026. Founder Mark Zuckerberg is reducing free cash flow to negative $2.12 billion through aggressive infrastructure investment in exchange for dominance of long-term AI strategies.
This fundamental shift in the financial structure stems from the dual drive of strong growth in the advertising business and improved efficiency of traffic monetization. Meta achieved total revenue of US$64.601 billion during the quarter, up 26% year-on-year from US$51.242 billion in the same period in 2025; even after excluding the impact of foreign exchange headwinds of US$113 million, the revenue growth rate was locked at 26% under the fixed exchange rate. The core engine comes from the advertising business. The quarter's ad revenue reached US$63.061 billion, up 26% from US$50.82 billion in the same period last year. The app family's other revenue was $1,097 million, driving the segment's total revenue to $64,158 million; while the Reality Labs division, which focuses on the metaverse and hardware, had revenue of $443 million. On the traffic side, the number of daily active users in the app family reached 3.64 billion, an increase of 3% over the previous year.
More importantly, the deepening of monetization capabilities: ad impressions increased 10% year over year, and the average price of a single ad increased 14%. Management attributed this performance to improvements in the AI recommendation system, believing that intelligent algorithms are significantly improving user engagement and advertising accuracy, thereby extracting incremental value in the traffic stock game.
However, the income statement was severely impacted by sky-high lawsuits and fines, which led to a sharp decline in net profit. In the third quarter, Meta included $9.89 billion in settlement expenses related to youth mental health in its financial report. This huge non-recurring expense directly dragged down overall profit performance. Affected by this, consolidated operating profit fell to $12.988 billion, down 37% year over year; reported net profit of $9.860 billion, or $3.83 per diluted share. More dramatically, on the day of the October 21 earnings report, U.S. District Court Judge Francis Matthew of New Mexico ruled on the state's lawsuit against Meta to mislead consumer data privacy, ordering Meta to pay the state a compensation fine of up to $25 billion.
This heavy punishment at the judicial level not only increased current financial pressure, but also cast a long-term shadow over the company's future compliance costs and potential liabilities, causing the originally strong advertising profits to be greatly eroded by legal expenses.
The core of the cash flow crisis is that capital expenditure surged far faster than operating cash flow, leading to negative free cash flow. Although Meta's operating cash flow for the quarter remained high at US$32.786 billion, up from US$29.99 billion in the same period in 2025, the expenditure side expanded even more rapidly. Cash expenses for property and equipment purchases rose to US$33.942 billion during the quarter, compared to only US$18.829 billion in the same period in 2025; when adding US$964 million in financial lease principal payments, the total related cash expenses reached US$34.906 billion. According to official Meta data calculation, free cash flow for the period fell to negative $2.120 billion. In contrast, free cash flow for the third quarter of 2025 was positive at US$10.625 billion, and cumulative free cash flow for the first nine months also narrowed from US$29.508 billion in the same period in 2025 to US$11.049 billion.
This structural reversal in cash flow indicates that Meta has officially moved from a profit accumulation period to an asset-heavy investment period, and its capital consumption pattern has fundamentally changed.
The expansion of the balance sheet further confirms the profound impact of infrastructure investment. By the end of the third quarter, the company's net property and equipment had increased from US$176.40 billion at the end of 2025 to US$253.22 billion, an increase reflecting large-scale construction of data centers, servers and network infrastructure.
At the same time, long-term debt also increased from US$58.744 billion to US$80,922 million, indicating that the company is using leverage to accelerate computing power deployment. Negative free cash flow does not mean that the company is experiencing a liquidity crisis because cash flow from operating activities is still abundant, but this trend indicates that Meta is rapidly transforming current advertising profits into future computing power assets. The key variables in the future are whether these additional assets can be gradually converted into revenue and profit, and whether operating cash flow can continue to grow to cover expanding capital requirements.
Data compiled by Woofun AI showed that in the face of capital consumption, Meta management did not give a signal of contraction, but instead clarified the strategic direction of continued aggressive investment in 2027. Chief Financial Officer Susan Lee notes that while the 2027 budget is still being drawn up, the company is expected to continue its aggressive investment pace. For the fourth quarter results, the company expects revenue to be between $73 billion and $76 billion, with a guideline midpoint of $74.5 billion, an increase of about 15.3% over the third quarter. With the reallocation of resources, the Reality Labs division's losses are expected to be contained, falling to $18 billion to $19 billion for the full year of 2026, and further narrowing to around $15 billion in 2027. In terms of capital expenditure, Meta raised the lower limit of capital expenditure guidance for the full year 2026 by $5 billion and updated to $135 billion to $145 billion.
The increase in this guideline reflects the company's firm optimism about the long-term value of AI infrastructure. Even when short-term cash flow is under pressure, it is necessary to ensure a leading position in computing power reserves.
In terms of AI product development, the release of Muse Assistant's commercial testing and model matrices became a new growth engine. In just a month and a half since it went live in the US and Canada, Muse has surpassed 10 million set up users, and about 25% of these users use it every day. The most common use case is shopping, accounting for over 25%, and the average spend is over $200. As first-week transaction volume far exceeded expectations, Meta had to work with Stripe to temporarily increase credit limits to cope with surging transaction demand. Next, Muse will launch on Meta smart glasses starting next month to expand the hardware portal. In terms of model development, Meta released the first closed-source cutting-edge model, Muse Wonder, which is at the most advanced level in the industry in terms of many key indicators;
At the same time, Muse Spark 1.2/1.3, Muse Code, and two open source open weighting models, Muse Glimmer and Muse Spark Open, were also released.
The release of this product matrix marks Meta's shift from simple infrastructure investment to application-level commercial exploration in an attempt to open up new revenue streams through personal intelligence.
To match its long-term strategic layout, Meta made major adjustments to its governance structure and issued non-voting Class C common shares to reserve capital space. The Board of Directors has approved the establishment of a new non-voting Class C common stock to pay stock dividends to existing Class A and Class B shareholders in the form of 1 Class C share for every 1 share.
This arrangement will not change the current shareholders' voting rights structure. The company's special committee has signed a relevant agreement with Zuckerberg to continue the current governance arrangement. This move followed the practice of Google's parent company Alphabet (GOOGL.US) in 2014. At that time, Alphabet completed the split by distributing non-voting Class C shares on a pro rata basis, enabling founders Larry Page and Serge Brin to successfully lock in control of the company while continuing to issue shares for employee equity incentives and mergers and acquisitions. For Meta, non-voting C-shares will mainly be used for future employee equity incentives, potential strategic acquisitions, or other capital operations.
The optimization of this governance structure aims to expand financing tools on the premise of maintaining the founder's leadership structure, give the company the ability to invest flexibly driven by AI, and reserve sufficient capital and governance space for huge investments over the year.
The clarification of the technology roadmap revealed computing power bottlenecks, cost efficiency improvements, and the long-term vision of smart commerce. Zuckerberg said that currently the biggest limitation of Muse is mainly computing power intensity, so it has not been widely introduced in regions other than the US and Canada; the company is working to improve computing efficiency and build sufficient computing power to support the scale of more than 1 billion users in the future.
Although currently a compute-intensive application, Muse is expected to increase its cost efficiency by more than 10 times in 2027, which will make it profitable when scaled up. In terms of corporate cooperation, Meta is developing a “Personal Agent Protocol” (Personal Agent Protocol) with partners such as Sierra, Walmart (WMT.US), Stripe, and Shopify (SHOP.US) to define a safe and reliable way to interact, so that consumer Muse can directly negotiate and trade with enterprise intelligence. In terms of model development, Meta plans to distill cutting-edge models into smaller, more efficient product models. It is expected that after a year, it will be able to deliver the same intelligence in a smaller model at a fraction of the current cost of Muse Wonder.
This technology path shows that Meta is trying to solve the high-cost problem of AI applications through model distillation and efficiency optimization, thereby building a sustainable closed loop of commerce.
Meta's long-term strategic bet is to prove that the new revenue and efficiency improvements brought about by AI are enough to support this infrastructure expansion that has continued for many years. In 2027, Zuckerberg will continue to invest aggressively, rely on advertising revenue to provide a funding base, and open up new growth points through the construction of AI infrastructure.
Although AI investments take longer to pay off than core businesses, they have generated significant incremental revenue for core advertising by improving recommendation and monetization efficiency. As investment continues to expand, companies need to prove that the new revenue and efficiency improvements brought by AI can cover growing infrastructure costs and be profitable over the next few years. This is yet another strategic transformation Meta is facing after the mobile internet era, and its success or failure will determine the company's ultimate position in the AI era.