[META] Meta: Q3 2026 Earnings Preview on Ad Pricing and AI Capex

Meta, the parent of Facebook, Instagram and WhatsApp, earns almost all of its revenue from advertising on those apps, and it will hold its earnings call on 2026-10-28 [1] to report results for the third quarter of 2026, ending September 30, 2026. Heading into this Meta Q3 2026 earnings preview, the latest disclosed quarter, the second quarter of 2026, showed revenue of $60.80 billion, up 28% year over year, with ad impressions up 14% and average price per ad up 12%; total costs and expenses, however, rose 55% to $42.03 billion, operating income fell 8% to $18.78 billion and free cash flow was only $784 million [2]. Management guided third-quarter total revenue to $61-64 billion, assuming foreign currency is a headwind of about 1 percentage point to year-over-year growth; it guided full-year total expenses to $165-169 billion and capital expenditures, including principal payments on finance leases, to $130-145 billion, and it kept its statement that full-year operating income will be above 2025 [3]. According to analyst estimates compiled by Drillr, the third-quarter revenue consensus from 28 analysts is $63.28 billion, in the upper half of the company's range; the operating income consensus is $22.15 billion but the lowest estimate is only $11.23 billion, and the earnings-per-share consensus from 27 analysts is $6.77, in a range of $3.33 to $7.52 [4].

Three things matter most in these results. The first is whether ad pricing can keep carrying revenue while impression growth slows: impression growth fell from 19% in the first quarter to 14% in the second [5], second-quarter price growth included a favorable currency effect [6], and currency turns into a headwind in the third quarter, so whether revenue lands above the $62.5 billion midpoint of guidance directly tests the explanation that AI is lifting ad performance and pushing prices up. The second is whether expenses stay within the full-year range: even after excluding $2.4 billion of legal charges and $1.18 billion of severance, second-quarter expense growth still clearly outpaced revenue growth [2], the third quarter is the first full quarter after the May layoffs, and newly deployed servers will keep pushing depreciation higher, so the expense path decides whether the promise of higher full-year operating income than in 2025 holds. The third is whether free cash flow turns negative for the first time and how any gap is funded: the full-year capex range implies roughly $39.5-47 billion per quarter in the second half, above the roughly $32 billion of quarterly operating cash flow in the first half [7], so the third quarter will show whether Meta's AI build-out has started to rely on new debt and third-party partnerships.

Company Background and Business Structure

Meta is a founder-controlled social platform company that is now pouring the cash from its advertising business into AI on a large scale. It started as Facebook, Inc., founded in 2004, renamed itself Meta Platforms in October 2021 and is headquartered in Menlo Park, California; founder Mark Zuckerberg serves as chairman and CEO and controls voting power through Class B shares. Meta had 78,865 employees at the end of 2025 [8]; after cutting about 8,000 jobs in May 2026, it reported 75,472 employees at the end of June, and most of the affected employees will only leave the headcount figure by the end of the third quarter [2]. In 2026 the company released its own Muse family of models and is building AI agents for consumers and businesses [9].

Meta's revenue and profit are concentrated in the Family of Apps segment, while the other reportable segment, Reality Labs, keeps losing money. In 2025 Family of Apps generated $198.76 billion of revenue and $102.47 billion of operating income, a 52% margin, while Reality Labs had $2.21 billion of revenue and a $19.19 billion operating loss [10]. Family of Apps revenue comes almost entirely from advertising: marketers buy ads directly or through agencies and resellers and pay per impression or per user action such as a click, and the company calculates average price per ad as total advertising revenue divided by the number of ads delivered [11]. Other revenue, mainly WhatsApp business messaging and subscriptions, reached $1.007 billion for the first time in the second quarter of 2026, up 73% [6]; Reality Labs sells Quest headsets and AI glasses such as Ray-Ban Meta and Oakley Meta, and the company plans to spend about 70% of Reality Labs operating expenses on wearables in 2026 [12].

By region and cost structure, North America is Meta's fastest-growing high-value market, while the infrastructure needed to deliver ads increasingly relies on outside resources. In the second quarter of 2026, revenue by user geography was $23.86 billion in the United States and Canada, $14.01 billion in Europe, $16.07 billion in Asia-Pacific and $6.86 billion in the rest of the world, with the United States and Canada up 32%, Europe up 24% and Asia-Pacific up 19% [13]; the US, Canadian and European ad markets are larger and more mature and earn more per unit, while impression growth comes mainly from lower-priced regions such as Asia-Pacific [14]. Delivering ads requires Meta's own data centers, servers and network, so cost of revenue consists mainly of depreciation on those assets, energy and bandwidth, and partner payments [15], and in 2026 the company is also renting more third-party cloud capacity and buying more third-party AI tokens [16]. Meta has also begun sharing the build-out through asset partnerships: it plans to contribute El Paso data center assets with a carrying value of $1.48 billion to a third party in the third quarter to co-develop data centers [17].

Financial History and Current Position

In 2025 Meta grew revenue and profit together, but capital spending had already started to absorb more cash. Full-year revenue was $200.97 billion, up 22%, operating income was $83.28 billion, up 20%, for a 41% operating margin, and net income was $60.46 billion; the full-year effective tax rate was about 30%, higher than in prior years mainly because of a one-time tax charge in the third quarter of 2025 [18]. Operating cash flow was $115.80 billion, capital expenditures including finance lease principal were $72.22 billion and free cash flow was $43.59 billion, down from $52.10 billion in 2024 [19]; the company bought back $26.26 billion of stock, paid $5.32 billion of dividends and ended the year with $81.59 billion of cash and marketable securities and $58.74 billion of long-term debt [8].

In the first half of 2026 Meta's revenue growth accelerated, but its margin dropped sharply in the second quarter. First-quarter revenue was $56.31 billion, up 33%, total expenses were $33.44 billion, up 35%, free cash flow was $12.39 billion, and net income also benefited from an $8.03 billion income tax benefit [5]; first-quarter operating income was $22.87 billion, a margin of about 41% [20]. Second-quarter revenue was $60.80 billion, up 28%, or 27% in constant currency, and total expenses were $42.03 billion, up 55%, including $2.40 billion of legal charges and $1.18 billion of severance [2]. Second-quarter operating income was $18.78 billion, down 8%, with the margin falling from 43% a year earlier to 31%; Family of Apps earned $23.39 billion at a 39% margin, while Reality Labs lost $4.62 billion [21]; net income was $15.85 billion, or $6.18 per diluted share [22].

Second-quarter cash flow shows that capital spending has caught up with operating cash flow. Operating cash flow was $31.86 billion, capital expenditures were $31.08 billion and free cash flow was only $784 million; at the end of June, cash and marketable securities were $90.26 billion and long-term debt was $83.66 billion, about $24.9 billion higher than at the end of 2025 [2]. Meta did not repurchase any stock in the first half, raised $24.91 billion net from fixed-rate senior unsecured notes in May, and reclassified $10.80 billion of money market funds as restricted cash under escrow requirements in multi-year infrastructure purchase agreements [7].

Operating Model

Meta's advertising revenue equals ad impressions multiplied by average price per ad, and right now one factor is slowing while the other stays high. Impressions depend on daily users, time spent per user and ad load per session: an average of 3.60 billion people used at least one Meta app each day in June 2026, up 3% [2], and part of second-quarter impression growth came from showing ads more frequently [6]. Price depends on advertiser bidding, which the company attributes mostly to better ad performance from AI targeting and measurement tools, while regional and product mix also matter because Asia-Pacific and Reels monetize at lower rates [6]; on the earnings call, management said LLM-powered ad matching lifted Facebook ad clicks by 8.3% and conversions by 15.7% [23]. Impressions and price grew 12% and 9% in 2025 [24], 19% and 12% in the first quarter of 2026 [5], and 14% and 12% in the second quarter [6]; because advertising revenue is recognized when an ad is shown or a user acts, better performance reaches revenue with almost no lag.

Operating income equals revenue minus cost of revenue, research and development, marketing and sales, and general and administrative expense, and the growth mix of these lines is shifting in 2026. Advertising has almost no variable cost of goods: in 2025 cost of revenue was $36.18 billion, about 18% of revenue, and R&D was $57.37 billion, about 29% of revenue [18]. In the second quarter cost of revenue rose 33%, R&D 67% and marketing and sales 15%, driven by data centers, third-party cloud, third-party AI tokens, share-based compensation and severance [25], while general and administrative expense roughly doubled from a year earlier to $5.61 billion because of the $2.40 billion legal charge [22]. Capital spending reaches profit with a clear lag: servers begin depreciating over 5.5 years only once they are placed in service [26], so this year's heavy equipment purchases will keep lifting cost of revenue and R&D over the following quarters; Reality Labs loses about $4.0-4.6 billion per quarter, and the company expects its full-year loss to be similar to 2025 [27].

Free cash flow equals operating cash flow minus purchases of property and equipment and finance lease principal, and the key issue now is that capital spending is growing much faster than operating cash flow. Advertisers pay quickly, and share-based compensation and depreciation are large non-cash charges, so operating cash flow runs well above net income: in 2025 operating cash flow was $115.80 billion against net income of $60.46 billion that year, and the company raised $29.91 billion net from debt [19]. Capital spending rises from $72.22 billion in 2025 to the $130-145 billion guided for 2026 [3], with the gap filled by cash on hand, new long-term debt and third-party partnerships, including another $24.91 billion of net debt proceeds in May 2026 [7]. The longer-term cash need is written into contracts: at the end of June Meta had $349.31 billion of non-cancelable contractual commitments and about $278.99 billion of data center and other leases that have not yet commenced and will start between the rest of 2026 and 2036 [28], and these obligations shape cash spending and expenses from 2027 onward.

Industry and Competitive Position

Meta is one of the leaders in global digital advertising, but it competes with several platforms for advertiser budgets and user time. It competes with Alphabet's Google Search and YouTube, Amazon Ads, TikTok, Snap and Pinterest for ad budgets, with YouTube and TikTok for time spent, and with OpenAI, Google, Anthropic and others in AI models and personal assistants [29]. Its advantages come from user scale, first-party behavioral data and a performance-priced ad system, and online commerce was the largest contributor to advertising revenue growth in the first half of 2026 [6]; in the second quarter, global time spent on Instagram grew by double digits and global video time on Facebook rose 9% [23].

Meta's constraints come mainly from platforms it does not control, regulation and the capital intensity of the AI race, and the available material cannot quantify these constraints against peers. Apple and Google control the mobile operating systems, and Apple's iOS changes limit ad targeting and measurement [29]; in the EU the company has had to offer free users a less personalized ads option that is less relevant and effective than its standard ads [30]. The AI race requires capital spending far above Meta's historical levels, while new businesses such as the Muse personal agent, an enterprise AI platform and compute sales are still early, and management has acknowledged that consumer behavior may not shift quickly toward personal agents [31]. The available material has no comparable peer financials for the same period, so it is not possible to say how much of Meta's price growth reflects share gains rather than a broader recovery in the ad market.

Core Debates

With ad impression growth already slowing from 19% to 14%, can higher prices per ad keep Meta's third-quarter revenue in the upper half of its guidance?

This question decides whether the only evidence so far that Meta's AI spending is paying off at scale can continue. Almost all of Meta's revenue comes from advertising on Facebook and Instagram [30], and advertising revenue equals impressions times price; advertising revenue grew 30% in the first half of 2026 [6], faster than the 22% for full-year 2025 [24]. The company attributes price growth mainly to AI-driven ad performance, which is also what supports operating income while expenses and capital spending climb quickly; with currency turning into a headwind and impressions facing a high base, the third quarter is the first stress test of that explanation [3].

Evidence for the pricing story remained clear in the second quarter, but the evidence against it is just as specific. Revenue was $60.80 billion, up 28%, or 27% in constant currency, with impressions up 14% and price per ad up 12% [2], and United States and Canada revenue grew 32%, faster than the company overall [13]; LLM-powered ad matching lifted Facebook ad clicks by 8.3% and conversions by 15.7%, and 9 million small businesses use at least one AI ad creative tool [23]. On the other side, impression growth slowed from 19% in the first quarter to 14% [5], part of impression growth came from showing ads more often, second-quarter price growth included a favorable currency effect [6], and the $63.28 billion revenue consensus already sits near the top half of guidance [4].

The financial transmission in this debate is direct; what remains unresolved is how much of the price increase comes from AI performance. AI recommendations raise time spent and ad frequency, which lifts impressions; AI targeting, measurement and ad matching raise conversions, which lifts advertiser bids and average price; the product of the two is advertising revenue, which flows into Family of Apps revenue and operating income. Meta discloses only year-over-year growth in impressions and price, so the contributions of ad frequency, user count and time spent cannot be separated, and price growth cannot be cleanly separated from currency and industry mix [11]; the alternative explanation is that the first-half acceleration came mainly from higher ad load and currency, neither of which is durable, and that price growth will return to roughly the 9% seen in 2025 [24].

The third quarter offers five signals and two clear falsifiers. Watch whether total revenue is at least the $62.5 billion guidance midpoint, whether price-per-ad growth stays at or above 11% despite the currency headwind, whether impression growth drops below 12%, whether United States and Canada revenue still grows faster than the company overall, and whether other revenue such as WhatsApp paid messaging keeps growing around 70% [6]. If revenue comes in below $61 billion with price growth under 9%, or if the company blames the slowdown on advertiser budgets rather than currency, the view that pricing can hold up revenue would be falsified [3].

With AI infrastructure and people costs rising faster than revenue, can Meta keep third-quarter expenses within its full-year guidance?

The expense question decides whether the statement that full-year operating income will exceed 2025 holds up. In the second quarter revenue grew 28% but total expenses grew 55%, operating income fell 8% and the operating margin dropped from 43% to 31% [21]. Management still guides full-year operating income above the $83.28 billion earned in 2025 [3], while first-half operating income totaled $41.65 billion [22], so the second half needs at least about $41.63 billion, roughly matching the first half; the third quarter is the first full quarter after the May layoffs and another quarter of rising depreciation, which makes it the best test of whether the ad business can absorb the cost of AI.

The evidence that expenses are under control comes mainly from one-time items and the layoffs, while the evidence against it comes from infrastructure. Excluding the $2.40 billion legal charge and $1.18 billion of severance, second-quarter total expenses were about $38.45 billion and the operating margin about 37%, and most of the roughly 8,000 employees affected by the layoffs will leave the headcount figure by the end of the third quarter [2]; the company also expects its full-year Reality Labs loss to be similar to 2025 [27]. Yet second-quarter cost of revenue rose 33%, R&D 67% and marketing and sales 15%, mainly because of data centers, third-party cloud and third-party AI tokens [25]; derived from the cash flow statements, second-quarter depreciation and amortization was about $6.36 billion, up about 46% [32][33], while most of the $50.92 billion of first-half capital spending [7] has not yet started depreciating.

The biggest unknown in the expense outlook is a legal charge that has not been confirmed. Media reports in late August 2026 said Meta had reached a settlement of about $16.68-17.1 billion with roughly 47 states over youth harm litigation, with Florida declining to join, but the company has not confirmed the amount or timing in its formal disclosures, and the charge is not part of the full-year expense range given in July, which includes only the $2.4 billion legal charge recognized in the second quarter [3]. The average analyst estimate for third-quarter operating income is about $22.15 billion, but the lowest is only about $11.23 billion, which may mean some analysts have already included a large charge [4]. The alternative explanation is that the jump in expense growth came mainly from one-time severance and legal charges and that underlying cost growth has already slowed with the layoffs; because the company does not disclose how those two items are split between segments, the Family of Apps margin excluding one-time items can only be approximated [21].

The third-quarter expense test has five checkpoints and two falsifiers. Watch whether total expenses excluding any new legal charge stay at or below about $46.8 billion, the average quarterly amount left for the second half after subtracting first-half expenses of $75.46 billion from the top of the full-year range [22]; whether the Family of Apps margin excluding one-time items holds around 44%; whether cost of revenue growth falls back below revenue growth; whether period-end headcount drops below 69,500; and whether the company confirms a multistate settlement charge while still stating that full-year operating income will exceed 2025 [3]. If the company raises its expense range excluding legal charges, or if depreciation grows more than 15% quarter over quarter, the view that expenses are under control would weaken.

As capital spending keeps accelerating, will Meta's free cash flow turn negative for the first time in the third quarter, and how will it fund the gap?

This question goes to whether Meta's AI build-out has started to depend on outside money. Meta used to fund buybacks, dividends and data center construction from advertising cash flow at the same time, generating $43.59 billion of free cash flow in 2025 [19] and buying back $26.26 billion of stock [8]. In the second quarter of 2026, free cash flow was only $784 million [2], there were no buybacks in the first half and the company raised another $24.91 billion net from debt in May [7]; the full-year capex range implies roughly $39.5-47 billion per quarter in the second half, above the roughly $32 billion of quarterly operating cash flow in the first half [3][32].

The evidence that funding is ample is Meta's cash balance and the financing mix management has described, while the evidence against it is that commitments are growing even faster. Cash and marketable securities were $90.26 billion at the end of June [2] and operating cash flow runs about $32 billion per quarter [32]; management said it will combine operating cash flow, cost-efficient long-term debt and third-party partnerships such as the BlackRock data center venture [34], and it plans to contribute El Paso data center assets to a co-development partner in the third quarter [17]. On the other side, first-half capital spending was $50.92 billion, the full-year range requires at least $79.08 billion more in the second half, and $10.80 billion of money market funds has been moved into restricted escrow for infrastructure purchases [7]; non-cancelable contractual commitments rose from $131.05 billion at the end of 2025 [35] to $349.31 billion at the end of June, on top of about $278.99 billion of leases not yet commenced and about $68 billion of new data center leases signed in July [28].

The cash flow chain is clear; what remains open is the timing of third-quarter spending and any extra draw on cash. The gap between capital spending and operating cash flow determines free cash flow, free cash flow in turn drives changes in cash, long-term debt and partner funding, and contractual commitments and uncommenced leases will turn into future capital spending and expenses [28]. Because capex timing depends on equipment deliveries, third-quarter spending could come in below the second-half quarterly average and push a negative free cash flow quarter into the fourth quarter; if the reported multistate settlement is paid on schedule, it would also draw on operating cash flow.

The third-quarter cash test covers five items and has two clear falsifiers. Watch whether free cash flow turns negative, whether quarterly capital spending falls within $39.5-47 billion, whether long-term debt rises by more than another $15 billion, whether contractual commitments keep rising sharply, and whether the El Paso asset contribution is completed in the third quarter [17]. If free cash flow is negative and filled mainly with new debt, or if the company raises the top of its full-year capex range, the view that advertising cash flow can fund the AI build-out on its own would be falsified.

Risks and Falsifiers

Youth-related litigation is the risk most likely to hit the income statement directly in the third quarter. A New Mexico jury ordered Meta to pay a $375 million civil penalty, the state's attorney general is seeking another $953 million in abatement costs, and the next two user bellwether trials are scheduled to begin on October 28, 2026 [36]; media have also reported a settlement of about $16.68-17.1 billion with roughly 47 states, which the company has not confirmed in formal disclosures. Such charges are booked in general and administrative expense and reduce operating income in the quarter they are recognized, settlement payments would draw on operating cash flow, and the full-year expense range includes only the $2.4 billion legal charge recognized in the second quarter [3]; if third-quarter disclosures make any new legal charge and payment schedule clear and the statement that full-year operating income will exceed 2025 is unchanged, the impact of this risk on the full-year picture would be bounded.

EU regulation threatens the performance and price of European ads. The European Commission found that Meta's "subscription for no ads" model breaches the Digital Markets Act and fined it €200 million; Meta has appealed, but further changes to its less personalized ads offering may still be imposed during the appeal, and in June 2026 the Commission also imposed an interim measure requiring WhatsApp to give general-purpose AI providers free access to its business API [37]. Europe generated $14.01 billion of revenue in the second quarter, about 23% of the total [13], and a forced expansion of less personalized ads would lower both ad performance and price in Europe; if European revenue growth in the third quarter is at least the 24% of the second quarter and the company discloses no new EU product changes, this risk has not materialized.

A pullback in ad pricing is the most direct revenue risk. Part of first-half price growth came from currency and concentrated spending by online commerce advertisers [6], and AI performance gains could be matched by competitors or eroded by platforms and regulators [29]. Based on second-quarter advertising revenue of $59.36 billion, each percentage point less of price growth cuts quarterly revenue by about $590 million, almost all of which falls to operating income [21]; if third-quarter price growth is at least 11% and United States and Canada revenue grows faster than the company overall, this risk has not shown up.

Front-loaded AI infrastructure costs mean depreciation, cloud capacity and AI token expenses rise ahead of revenue. First-half capital spending was $50.92 billion against a full-year range of $130-145 billion [7]; with a 5.5-year server useful life [26], every $10 billion of server spending adds about $1.8 billion of annual depreciation, all of which lands in cost of revenue and R&D [25]. If third-quarter cost of revenue grows no faster than revenue and depreciation rises no more than 10% quarter over quarter, the pressure from front-loaded costs remains absorbable.

Externalizing the funding gap is a longer-term risk, meaning the AI build-out relies more and more on debt and leases that have not yet commenced. Long-term debt has risen from $58.74 billion at the end of 2025 [8] to $83.66 billion at the end of June [2], and about $278.99 billion of leases have not yet commenced and will generate lease expense or depreciation and interest each year once they start [28]. If third-quarter free cash flow is at least zero and long-term debt does not increase, this risk would be falsified.

What to Watch Next

• Ad pricing and revenue: second-quarter revenue was $60.80 billion. Watch whether third-quarter revenue reaches at least the $62.5 billion guidance midpoint; revenue below $61 billion with price growth under 9% would falsify the pricing view.
• Impressions and price: second-quarter impressions grew 14% and price per ad 12%. Watch whether price growth stays at or above 11% despite the currency headwind and whether impression growth falls below 12%; blaming the slowdown on advertiser budgets rather than currency would weaken the view.
• United States and Canada: revenue grew 32% in the second quarter. Watch whether it still grows faster than the company overall.
• Expenses: second-quarter total expenses were $42.03 billion including one-time items. Watch whether expenses excluding any new legal charge stay at or below about $46.8 billion; a higher expense range excluding legal charges would weaken the view.
• Depreciation and cost of revenue: cost of revenue grew 33% in the second quarter. Watch whether it falls back below revenue growth; depreciation growth above 15% quarter over quarter would weaken the view.
• Headcount and legal charges: headcount was 75,472 at the end of June. Watch whether it drops below 69,500 and whether a multistate settlement charge is confirmed while the company keeps its statement that full-year operating income will exceed 2025.
• Free cash flow and capex: second-quarter free cash flow was $784 million on $31.08 billion of capex. Watch whether free cash flow turns negative and whether capex lands within $39.5-47 billion; negative free cash flow filled mainly with new debt would falsify self-funding.
• Debt and commitments: long-term debt was $83.66 billion at the end of June. Watch whether it rises by more than another $15 billion and whether the El Paso contribution closes; a higher top end of the capex range would weaken the view.

Conclusion

Meta's business is still driven by a high-margin advertising engine, but in 2026 it is absorbing a simultaneous expansion in AI costs and capital spending. Second-quarter revenue was $60.80 billion, up 28%, with impressions up 14% and price per ad up 12%, yet free cash flow fell to $784 million and long-term debt rose to $83.66 billion [2]; in the same quarter total expenses rose 55% and operating income fell 8% [21]. The central unresolved relationship is whether AI-driven ad pricing can outrun expenses pushed up by AI infrastructure, people and legal charges while still generating enough cash for full-year capital spending of $130-145 billion [3].

Two independent commentaries published after the second-quarter results both focus on the return on AI spending rather than on third-quarter revenue itself. Writing for The Motley Fool on August 6, Adam Spatacco argued that because Meta's capital spending is growing faster than its profits, free cash flow must shrink, that the company is "effectively investing tomorrow's profits today," and that if the pattern continues, self-funding its AI plans will get harder and the balance sheet will have to absorb the shortfall [38], which maps directly onto the free cash flow and funding debate. Business Insider reported on September 28 that MoffettNathanson analysts estimate consumers make about 47 payments a month on average, and that after stripping out automatic bill payments, gas stations and in-store shopping only about 11 remain that an agent such as Muse might handle, so thin transaction volume would make a per-transaction model hard to monetize and Meta may end up selling ads inside Muse [39]. Both doubt that new spending will produce new sources of cash soon; the first looks at the size of cash flow and the second at the monetization path of new products. They are outside interpretations rather than facts or a majority view, but both make the third-quarter verdict depend more on whether existing ad pricing and impressions can keep growing.

Over the next few quarters, the view that the ad engine can absorb both AI costs and capital spending would be strengthened if price-per-ad growth stays above 11% despite the currency headwind, expenses excluding new legal charges stay within the full-year range, free cash flow stays positive and long-term debt stops rising. Conversely, if revenue falls below $61 billion with price growth under 9%, the company raises its expense or capex range, or free cash flow turns negative and is filled mainly with new debt, that view would be materially weakened.

Sources

[1] Drillr earnings calendar (updated 2026-09-28) · META 2026-10-28 call · 2026-09-28 · Drillr earnings calendar

[2] META 8-K filed 2026-07-29 · Q2 2026 highlights · 2026-07-29 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001326801&type=8-K&dateb=&owner=include&count=40

[3] META 8-K filed 2026-07-29 · Q3 2026 and full-year outlook · 2026-07-29 · Meta · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001326801&type=8-K&dateb=&owner=include&count=40

[4] Drillr analyst_financial_estimates (updated 2026-09-28) · META quarter ending 2026-09-30 · 2026-09-28 · Drillr analyst_financial_estimates · https://gateway.drillr.ai/mcp/private

[5] META 8-K filed 2026-04-29 · Q1 2026 highlights · 2026-04-29 · 8-K · https://www.sec.gov/Archives/edgar/data/0001326801/000162828026028364/meta-03312026xexhibit991.htm

[6] META 10-Q filed 2026-07-30 · Q2 2026 advertising drivers · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[7] META 10-Q filed 2026-07-30 · liquidity, capex and capital return · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[8] META 10-K filed 2026-01-29 · FY2025 key metrics and capital · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[9] META Q1 2026 earnings call 2026-04-29 · segment results and capex commentary · 2026-04-29 · earnings-call · https://investor.atmeta.com/

[10] META 10-K filed 2026-01-29 · FY2025 segment results · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[11] META 10-Q filed 2026-07-30 · how advertising is sold and priced · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[12] META 10-K filed 2026-01-29 · Reality Labs products and 2026 spend mix · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[13] META 10-Q filed 2026-07-30 · Q2 2026 revenue by geography · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[14] META 10-K filed 2026-01-29 · FY2025 revenue by geography · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[15] META 10-K filed 2026-01-29 · cost of revenue composition · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[16] META 10-Q filed 2026-07-30 · Q2 2026 results overview · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[17] META 10-Q filed 2026-07-30 · El Paso data center held-for-sale assets · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[18] META 10-K filed 2026-01-29 · FY2025 income statement · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[19] META 10-K filed 2026-01-29 · FY2025 cash flow and free cash flow · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[20] META 8-K filed 2026-04-29 · Q1 2026 segment results · 2026-04-29 · 8-K · https://www.sec.gov/Archives/edgar/data/0001326801/000162828026028364/meta-03312026xexhibit991.htm

[21] META 10-Q filed 2026-07-30 · Q2 2026 segment results · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[22] META 10-Q filed 2026-07-30 · Q2 2026 income statement · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[23] META Q2 2026 earnings call 2026-07-29 · AI engagement and ad performance · 2026-07-29 · earnings-call · https://investor.atmeta.com/

[24] META 10-K filed 2026-01-29 · FY2025 advertising drivers · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[25] META 10-Q filed 2026-07-30 · Q2 2026 cost line drivers · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[26] META 10-K filed 2026-01-29 · cost of revenue and depreciation FY2025 · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[27] META 10-Q filed 2026-07-30 · investment philosophy and RL outlook · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[28] META 10-Q filed 2026-07-30 · leases and contractual commitments · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[29] META 10-K filed 2026-01-29 · competition · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[30] META 10-K filed 2026-01-29 · ad targeting and EU regulation · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[31] META Q2 2026 earnings call 2026-07-29 · stated risks and new revenue lines · 2026-07-29 · earnings-call · https://investor.atmeta.com/

[32] META 10-Q filed 2026-07-30 · H1 2026 cash flow statement · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[33] META 10-Q filed 2026-04-30 · Q1 2026 cash flow and depreciation · 2026-04-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026028526/

[34] META Q2 2026 earnings call 2026-07-29 · guidance and capacity planning · 2026-07-29 · earnings-call · https://investor.atmeta.com/

[35] META 10-K filed 2026-01-29 · 2026 capex outlook and commitments · 2026-01-29 · 10-K · https://www.sec.gov/Archives/edgar/data/1326801/000162828026003942/

[36] META 10-Q filed 2026-07-30 · youth-related litigation · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[37] META 10-Q filed 2026-07-30 · EU consent model and WhatsApp API · 2026-07-30 · 10-Q · https://www.sec.gov/Archives/edgar/data/1326801/000162828026050705/

[38] The Motley Fool 2026-08-06 · Meta Platforms Saw Free Cash Flow Plunge 91% to Just $784 Million · 2026-08-06 · The Motley Fool · https://www.fool.com/investing/2026/08/06/meta-platforms-free-cash-flow-plunge-ai-spending/

[39] Business Insider 2026-09-28 · Muse is supposed to make money when you use it to shop · 2026-09-28 · Business Insider(引述 MoffettNathanson) · https://www.aol.com/articles/muse-supposed-money-shop-happens-201339000.html