Meta increases its baseline for AI expenditure as revenues rise, although cash flow decreases significantly.
Meta informed investors that it will increase its spending on artificial intelligence this year, adjusting its capital-expenditure forecast for 2026 by raising the lower limit rather than the upper limit. The company now projects expenditures between $130 billion and $145 billion, an increase from a previous range starting at $125 billion, which is roughly twice the amount outlined a year ago.
This expenditure supports a rapidly growing sector. Meta reported second-quarter revenue of $60.8 billion, a 28% increase from a year prior, marking its fastest growth since late 2021, driven by resilient advertising and AI-enhanced recommendations that kept users engaged.
The advertising sector played a crucial role in this growth. Improved AI recommendations increased engagement on Instagram and Facebook, and a stronger advertising market allowed Meta to convert that heightened user interest into its fastest revenue growth in years.
However, this growth isn’t reflected at the bottom of the cash flow statement. Free cash flow plummeted to $784 million, down 91% from $8.55 billion a year earlier, highlighting how the AI expansion is consuming the funds generated by the advertising business.
Earnings also fell short, with profits of $6.18 per share missing analyst expectations of $7.22, hindered partly by legal expenses, even though revenue exceeded forecasts. These legal costs are significant, with Meta reporting a $2.4 billion charge for the quarter, and a group of US states aiming to impose penalties of up to $1.4 trillion over claims that its products are designed to exploit young users, a legal issue that looms over each quarter.
On a positive note, the user base continues to grow. Meta reported 3.6 billion daily active users across its apps, a 3% increase year on year, which is reassuring after a previous quarter where daily user numbers declined.
During the earnings call, Zuckerberg clarified the allocation of spending. A substantial portion of computing resources will be dedicated to training Meta’s models, supporting the core advertising business, and facilitating the development of “personal agents,” which he anticipates will become a significant consumer product.
He also outlined a future strategy. Meta plans to “develop a substantial business catering to large customers,” referring to monetizing AI compute in a cloud-like model that would transform its extensive infrastructure from merely a cost center into a revenue-generating asset.
This perspective addresses investors' concerns. As capital expenditures rise and cash flow diminishes, stakeholders are curious about when this spending will translate into earnings, with Zuckerberg suggesting that the data centers will eventually sustain themselves financially.
The adjustment of the forecast itself is telling. By increasing the lower end of the range instead of the upper, Meta is essentially assuring investors that spending will not fall short, indicating a commitment to the expansion even as returns remain uncertain.
This trend is not exclusive to Meta. Across Big Tech, the numbers from the second quarter reveal a similar narrative of soaring AI investments surpassing the cash they generate, reflecting an industry-wide gamble.
What sets Meta apart is its approach. It is spending like a cloud service provider, although it is not one yet, using an advertising engine to fund its aspirations for superintelligence and asking shareholders to trust that these two aspects will eventually converge.
For the moment, the market seems accustomed to this pattern: strong revenue, dwindling profit, increasing capital expenditure, and a CEO asserting that the investment represents not a risk but a competitive advantage, time after time.
Wall Street has adapted to this trade-off for now. Meta’s stock has largely remained stable during this investment phase, betting that a company growing revenue at 28% can sustain heavy investments as long as growth continues.
The eventual reckoning, if it occurs, will hinge on timing. Meta can handle this expansion today; however, the significant decline in free cash flow raises the question of how many more quarters it can maintain this level of spending before the returns need to materialize.
Other articles
Meta increases its baseline for AI expenditure as revenues rise, although cash flow decreases significantly.
Meta has revised its 2026 capital expenditure forecast to a range of $130–145 billion, increasing the lower limit. In the second quarter, revenue increased by 28%, while free cash flow dropped by 91% due to rising expenses related to AI.
