Meta Stock Q3 Earnings Analysis: Strong Ads, Costly AI

Meta’s advertising business has given Mark Zuckerberg more room to build his AI future. Revenue beat expectations, the holiday-quarter outlook came in stronger than Wall Street expected, and more than 10 million people have set up its Muse AI agent. 

Meta stock was up 2.17% in the after-hours snapshot following its October 21 results. But behind the encouraging headlines sits an expensive catch: profits are growing much more slowly than sales, and the year’s biggest investment bill may still be ahead.

Let’s break down Meta’s Q3 2026 earnings, what Zuckerberg revealed about Muse, and why a smaller cash shortfall does not yet mean its AI expansion is paying for itself.

Meta Q3 2026 Results: Revenue, Earnings and Cash Flow

Meta generated $64.60 billion in revenue, up 26.1% from a year earlier. Its advertising business remained the main source of income, while the legal settlement announced in August distorted reported profit.

Sources: Meta Q3 earnings release and presentation; MarketBeat; Yahoo Finance/StockStory. The $7.40 figure is Meta’s disclosed exclusion calculation, rather than reported GAAP EPS.

The result was stronger than the market’s revenue forecast and Meta’s own previous $64 billion guidance ceiling. However, the $3.83 and $7.40 EPS figures answer different questions. Comparing either one with a forecast using another accounting basis can produce a misleading earnings headline.

Why Did Meta Stock Rise After Earnings?

Three developments help explain the positive reaction in Meta stock.

• First, advertising delivered. Meta’s revenue beat was supported by both more ad displays and higher average pricing. 
• Second, the company forecast Q4 revenue of $73 billion to $76 billion. The $74.5 billion midpoint was about 1.2% above the $73.65 billion analysts expected, according to Associated Press. 
• Third, Zuckerberg supplied concrete Muse usage figures alongside his plan to reduce its operating costs.

Investors received evidence that the existing business remains strong and that the new product has attracted users. That combination improves the argument for continuing investment, even though it does not establish the eventual return.

Did Meta’s Underlying Earnings Really Grow 265%?

No. Reported EPS rose from $1.05 to $3.83, approximately 265%, but both years contain major unusual items.

Last year, a $15.93 billion one-time, non-cash tax charge depressed EPS to $1.05. Meta said EPS would have been $7.25 without it. This year, the $9.89 billion teen mental-health settlement charge reduced EPS by $3.57, leaving reported EPS at $3.83.

Sources: Meta’s respective Q3 releases; growth calculations by the author. The exclusions remove the specified tax and legal charges, rather than every possible difference between the periods.

That second row is the more useful starting point for assessing earnings growth. Meta exceeded the pre-results earnings reference, yet its earnings per share grew only modestly against a cleaner year-ago comparison.

The operating-profit calculation tells a similar story:

Sources: Meta’s financial statements; author calculations.

Removing the settlement changes the picture from falling operating profit to growing operating profit. It still leaves a lower margin. Costs excluding that charge increased about 35.9%, ahead of revenue growth of 26.1%.

The business is expanding, but each dollar of sales is leaving less operating profit than a year ago. Excusing the entire margin decline as a legal accounting issue would miss the rising cost of running and expanding the company.

What Went Right in Meta’s Advertising Business?

Advertising contributed approximately 97.6% of revenue. It remains the business financing Meta’s experimentation.

Ad impressions, meaning the number of times advertisements are displayed, rose 10%. Average revenue per ad increased 14%. Multiplying those rounded growth rates gives approximately 25.4% growth, broadly consistent with the reported advertising increase.

Think of a cinema filling more seats while earning more per ticket. Both help revenue, but the ticket price deserves closer inspection: it can reflect a different mix of customers and screenings as well as stronger demand.

Meta’s CFO Susan Li attributed pricing growth to better advertising performance and improved economic conditions, partly offset by growth in less lucrative regions and ad placements. She also reported record engagement on Instagram and Facebook Reels.

These results support a practical investment argument: AI can earn money by improving an established advertising product before a new assistant becomes profitable. Better recommendations can increase engagement, while more effective ads can give businesses a reason to increase spending.

However, advertising growth alone cannot prove that every AI project earns an attractive return. Meta has not provided a separate financial statement linking each dollar of infrastructure investment to incremental advertising profit.

What Did Mark Zuckerberg Say About Muse?

Zuckerberg’s most useful update went beyond download rankings. More than 10 million people in the US and Canada had set up Muse agents in about a month and a half, and approximately 25% of those people used Muse on a given day.

Applying that percentage to a 10 million-user base suggests roughly 2.5 million daily users. This is an indicative calculation from management’s rounded figures, rather than a separately disclosed daily-user total. It also describes daily engagement, rather than a conventional 30-day retention rate.

Muse is an AI agent, meaning software that can take several steps to complete a task. Its promise is closer to an assistant arranging a trip than a search box listing travel websites.

The Revenue Plan Has Several Parts

Source: Zuckerberg’s earnings-call remarks. These are routes management described, with different stages of development.

The current position is narrower than the full ambition. On the follow-up call, Li said subscriptions are currently Muse’s only monetisation route. Transaction monetisation requires greater scale among users and business partners.

Shopping activity can therefore demonstrate commercial intent without already generating a transaction fee for Meta. A purchase facilitated by an agent is not automatically Meta revenue.

The Profitability Plan Depends on Lower Computing Costs

Mark Zuckerberg acknowledged that Muse is expensive to operate. He expects more than a tenfold improvement in cost efficiency over the next year, which he said could enable profitability once revenue reaches sufficient scale. That remains a conditional management expectation.

Li added a useful qualification on the follow-up call: the full improvement requires further technological advances. Routine improvements alone would not deliver the entire target.

Here is why the target needs careful interpretation:

Author illustration assuming a tenfold cost-efficiency improvement for the same work. These are not Meta forecasts. Model quality, task complexity and usage can change the outcome.

Even a major efficiency breakthrough can accompany a larger total computing bill. Investors need to track revenue per unit of activity alongside cost per unit. More users and cheaper tasks create an opportunity; profitable activity has to connect the two.

Meta’s Cash Flow Beat Has a Q4 Catch

Our published Meta Q3 earnings preview asked whether advertising could finance the AI expansion. It used Reuters-reported estimates of $41.68 billion in capital expenditure and negative $8.52 billion in free cash flow.

The actual shortfall was much smaller. The explanation is revealing.

Sources: INDmoney published preview, Reuters’ October 20 estimates and Meta’s Q3 cash-flow statement. The preview’s operating-cash reference was inferred from the other two estimates, assuming compatible definitions. Differences use unrounded reported results.

In this comparison, the entire cash-flow improvement is explained by lighter investment spending, partly offset by weaker operating cash generation. Meta did not exceed the preview’s implied operating-cash reference.

That is useful relief for the quarter, but it does not establish a permanent reduction in investment needs.

The Updated Outlook Implies a Much Larger Q4 Bill

Meta spent $85.824 billion on capital expenditure in the first nine months. Its updated annual outlook is $135 billion to $145 billion. Subtraction gives the spending needed to reach that outlook:

Author calculations using Meta’s presentation and updated guidance. These are implied amounts, rather than formal quarterly forecasts.

Even the lower end requires approximately 41% more spending than Q3. The smaller Q3 cash deficit therefore comes before a potentially much heavier quarter of investment.

This is the clearest test to carry forward: does operating cash generation catch up, or does the funding gap widen as the planned infrastructure bill arrives?

What Does Meta’s Guidance Say About 2027?

The near-term revenue outlook is encouraging. Against Q4 2025 revenue of $59.893 billion, the new $74.5 billion midpoint implies approximately 24.4% year-over-year growth. That suggests continued momentum after Q3.

Source: Meta’s earnings release and call.

The expense increase needs context. July’s $165 billion to $169 billion outlook plus approximately $10 billion of legal expense already pointed to the new range. The revised figure should not be interpreted as an additional $10 billion AI operating-cost overrun.

Management expects strong revenue growth to continue into 2027, while acknowledging that newer AI investments will take longer to produce returns than investments in its core business. It has not supplied a numerical 2027 revenue or capital-spending forecast.

Reality Labs Offers Some Relief

Reality Labs generated $443 million in Q3 revenue and lost $4.13 billion at the operating level. Meta expects annual losses of $18 billion to $19 billion in 2026, narrowing to approximately $15 billion in 2027.

That would reduce the division’s loss by $3 billion to $4 billion. However, Meta also plans to redirect investment toward AI. The improvement should not automatically be added to group profit as if all of the spending disappears.

What Do Meta’s New Class C Shares Mean for Investors?

Meta’s board approved a new publicly listed, non-voting share class, with the corporate amendment intended to take effect on January 19, 2027. Existing investors would receive one Class C share for each Class A or Class B share, with distribution dates to be determined.

The company expects a split-like effect on its Class A share price. Existing investors’ voting power would remain unchanged.

Receiving an additional share does not double the underlying business or an investor’s proportional economic interest. It divides the existing claim across more securities with different voting rights.

The larger strategic purpose is financing flexibility. Meta can use the new class for capital raising, employee compensation and acquisitions while preserving its founder-led voting structure.

Two events should be evaluated separately:

• The initial distribution: allocates shares to existing investors and does not itself raise cash.
• Potential future issuance: can raise capital or fund other commitments, but can also dilute existing economic ownership.

Li clarified on the follow-up call that a reference to hundreds of billions of dollars in financing was a theoretical control calculation, rather than an announcement of current financing needs or plans.

Additional funding options can support growth. Their value to existing investors ultimately depends on what Meta earns from the capital, the terms of any issuance and the resulting ownership changes.

The Separate $25 Billion Legal Risk Investors Should Watch

On October 21, a New Mexico court ordered Meta to pay $25 billion in civil penalties in a privacy case connected to Cambridge Analytica. The state’s Department of Justice confirmed the order. Meta plans to appeal.

This is a separate matter from the teen mental-health agreement behind the $9.89 billion Q3 charge discussed above.

It would be wrong to mechanically subtract the headline penalty from the reported quarter’s earnings, or to assume an appeal removes the financial exposure. Further disclosures on provisions, the appeal process and payment obligations will determine how investors should incorporate it.

The broader lesson is that excluding an unusual legal charge can improve an operating comparison while leaving a real cost for investors to assess.

How Should Investors Read Meta Stock After These Results?

The quarter strengthens the case for Meta’s advertising business. It also makes the Muse opportunity more measurable: users have arrived, daily engagement exists, and management has identified revenue routes.

The harder financial questions remain open. Earnings growth after the identified exclusions lagged sales growth, the cash-flow improvement came from lower quarterly capital spending, and the annual outlook implies a much larger Q4 investment requirement.

For an existing investor, the useful question is whether new profits and cash justify the additional investment commitments. For a potential investor, it is how much unproven Muse profitability the share price already assumes.

The after-hours gain is understandable. Advertising delivered and the next-quarter outlook improved. But a lasting improvement in Meta’s investment case needs more of that growth to reach profits and cash, rather than simply financing the next increase in computing capacity.