Meta Q2 2026: Revenue jumps 28% to $60.8bn but margins slide on AI costs

Meta Platforms, Inc. (Nasdaq: META) reported second-quarter 2026 revenue of $60.80 billion, up 28% year-over-year and modestly ahead of consensus, but the print landed with a distinctly different signal than the top line alone would suggest. Diluted earnings per share came in at $6.18, down 13% year-over-year, undershooting Street expectations after the company absorbed $2.40 billion of charges related to legal proceedings and $1.18 billion of severance costs tied to its May 2026 headcount reduction of roughly 8,000 employees. Operating margin fell to 31% from 43% a year earlier, free cash flow collapsed to $784 million from $8.55 billion in the prior-year quarter, and Meta bought back zero shares during the period after repurchasing more than $10 billion of stock in the same quarter of 2025. The central tension is now unusually sharp for a company of Meta’s scale: artificial-intelligence investment is visibly accelerating advertising monetisation, yet the cost curve, capital-expenditure profile and litigation exposure are simultaneously reshaping the earnings and cash-return profile that supported the equity’s re-rating over the past three years.

How is artificial intelligence translating into measurable advertising demand across Meta’s Family of Apps?

Advertising revenue reached $59.36 billion, an increase of 27% year-over-year and 26% on a constant-currency basis. The composition of that growth is important. Ad impressions delivered across the Family of Apps rose 14% year-over-year, while the average price per ad rose 12%. That combination indicates continued expansion of the inventory base, but also a genuine pricing dynamic that has been elusive for large parts of the digital advertising market since 2022. Meta has consistently attributed the price improvement to targeting and ranking gains driven by its Andromeda ads model and its Advantage+ automation suite, both of which rely on the artificial-intelligence infrastructure the company has been building out.

Family daily active people averaged 3.60 billion in June 2026, up 3% year-over-year. Engagement growth of this scale on a platform base this large is itself analytically significant, because incremental impressions must be monetised at similar or better yields to preserve unit economics. Family of Apps revenue reached $60.37 billion for the quarter, with Family of Apps operating income of $23.39 billion. That operating income was actually lower than the $24.97 billion generated in the prior-year quarter, meaning the advertising business is running faster on the revenue line but is carrying materially more cost, largely from research and development and the general and administrative line that houses the legal charge.

Why did operating margin fall to 31% even as revenue accelerated by 28% year-over-year?

The margin compression reflects two distinct forces that investors need to separate. The first is structural. Research and development expenses climbed to $21.66 billion, up 67% year-over-year, reflecting continued expansion of the AI research and infrastructure programme, higher personnel costs and depreciation from the enlarged property and equipment base. Property and equipment, net, rose to $225.72 billion at 30 June 2026 from $176.40 billion at 31 December 2025, a $49.3 billion increase in six months that is beginning to flow through the income statement as depreciation. Cost of revenue also rose 33% to $11.33 billion, consistent with higher infrastructure operating costs. These are not one-off items. They are the running cost of the AI capital cycle.

The second force is more discrete. General and administrative expenses jumped to $5.61 billion from $2.66 billion, and Meta disclosed that the second-quarter figure included $2.40 billion of charges related to legal proceedings. The company did not itemise the specific matter or matters in the release. Chief Financial Officer Susan Li noted separately that Meta continues to see scrutiny on youth-related issues in several markets and has a number of youth-related trials scheduled for this year in the United States, which she said may ultimately result in a material loss. On top of the legal charge, general and administrative expenses were also carrying the $1.18 billion severance from the May 2026 workforce reduction. Stripping both items would still leave underlying operating margin below the prior-year level, but the reported 31% figure overstates the run-rate margin compression.

What does the raised full-year capital expenditure range signal about Meta’s artificial-intelligence build-out?

Meta narrowed its 2026 capital expenditure guidance to a range of $130 billion to $145 billion, from a previous range of $125 billion to $145 billion. The lower end has moved up, and the ceiling is unchanged. The company spent $31.08 billion on capital expenditure in the second quarter alone, including principal payments on finance leases. For context, full-year 2025 capital expenditure was $72.2 billion, meaning the 2026 mid-point of $137.5 billion implies close to a doubling of the annual capital-spending base within a single year.

The direction of that spending was reinforced on 28 July 2026, when Meta announced a strategic venture with BlackRock, Inc. to develop and build a data centre campus in El Paso, Texas, with Meta committing more than $10 billion. That transaction fits a pattern of large hyperscalers partnering with financial sponsors to co-fund data-centre buildouts, effectively bringing external capital into the AI infrastructure race while retaining operational control. For Meta, the arrangement extends the capital envelope beyond internally generated cash flow at a moment when free cash flow is running well below prior-year levels. The strategic logic is straightforward: compute capacity is now the binding input for both consumer AI features and the enterprise AI opportunity Mark Zuckerberg referenced in his prepared statement. The question for investors is not whether Meta should be building capacity, but whether it can deploy that capacity at incremental returns that exceed its rising cost of capital.

How significant is the capital-allocation pivot from buybacks to debt-funded investment?

The most striking capital-allocation signal in the release is what did not happen. Meta repurchased zero Class A shares during the second quarter of 2026. In the same quarter of 2025 it repurchased $10.17 billion of stock, and in the first half of 2025 it repurchased almost $23 billion. Dividend and dividend-equivalent payments were unchanged at $1.35 billion for the quarter, meaning the entire cash-return pause is concentrated in the buyback line.

Simultaneously, Meta raised $24.91 billion in net proceeds from long-term debt issuance during the quarter. Long-term debt on the balance sheet rose to $83.66 billion at 30 June 2026 from $58.74 billion at year-end 2025. Cash, cash equivalents and marketable securities stood at $90.26 billion. Net cash and marketable securities relative to long-term debt is now finely balanced in a way it has not been for several years. For a business that historically generated large surplus cash flow and returned most of it, this represents a deliberate shift toward debt-funded reinvestment. Investors evaluating the equity now need to model Meta more like an infrastructure company with an advertising engine attached, rather than an advertising company with excess cash. The change in balance-sheet posture also raises the sensitivity of the equity to real rates and to any perceived slippage in the AI monetisation thesis.

What do Meta’s third-quarter guidance and full-year outlook reveal about management’s confidence?

Meta guided third-quarter 2026 total revenue to a range of $61 billion to $64 billion, implying year-over-year growth of roughly 21% to 27% depending on where in the range results land, and assuming a foreign-currency headwind of about 1%. Full-year 2026 total expenses are now expected in the range of $165 billion to $169 billion, with the low end raised to incorporate the $2.4 billion legal charge. The company continues to expect full-year 2026 operating income to exceed the 2025 level, an important reassurance given the reported second-quarter operating income decline.

The tax-rate outlook was also raised. Meta now expects a tax rate of 15% to 17% for the remaining quarters of 2026, compared with a prior outlook of 13% to 16%. That upward shift, together with the higher expense base, means the pathway to earnings-per-share growth in the second half depends materially on ongoing revenue acceleration and on the absence of further large discrete items. The Q3 guidance mid-point of $62.5 billion, if achieved, would represent one of the largest sequential revenue expansions in the company’s history and would provide meaningful support for the AI-driven monetisation thesis.

What role does Reality Labs still play in the investment case after another quarter of heavy losses?

Reality Labs generated $431 million in second-quarter revenue and reported an operating loss of $4.62 billion, compared with $370 million of revenue and a $4.53 billion loss in the prior-year quarter. The unit remains loss-making at approximately the same rate, meaning it consumes roughly $18 billion to $19 billion of operating income on an annualised run rate. In the current environment, where consolidated operating margin is under pressure from AI investment, Reality Labs occupies a delicate position. It is neither growing quickly enough to reframe itself as a near-term revenue driver nor shrinking losses in a way that would materially expand consolidated margins. For long-only investors focused on Meta’s AI story, Reality Labs is increasingly treated as an option on the eventual convergence of virtual, augmented and AI-native computing, rather than a near-term financial contributor.

How should investors think about the litigation and regulatory overhang flagged by the chief financial officer?

The $2.40 billion legal charge is a one-quarter figure, but the surrounding disclosure suggests continued exposure. Chief Financial Officer Susan Li’s specific reference to youth-related trials scheduled in the United States for 2026, with a potential material loss, is important because it establishes that the current quarter’s charge cannot be treated as a clean closure of legal risk. The company has not publicly linked the second-quarter charge to any specific matter, and it should not be presented as an admission of liability on any particular allegation. What can be said with more confidence is that legal and regulatory costs are moving from being an occasional line item to a recurring pressure on the general and administrative line, at a scale that is beginning to matter for consolidated margin. Any incremental material loss in coming quarters would compound the effect of the ongoing AI-investment cost base.

What has strengthened, what remains unresolved, and what would test the thesis next?

What has clearly strengthened is the underlying advertising engine. Impressions and pricing are both growing at double digits year-over-year, and the guidance implies continued top-line acceleration into the third quarter. What remains unresolved is whether the current cost trajectory can normalise on the far side of the AI capital cycle, or whether operating margin will compress structurally into the low thirties or lower. The next measurable proof point is the third-quarter print. A revenue outcome near the top of the $61 billion to $64 billion range, combined with an underlying margin that does not deteriorate further once discrete legal and severance items are excluded, would strengthen the thesis that Meta is buying growth rather than defending market share. Conversely, a further margin step-down or an additional material legal loss would raise legitimate questions about the earnings power that supports the current market capitalisation. The equity’s reaction in extended trading following the release, which coincided with the announcement, suggests that the market is already recalibrating for a higher-cost, higher-capex Meta, rather than the buyback-heavy free-cash-flow story of the prior cycle.

Key takeaways from Meta’s Q2 2026 results, artificial-intelligence capex programme and legal-charge disclosure

• Meta reported Q2 2026 revenue of $60.80 billion, up 28% year-over-year, with advertising revenue of $59.36 billion, up 27%.
• Ad impressions rose 14% and average price per ad rose 12%, indicating that both inventory expansion and yield are contributing to advertising growth.
• Operating margin fell to 31% from 43%, and diluted earnings per share fell 13% to $6.18, weighed by a $2.40 billion legal charge and $1.18 billion of severance costs.
• Second-quarter capital expenditure reached $31.08 billion, and full-year 2026 capex guidance was narrowed to $130 billion to $145 billion, roughly double 2025 levels.
• Meta repurchased no shares during the quarter, compared with more than $10 billion of buybacks in the same period of 2025, while issuing $24.91 billion in net long-term debt.
• Reality Labs generated $431 million in revenue and a $4.62 billion operating loss, broadly unchanged from a year earlier.
• Third-quarter 2026 revenue guidance of $61 billion to $64 billion implies continued acceleration, and full-year operating income is still expected to exceed the 2025 level.
• Chief Financial Officer Susan Li flagged youth-related trials scheduled for 2026 in the United States that may ultimately result in a material loss, keeping the litigation overhang live.
• The strategic venture with BlackRock to develop an El Paso data centre campus, announced 28 July 2026, illustrates Meta’s move toward externally co-funded AI infrastructure.
• Shares declined in extended trading following the release, with the market response reflecting concern about margin trajectory rather than the top-line beat.