Meta Earnings Call: Muse Monetization to Arrive Earlier, Plans to Issue Non-Voting Class C Shares to Support Compute

After a long period of concern over massive capital expenditures, sentiment on Wall Street toward Meta is undergoing a subtle shift: the market is starting to reward Meta for having a genuinely implemented AI product story, rather than punishing it for spending money.

As Wall Street Journal noted, after the U.S. market close on October 21, Meta reported third-quarter revenue of $64.6 billion, up 26% year-over-year, beating analysts' expectations of $63.3 billion. The company expects fourth-quarter revenue to reach $73 billion to $76 billion, with the midpoint of the range above market expectations.

Then, during a one-hour earnings call, management spent the vast majority of the time answering the questions the market cares about most right now: How exactly does the Agent make money? How will the high computing costs be covered? And to support this massive system, where will Meta's money come from?

Management disclosed that in the U.S. and Canada alone, more than 10 million people have set up the Muse personal Agent; among these users, about 25% use it on any given day. Zuckerberg revealed on the call:

On the business model, management provided a relatively clear framework this time: most users use it for free; in transaction scenarios, merchants are planned to pay the fees; individual users who need a stronger model or more tokens can choose a subscription; enterprise users can pay for additional features.

Evercore pressed in the Q&A: Meta has typically prioritized user growth and retention before considering monetization, but the Muse service costs are clearly higher—does this time require earlier profitability? Zuckerberg gave a direct response: commercialization will be pushed forward earlier than in the past, both the costs and the scale of the opportunity are large, users are already using Muse to shop, and a charging model needs to be established accordingly.

To support this unceasing arms race in computing power, Meta threw out a major decision in its earnings report that directly affects its capital structure: issuing new non-voting capital stock (Class C shares), distributed to all existing shareholders in the form of a stock dividend, effectively equivalent to a "1-for-2" stock split.

Although management claims "there are no specific financing plans to announce at this time," the market knows full well that, against the backdrop of full-year capital expenditure guidance of up to $135 billion to $145 billion, this is building the scaffolding in advance for future massive equity financing. As the relevant special committee meeting records imply, credit conditions for mega-cap tech companies are tightening, issuing massive debt could hurt credit ratings, and equity financing has become the inevitable choice.

Since launching about six weeks ago, Muse has demonstrated astonishing penetration speed and user stickiness. Management disclosed that in the U.S. and Canada alone, more than 10 million people have set up the Muse personal Agent; among these users, about 25% use it on any given day.

Meta CEO Mark Zuckerberg made no secret of his surprise at this data:

The most common use case for Muse is shopping; more than a quarter of people using Muse have engaged in shopping, with an average purchase amount exceeding $200. Plaid told us they had never seen activity on any other service as active as what they are now seeing on their connector. And in the first week, we had to work with Stripe to raise credit limits because transaction volume exceeded all of our expectations.

Faced with real money transaction flows, Meta decided to break with the traditional internet path of the Facebook and Instagram era—"first stake out territory to build massive users, then slowly pursue commercialization."

When asked by Wall Street investment banks whether it would push monetization earlier than in the past, Zuckerberg gave an extremely direct answer:

I think we will push commercialization on this issue earlier than our usual playbook.

He further explained the two major reasons for the acceleration:

First is the scale of the numbers we are discussing—whether from the perspective of cost or opportunity, it makes progress on commercialization extremely important in order to scale up; on the other hand, the large amount of activity people are engaging in is itself commercial activity. In order to deliver the consumer value people want, we must build the relevant business model.

Specifically, Meta laid out a clearly tiered commercialization map:

Merchant commissions: Build business tools and charge merchants based on facilitated transactions.

Subscription fees for premium consumer users: Users who want to use the most advanced models or additional tokens can pay for a subscription directly.

Charging enterprise customers: Business owners who need additional features to grow their business, or companies that want employees to use Muse at work, will pay subscription fees.

Issuing non-voting Class C shares: the long-term capital game behind the de facto "1-for-2 split"

To support this unceasing arms race in computing power, Meta threw out a major decision in its earnings report that directly affects its capital structure.

Each existing Class A or Class B shareholder will receive one Class C share, which carries no voting rights, effective January 19 next year, with trading beginning in late January. The economic effect is equivalent to a 1-for-2 split, but management explicitly stated that the purpose of establishing Class C shares is financing, employee compensation and acquisitions.

Goldman Sachs analyst Eric Sheridan directly pressed on the financing plans during the call; CFO Susan Li said there are currently no specific financing arrangements to announce, and that the company simply believes it is more prudent to set up the capital structure framework in advance.

But the SEC filing provides clearer context. According to the PRE 14C filed by Meta, on May 29 management told the special committee that additional financing pressure was expected to rise substantially; on June 11, financial adviser Moelis relayed that all three long-term forecasts discussed by the board in May would each require raising hundreds of billions of dollars within a few years; by early August, management's language had shifted to the company "urgently needing equity financing in the near term."

The filing also explains why bonds were not chosen: debt of that magnitude would damage the credit rating, and subscription for the bond issuance of Meta's Texas data center joint venture in July was already weaker than comparable offerings; Moelis's judgment was that credit conditions for mega-cap tech companies are tightening.

Although management claims that "there is currently no specific financing plan to announce," the market knows full well that, against the backdrop of full-year capital expenditure guidance of as much as $135 billion to $145 billion, this is laying the scaffolding in advance for a massive future equity financing.

As the relevant special committee meeting minutes imply, credit conditions for mega-scale tech factories are tightening, issuing massive debt could hurt the credit rating, and equity financing has become the inevitable choice. Susan Li explained this in a very objective and neutral way:

We believe this share class will maintain our founder-led structure, it will expand our financing instruments and provide us with flexibility to support the unprecedented momentum we are seeing across all parts of our business driven by AI... Establishing an appropriate capital structure framework to ensure we have access to diversified sources of funding is the prudent thing to do.

Computing cost is key to scale; Meta expects efficiency to improve more than tenfold next year

Muse's commercialization prospects are directly tied to computing costs. Meta management has repeatedly emphasized that AI agents are currently highly compute-intensive, and that to cover hundreds of millions or even billions of users, computing efficiency must be raised and infrastructure capacity expanded in tandem.

Zuckerberg said the company is building computing capacity capable of serving more than 1 billion people in the coming years. He also noted:

"We expect its cost efficiency to improve more than 10x next year, which will allow it to be profitable once it reaches scale on the revenue line. If compute efficiency continues to improve rapidly over the next few years, then Muse will be highly profitable for many years."

Li added that the company is improving efficiency in several directions, including using off-peak hours or idle computing capacity, adopting smaller models, and optimizing the inference process and improving GPU utilization. Over the long term, the company also hopes to leverage advances in model capabilities to accomplish more tasks with fewer tokens, and to use AI to improve the efficiency of the software that runs the models.

Zuckerberg also said:

"I expect that a year from now we will be able to deliver equivalent intelligence in a smaller model at a fraction of the cost of the current Muse Wonder or any current frontier model."

Zuckerberg said Meta wants to advance revenue growth and cost optimization while expanding its user network, "and we plan to push quite aggressively." He also said Meta believes it has the conditions to bring personal agents to billions of people, but the business still needs to continue refining its products, computing capacity and business model.

In addition, Reality Labs, Meta's long-standing "profit drag," has also reached an inflection point. Because of changes to the release cadence of VR devices and a shift in focus toward prioritizing delivery of personal superintelligence (such as AI smart glasses), the division's losses are expected to narrow.

Susan Li said: "We expect operating losses to narrow to about $15 billion next year (2027) and to trend toward profitability from there."

Disrupting the B2B ecosystem: from "human-computer interaction" to "agent-to-agent conversations"

In the enterprise (B2B) market, Meta is equally ambitious. Management disclosed that at least 500 million self-identified small businesses currently use Meta's family of apps. Susan Li painted a highly disruptive commercial future:

"We expect the market will eventually no longer have agents navigate every digital experience like a human. Over the long term, this will shift to more permissioned, structured interactions... direct communication between a consumer's personal agent and a business's own agent."

In short, future e-commerce and enterprise services may no longer involve consumers personally browsing web pages, but rather "the consumer's Muse" directly negotiating and transacting with "the enterprise's AI agent." As Susan Li put it:

When summarizing Meta's strategic cadence, Zuckerberg expressed amazement at the current pace of evolution:

"It took only about eight months for Muse to go from an idea on a whiteboard to a successful product... Things have accelerated recently. At Meta, we believe the future belongs to everyone, and putting power in people's hands is how we get there."

The full transcript of Meta's Q3 2026 earnings call is as follows:

Mark Zuckerberg, Founder, Chairman and Chief Executive Officer:

Hello everyone. Thanks for joining today. The past few months have been busy and exciting. Our family of apps now has 3.64 billion people using them every day, and we're seeing record engagement on both Facebook and Instagram, driven by strong growth in video.

Threads' monthly active users have now surpassed X and continue to grow rapidly. WhatsApp's message volume has hit an all-time high, and it is becoming an increasingly important platform for agents. Over the past few months, we have reached several major milestones toward our goal of "personal superintelligence for everyone." Our Muse personal agent is one of the most successful launches ever.

We released Muse Wonder as our first frontier model, and launched several new glasses and smart devices with Muse integration, including coming soon to our new audio glasses. We have been discussing the vision of personal superintelligence for more than a year. We want to help everyone have an extremely capable personal agent that understands your goals, can help you achieve them, and improves your life in whatever way you want—whether it's improving relationships, helping save time, saving money, starting a business, creating new things, or anything else you can think of. Muse is the product form of this vision.

So we now have the foundational technology to deliver this vision. There is a lot of work ahead, but our starting position is very strong. In just a month and a half, more than 10 million people in the US and Canada have set up a Muse personal agent.

Growth continues to be strong, and we are the number one app in the app stores almost every day. Engagement and retention have also had a strong start. About 25% of people who set up Muse use it every day. For the earliest adopters, the longer they use Muse, the better retention seems to get.

To put this in perspective, this is more than twice the level Threads was at when it started, or what we believe other AI apps were at when they started. The most common use case for Muse is shopping. More than a quarter of Muse users engage in shopping, with an average purchase amount of more than $200. Plaid told us they have never seen this kind of activity on their connectors for any other service. In the first week, because transaction volume exceeded all of our expectations, we had to work with Stripe to raise our credit limit.

Beyond shopping, the most common connectors are for personal productivity and for connecting Muse to other services like email and calendar. This shows the importance of the Muse Secure VM we built to provide state-of-the-art privacy and security. We look forward to launching the Muse Confidential VM soon, which will make it so that even Meta cannot see your content. We have also seen Muse Wonder further improve quality, engagement, and retention, which shows that Muse will perform better as we continue to improve intelligence. Now that Muse has clear product-market fit, we are entering a new phase, a more operational phase focused on scaling. The goal of the next phase is to fully leverage many of Meta's strengths. We need to do three things: scale the product, grow revenue, and optimize compute costs. We expect to grow Muse to serve billions of people.

This is both our mission and our strategy. As a mission, we believe in putting the power of personal superintelligence in everyone's hands. We believe this is the path to ensuring a positive future for all, with individual empowerment as the source of prosperity and a balance of power as the foundation of safety.

At the values level, we believe that delivering this capability broadly is the best thing we can do. This is also our strategy, because we believe that billions of Muses working together in a network at scale will be able to accomplish more for people and businesses. Therefore, we want to create a large-scale Muse network as quickly as possible. From a business perspective, the average revenue per US user of our existing family of apps is several hundred dollars per year, and we expect Muse to reach a higher level over time given the volume of commercial activity we are already seeing.

This will take some time, but we are starting now. Our business model will have several components. For most people, the product should be free, and we will build commercial tools that charge based on the transactions we help facilitate.

We expect this to be paid by businesses. For people who want the most advanced models or to use extra tokens, they can subscribe and pay directly. For business owners who want extra features to grow their business with Muse, or companies that want their employees to use Muse, they can also subscribe and pay directly. For those who want Muse to help improve the personalization and quality of content and ads across Meta's platforms, they can also connect to those experiences.

To clarify, the average revenue I talked about earlier is the family-of-apps revenue corresponding to existing Muse users. On the compute side, we are building capacity to serve more than 1 billion people in the coming years. In the near term, Muse is very compute-intensive to run, but we expect its cost efficiency to improve by more than 10x next year, which will make it profitable once the revenue line reaches scale. If compute efficiency continues to improve rapidly in the coming years—which we think is highly likely—then Muse will be highly profitable for many years.

Recently, the strength of our core business has allowed us to scale the Muse network before profitability, and we believe we are the best company to pursue this strategic position. Another point about the compute economy is that, since training is a huge fixed cost, and since we have leading models, we want to serve as much high-margin inference demand as possible. That is why we are building the Meta Enterprise Platform led by CJ Desai. Muse Wonder, as well as the even more advanced models we are training, can serve additional high-margin services outside the Muse network, including our API, Muse Code, and other enterprise services.

A fundamental feature of our business is that we have many great uses for compute. It is hard to predict exactly how the Muse network, the enterprise platform, and Muse Wonder and beyond will accelerate the growth rate of our core business. But by building multiple ways to deploy these advanced models, we increase the chance that one or more of them significantly exceeds expectations. Beyond serving other enterprises, we have not yet found an upper limit to how effectively compute can be applied to accelerating our own core business.

We have implemented an evaluation using Muse Wonder to improve its effectiveness at autonomously improving our recommendation and ads systems. As our upcoming models become more advanced, I expect these systems to improve at an accelerating rate. OK, now on to the models.

Muse Wonder is state of the art across many dimensions that are critical to personal superintelligence, which validates the importance of training models ourselves to deliver the best experience. In addition to Muse, we have also seen that Meta AI's per-person engagement has more than doubled since we began rolling out new models in April. While Muse Wonder is our most important release, I am very pleased with our execution speed here. Over the past few months, we have released Muse Spark 1.2 and 1.3, Muse Code, two open-weight models Muse Glimmer and Muse Spark Open, and Muse Voice. Looking ahead, as the field advances rapidly, we plan both to push the frontier of intelligence and to distill these models into smaller, more efficient product models optimized for personal agents. My view is that we will be able to build superintelligence through scaling, so we are methodically climbing this ladder. We are already working on larger models after Muse Wonder, and they will be very impressive and capable. We are also developing new smaller models, coming soon, that will deliver Muse Wonder-level intelligence in a more efficient way.

Next, our work on personal superintelligence is now also the focus of Reality Labs. Muse will come to Meta Glasses starting next month, which will be an important step toward this goal. People already love many features of the glasses—music, calls, capturing moments, hearing assistance, and more. But over time, we expect Muse, and having personal superintelligence with you all day, to become the most important part of the experience.

Muse is also central to the glasses business model. Again, our mission and strategy are to serve everyone. We believe in putting the power of this technology in everyone's hands. So we are pricing the devices not to make large profits on hardware, but to build a large software and superintelligence business on top of them over time.

Glasses sales continue to grow rapidly. Now with Muse, we expect paid subscriptions to increase, and this business to begin to improve meaningfully from here. I also shared at Connect that we are changing our approach to virtual reality, releasing new devices only when there is a major breakthrough, such as this year's Meta VR glasses, which of course is the first-ever VR glasses experience.

I still believe that the path to delivering holographic experiences, blending the physical and digital worlds into everyday glasses, will come first through VR, but we will pace our investments in this area more moderately to prioritize delivering personal superintelligence. So, taking these factors together, we expect Reality Labs' P&L to improve from here, with operating losses dropping to about $15 billion next year and trending toward profitability from there. Before I close, I want to take a moment to talk about the amazing speed we are seeing. An important part of our approach has always been to move fast, so we can learn quickly and better serve people at scale.

But recently things have accelerated. Over the past year, we released Edits, which has already surpassed 50 million active users. New apps like Forum and Seller, built by small teams, are scaling quickly to serve advanced users. Muse went from an idea on a whiteboard to a successful product in only about eight months.

We are about to release Muse Charm, which will go from concept to a hardware device in people's pockets in less than a year. I am proud of what we are building. People have many different views of the future, but at Meta, we believe the future belongs to everyone, and putting power in people's hands is how we get there. I think that matters.

There needs to be a company that represents these ideas. But it's not just about ideas, it's also about execution. So I'm grateful to everyone who has worked so hard to deliver all of this to billions of people around the world. And as always, thank you to everyone who is on this journey with us.

Susan Li, Chief Financial Officer:

Thanks, Mark, and good afternoon everyone. Let's start with segment results. Unless otherwise noted, all comparisons are year-over-year.

Third quarter total Family of Apps revenue was $64.2 billion, up 26%. Third quarter Family of Apps advertising revenue was $63.1 billion, up 26% on both a reported and constant currency basis. Note: the original text here reads "$63.1 million," which is clearly inconsistent with the context and appears to be a typo. In the third quarter, the total number of ad impressions delivered across our services grew 10%. Impression growth was healthy in all regions, driven by engagement and user growth, as well as ad load optimization.

The global average price per ad grew 14% year-over-year, driven by improved ad performance and an improvement in the macro environment relative to the third quarter of last year. This was partially offset by strong impression growth, particularly from lower-monetizing surfaces and regions. Other Family of Apps revenue was $1.1 billion, up 59%, driven primarily by WhatsApp paid messaging and subscription revenue. In the Reality Labs segment, third quarter revenue was $443 million, down 6%, due to lower Quest headset sales, partially offset by continued strong growth in AI glasses revenue.

Now turning to consolidated results. Third quarter total revenue was $64.6 billion, up 26% on both a reported and constant currency basis. Third quarter total expenses were $51.6 billion, up 68% year-over-year, including a $9.9 billion charge related to the August 2026 multistate attorneys general agreement. The year-over-year increase was driven primarily by legal-related costs, infrastructure costs, and third-party AI token costs.

Infrastructure cost growth was driven by higher third-party cloud spending, depreciation, and data center operating costs. At the end of the third quarter, we had more than 66,000 employees, down 12% from the second quarter following the May 2026 layoffs. Third quarter GAAP operating income was $13 billion, down 37% year-over-year, representing a 20% operating margin. Excluding the third quarter legal expenses related to the aforementioned legal settlement, our third quarter operating income would have grown 11% year-over-year.

The tax rate for the quarter was 24%, above our outlook range, primarily due to the non-deductible portion of the legal expenses. Net income was $9.9 billion, or $3.83 per share. Capital expenditures, including principal payments on finance leases, were $34.9 billion, driven by investments in servers, data centers, and network infrastructure. Free cash flow was negative $2.1 billion.

We ended the quarter with $82.4 billion in cash and marketable securities and $83.7 billion in debt. Now turning to business performance. There are two main factors driving our revenue performance: our ability to deliver compelling experiences to our community, and the efficiency with which we monetize that engagement over time. On the first, we continue to see significant gains from our content recommendation initiatives.

On Instagram, time spent per user globally reached an all-time high, growing double digits year-over-year in the third quarter, with both Reels and Feed performing strongly. On Facebook, Reels time spent reached an all-time high both globally and in the US, and grew double digits year-over-year in both. Threads also continued to scale, with time spent per user growing more than 20% year-over-year. Engagement growth across our services continues to be driven by our improvements in the relevance, personalization, and timeliness of the content we recommend.

Our investments in retrieval and ranking recommendation engines on Instagram and Facebook continue to produce positive results. For example, on Instagram Reels, our relevance launches in the third quarter increased Reels time spent by more than 4%, thanks in part to our upgraded model architecture, which ranks a series of Reels at once rather than evaluating each Reel independently, as well as improvements we made to incorporate richer watch-time signals. On Facebook Feed and Reels, we completed the global rollout of our new generative retrieval architecture, which is inspired by LLMs and uses semantic representations to tag content segments and generate latent recommendations, helping our systems surface relevant new posts before they accumulate significant engagement. This new architecture is now rolled out to about a quarter of unconnected Feed views and half of Reels views, and has increased Reels watch time by 1%.

Our recommendation work is also helping deliver fresher, more personalized content. Nearly 45% of recommended Instagram Reels are now less than a day old, up more than 11 percentage points from the second quarter. On Facebook Reels, we doubled the amount of history our ranking models can consider, increasing US Reels watch time by 1%. LLMs are also helping us understand content more deeply. We continue to process public Instagram posts every day through our LLM-based content understanding systems.

By the end of the quarter, these signals increased engagement with Instagram search results by 1%. During the quarter, we also began migrating our content understanding models to our first-party Muse model family. In offline evaluations, these models outperformed our previous production models and reduced the cost per unit of content processed across Instagram, Threads, Facebook, and ads by more than 80%. In addition to improving quality, having more of this stack gives us greater flexibility to develop models around the signals that matter most for personalization. Moving to the second driver of revenue performance: improving monetization efficiency. We see a significant opportunity to continue improving the performance of businesses that use our services. In our ads system, we continue to drive performance gains by deploying more predictive models that understand users more deeply and deliver ad experiences that better match their unique preferences, while also making our infrastructure more efficient. Last quarter, we launched Meta Generative Recommender, which uses LLMs to reason about ad content and user preferences to predict the most suitable ad for each person.

In the third quarter, we launched this generative retrieval approach on Facebook Feed, increasing conversion rates for target segments by nearly 2%. In addition, we expanded the GEM model used for ads ranking to more services/surfaces, increasing conversion rates for off-site conversion campaigns by 1% to 1.7% and app ads conversion rates by 9.6%. We continue to scale our retrieval systems efficiently. In the third quarter, improvements to Andromeda, such as scaling model complexity and throughput, increased Facebook conversion rates by 1.4%.

We are also using AI to make it easier for businesses to manage campaigns, develop creative, and capture performance opportunities. Meta AI Business Assistant is becoming a more proactive and capable partner for advertisers. We estimate that advertisers who adopt its recommendations drive nearly 5% more conversions at the same cost per conversion, and increase their ad investment by 9% compared with similar non-adopters. We are rolling out more capabilities, including creative insights, stronger business context, and more agentic actions.

These features will enable marketers to plan, create, and optimize campaigns, and to schedule tasks and reminders through a conversational experience, while keeping them in control through approvals and previews. I'll share one example of Meta AI Business Assistant empowering an advertiser. Simply Organic Maple is a family-owned maple syrup producer in Vermont that used Meta AI Business Assistant to improve its campaign messaging. After Business Assistant suggested testing longer, more descriptive ad copy, the business saw a 57% increase in purchases and a 36% decrease in cost per purchase. The strongest revised ad achieved a 3x return on ad spend, 52% higher than the original. Adoption of our generative AI ad creative tools also continues to expand.

In the third quarter, nearly 10 million advertisers used at least one of these tools. Advertiser adoption of our video generation features grew 25% quarter over quarter, thanks in part to pre-made videos, which cut generation time from a single image upload to just a few seconds. We are also expanding opportunities at the intersection of creators and ads. Collaborative ads, which allow advertisers to run ads together with creators, brands, and other businesses, continued to scale rapidly in the third quarter. Our collaborative ads product grew more than 97% year over year, with an annual revenue run rate of more than $15 billion.

In the third quarter, we also completed the global rollout of Creator Marketing Hub to all advertisers, bringing creator discovery, content organization, and collaborative ad activation together in one place, with more than 10 million creators now discoverable. We are integrating Meta AI Business Assistant into the Hub so that advertisers can search for creators conversationally, identify content for collaborative ads, and get guidance throughout their creator marketing workflows. Beyond ads, we are also building more ways to monetize engagement across our ecosystem. In September, we announced the global expansion of Meta One, our unified subscription service for consumers, creators, and businesses. Meta One is rolling out gradually, with more than 50 features across Instagram, Facebook, and WhatsApp, including Meta Business Agent and Meta AI, and has surpassed 15 million subscriptions and trials to date.

The core experiences of our apps and Meta AI will remain free, while subscriptions offer more specialized capabilities and expanded AI usage. In addition to monetizing engagement in our existing services, we are also building new experiences and revenue streams around consumer agents. Mark highlighted Muse's impressive early momentum. Users have found many reasons to use Muse, with the top three categories being shopping, personal productivity, and creativity.

We are encouraged by the breadth of use cases, especially shopping, with more than 25% of Muse's weekly active users using shopping. While our near-term priority is expanding adoption and extending Muse's reach, we see multiple paths to monetization, starting with subscriptions. We also see a significant opportunity to extend Muse to small businesses. Based on a recent user survey, we estimate that at least 500 million self-identified small businesses use Meta's family of apps to reach customers and generate revenue. Muse can help these businesses acquire customers and operate more efficiently, while also helping people complete transactions. Over time, consumer and business agents can connect and transact directly, creating opportunities around paid usage, commerce, and enterprise services. In addition to growth, we are also focused on building sustainable unit economics. Muse's compute intensity means that expanding its user base requires significant investment, and we expect total cost of service to rise.

At the same time, we expect compute cost efficiency to improve more than tenfold next year, driven by better utilization of off-peak capacity, more efficient models, and inference optimization. As revenue scales, these efficiency gains give us confidence in our ability to build a highly profitable business over time. Now turning to the financial outlook, we expect fourth quarter 2026 total revenue to be in the range of $73 billion to $76 billion. Our guidance assumes an approximately 1% headwind to year-over-year total revenue growth from foreign exchange based on current exchange rates.

Turning to the expense and capital expenditure outlook, we are raising our expense outlook to incorporate the $9.9 billion in legal proceedings-related expenses recognized in the third quarter. We now expect full-year 2026 total expenses to be in the range of $175 billion to $179 billion. Excluding the third quarter legal expenses, we continue to expect 2026 operating income to be higher than 2025 operating income. At the segment level, we now expect Reality Labs losses to decrease from 2025 to the $18 billion to $19 billion range.

We expect 2026 capital expenditures, including principal payments on finance leases, to be in the range of $135 billion to $145 billion, narrowed from our prior outlook of $130 billion to $145 billion. Assuming no change in the tax environment, we expect our fourth quarter 2026 tax rate to be between 15% and 17%. In addition, we continue to monitor active legal and regulatory matters that could materially affect our business and financial results. For example, we continue to see scrutiny of teen-related issues in multiple markets, with several teen-related trials scheduled in the U.S. in the coming months.

Looking beyond 2026, we are still finalizing next year's budget and capacity plans, but we expect to continue investing aggressively to support our core business and emerging AI initiatives. Our current capacity and planned capacity coming online give us the flexibility to allocate resources among model training, core business growth, and new product development and support. Our AI investments are already driving substantial incremental advertising revenue through stronger engagement, better recommendations, and more efficient monetization. Building on this progress, our MSL investments are creating the technology and infrastructure needed to unlock significant new revenue opportunities and strengthen capabilities across the business.

While these investments will take longer to generate returns than core business investments, the breadth of the monetization opportunity and internal efficiencies give us confidence in the long-term potential. Looking ahead, while we are not providing a full-year 2027 revenue outlook, the benefits our AI investments are bringing to the core business, combined with emerging opportunities from model advancements and product launches, position us to maintain strong overall revenue growth throughout 2027. Additionally, we remain focused on improving Reality Labs' profitability.

As Mark mentioned, we expect losses to narrow to approximately $15 billion next year as we reallocate investment to other priority areas, primarily AI. Finally, today we announced plans to create a new class of non-voting capital stock to be distributed to all existing shareholders through a stock dividend. For each existing share held, shareholders will receive one new non-voting share, effectively equivalent to a two-for-one stock split. We expect trading to begin in late January 2027. In summary, our business momentum continued in the third quarter, with strong execution on our core advertising and engagement initiatives, and we are also excited about the progress we are making in bringing personal superintelligence to everyone. Krista, let's open it up for questions.

Q&A Session

Operator: Thank you. We will now open the line for questions. (Operator instructions) Our first question comes from Brian Nowak of Morgan Stanley. Please go ahead.

Brian Nowak: Thanks for taking my question. I have two questions, one on Muse and one on the enterprise business. Maybe first, congratulations on Muse's success, but as you analyze user behavior and the product, can you talk about the future user growth slope constraints you are most focused on solving, whether capability, model efficiency, capacity, etc., so we can understand the forward-looking user growth constraints you want to address? The second is on the enterprise business. Mark, can you walk us through your vision for Meta's enterprise product offerings and the key value drivers you want to provide to businesses that adopt the service?

Mark Zuckerberg, Founder, Chairman and Chief Executive Officer: Yes, I think I can talk about those two, and then Susan can add. Okay, so for Muse right now, one point I made is that we have more than 10 million people in the US and Canada who have set it up, but I think one of the biggest constraints so far is that, because this is a compute-intensive service, we haven't rolled it out globally yet. Part of that is wanting to optimize for each place, and there is some work to do around connectors.

But part of it is that we want to make sure we do this efficiently. So there are pros and cons. The upside, of course, is that it makes the 10 million people using and signing up in the first six weeks even more impressive, because those people are all in the US and Canada and concentrated in a few markets. If we actually rolled out more broadly, you would expect that number to be significantly larger than what we see today.

But on the other hand, we actually have to do the work to get efficiency to a reasonable level. So that is an important part we need to do. The compute efficiency work will be important. Also because we want to make sure we have built enough compute capacity to serve what we think is more than 1 billion people, if I can hit some of the compute efficiency targets. But obviously we have a lot of work to do there as well, and that will be one of the limiting factors next year. We want to build a business around it. I talked about this, and I think doing that while lowering compute costs or improving compute efficiency is important to making this successful from a business perspective, and I am optimistic about it.

But I think that is one of the things we need to do. Then there is continuing to increase product value, which I think we do mainly by continuing to develop new models. That is why I mentioned that Muse first launched on Muse Spark, which is currently the default model on the free tier. We launched Muse Wonder, which people can access on today's paid tier.

We know it performs much better. So we are working to get a more efficient version of Muse Wonder so we can roll it out more broadly. But you can imagine that as we iterate forward, two or three models later, when you have significantly higher intelligence and the ability to solve more problems and can deliver that at a more affordable price point, I think that will unlock more value for people.

So we are very excited about that. Turning to the enterprise part, because I think the two parts are somewhat connected, the main enterprise products in the near term will be the model API and Muse Code. We also have enterprise agents that we have been expanding for a while. All three are growing quickly.

So we are excited about that. Over time, we want to build a full Muse agent commerce network. There are also enterprise agents, which we talked about on previous earnings calls, helping businesses with sales and customer support. But we see a lot of people actually using Muse today to help run their businesses, especially small businesses.

We think there is an opportunity to help create an agent that people use to automate and complete many of the difficult and time-consuming parts of starting a small business that are not the core things people want to do. We think that is a big deal. And then making a version of Muse that people can bring into different business environments and work there, we think that will also be a big deal. Beyond that, building what I call the Muse network, where your Muse can not only collaborate with friends and family, but if you want to buy something, instead of just navigating to that business's website, you talk to their agent. I think making all of this work well together will unlock enormous value at scale. So I am very excited about that. I think enterprise use cases will be an important part of it, but it will also bring a very valuable utility boost to consumers. So I am very excited about both.

Susan Li, Chief Financial Officer:

I want to add to Mark's comments on enterprise. AI is creating enormous disruption in the enterprise space today. From what we can see, we really do not think traditional enterprise companies have locked up this market. In fact, some of the leading companies in this space today did not exist a few years ago.

Those companies made large investments in model development, and we have done the same, and we have increasingly competitive models that can benefit all users, including small businesses and enterprises. Already today, 500 million businesses use our products. So we really believe AI will reshape how businesses operate. We expect businesses of all sizes to use AI agents in software development, productivity, sales, marketing, customer service, and more. We have differentiated assets to serve this market, combining models, infrastructure, global distribution, and long-term relationships with businesses.

We want this to become our new diversified recurring revenue stream that can support multiple business models, including subscriptions, usage-based APIs, and other revenue lines such as messaging. We expect this to be a multi-year opportunity, and we intend to scale investment based on proven customer demand and product-market fit.

Operator: The next question comes from Eric Sheridan of Goldman Sachs. Please go ahead.

Eric Sheridan: Thank you very much for taking the questions. Maybe the first question builds on Brian's question: when you consider the friction points around agentic commerce, how do you as a company think about addressing the rising friction, namely that some aggregators or providers do not want agents to interact with their platforms or services; how do you think about reducing that friction, or creating the right balance between the existing services consumers access today and how agent behavior may change all of this in the medium to long term?

Mark Zuckerberg, Founder, Chairman and CEO: I can answer the first one. Okay, so, the question, sorry, what was the first question?

Susan Li, CFO: The friction points around agentic commerce.

Mark Zuckerberg, Founder, Chairman and CEO: Got it. Okay. Yes. So what we are seeing right now is that the vast majority of businesses are very excited to support this, because they recognize that this will be a new opportunity for them, not only to increase sales and improve the quality of experience for existing customers, but potentially to bring in new customers.

We will work with people on terms that make sense for them. But I do think that over time, making this work well is quite valuable for everyone. I think we are on a good trajectory so far.

Susan Li, CFO: Yes, I want to add, and I will also answer your second question. I would say agents will be an important part of how reality evolves. I think everyone will have to adapt to what this means, namely the new era of agentic commerce. We clearly have many businesses that are very excited to work with us. There are also some businesses that are understandably focused on protecting customer relationships. We expect the market will eventually move away from having agents navigate every digital experience like a human.

I think in the long run, this will shift toward more permissioned, structured interactions, including authenticated APIs, approved agent partnerships, or direct communication between a consumer's personal agent and a business's own agent. We are helping to build the standards needed to support that model. Earlier this month, we previewed the Personal Agent Protocol, an open standard across Muse and Meta business agents, which we are developing together with industry partners such as Sierra, Walmart, Stripe, and Shopify. The protocol is designed to define how consumer personal agents can interact with business agents safely and reliably, while preserving appropriate control for businesses and customers.

So adoption may be uneven and will take time. In the near term, we expect many experiences will still be AI-assisted rather than fully autonomous. But as trusted integrations develop and agents prove they can generate incremental value and demand and operate safely, we expect more and more businesses will be excited to participate. Your second question was about the announcement of the new share class. I should say that we created this new share class with the goal of creating publicly tradable non-voting stock that can be used for financing, employee compensation, acquisitions, and other general corporate purposes, while preserving the voting rights of existing Class A and Class B shareholders.

We believe this share class will maintain our founder-led structure, expand our financing tools, and give us flexibility to invest behind the unprecedented AI-driven momentum in every part of our business. We currently have no specific financing plans to announce, but we believe it is prudent to establish the appropriate capital structure framework to ensure we can access diversified sources of funding.

Operator: The next question comes from Doug Anmuth of JP Morgan. Please go ahead.

Douglas Anmuth: Okay. Thank you for taking the questions. I have two. First, I want to follow up on your last comments, Susan, about the evolution of Muse and more businesses adopting agents, and I want to understand how things change and shift as we move toward more permission-based and agent-to-agent transactions, and what that really unlocks over time? Second, on expenses and capex, thinking about 2027, I know you are not really there yet, but looking back at some earlier comments, you talked about maximizing compute in 2026 and 2027, and then having more flexibility in 2028.

A year ago, you talked about the dollar increase in 2026 being significantly greater than in 2025. Then you also used the phrase annual operating income growth. So I am curious whether any of that can be connected or help us think about 2027? Thank you.

Susan Li, CFO: Thank you, Doug. On your first question, how the agent ecosystem evolves, I wish I could predict it more perfectly; I think it will take some time. Again, I think it will be somewhat uneven, but I absolutely think that when I use Muse today, it is clear that many parts of the experience would be better if Muse could interact more seamlessly with the business's agent on the other side. For example, if Muse has to call a business on my behalf, I think that call would be easier for the business if it were handled by a business agent representing the business.

I also think one thing that is truly unique about Muse and makes the product experience so good is the degree of initiative it shows on my behalf, right? I can give it a path, I can give it a task, and Muse will execute that task for me, and even if it hits obstacles or cannot find the right thing, it needs to forge a new path to help me look for what I want.

So I think this dimension of proactivity that consumer agents or enterprise agents exhibit will cause more transactions to happen or a broader range of activities to happen than today. Today more actions are mainly taken when latent demand is expressed, whereas I can have Muse continuously search for things for me, regularly update what I'm looking for, and send me 10 new books every month. Agents should do similar things for enterprises. I think this creates an opportunity for a new, more proactive framework for what agent-to-agent interactions look like over time. On your second question, 2027, we are still early in formulating next year's budget and capacity plans.

So, I would say we are in the stage of evaluating many investment proposals with attractive return profiles across different time horizons. We want to figure out how to fund as many proposals as possible within the overall budget and financial targets framework. But this is very dynamic and still early. So before the Q4 earnings call, we don't have anything more specific or quantitative to share, which is usually when we share guidance for the next year.

Operator: The next question comes from Justin Post of Bank of America. Please go ahead.

Justin Post: Great. Thank you. One for Mark and one for Susan. Mark, you now have this powerful new model, and it seems there is a lot that can be done internally as well as externally, especially for the advertising business, where you have already re-architected the systems around AI. How much room for improvement is there in deploying this new model? How are you thinking about it? And then for Susan, you mentioned strong revenue growth next year. I want to know whether this is advertising, or whether it includes some incremental AI revenue next year? Thank you.

Susan Li, CFO: I'd be happy to answer both.

Mark Zuckerberg, Founder, Chairman and CEO: Go ahead.

Susan Li, CFO: Okay. On your first question, I think our advertising business has benefited for many years from our continued investment in models we used to call traditional AI machine learning, and of course also from the recent generation of models. I would say there are three themes worth outlining.

First is the traditional performance improvement roadmap in our pipeline, which I would more call routine machine learning work, and this continues to be a robust and healthy pipeline. Looking ahead to 2027, again, we are not providing guidance, but I would say our investments and product improvements in engagement and advertising initiatives give us confidence that we have meaningful room to grow the core business throughout 2027. This also relates to your second question, but I should say that basically I think the ranking and recommendation pipeline is healthy.

Second, I would say we are excited about some new research bets, as we shift to a more LLM-oriented AI infrastructure, to help us reason more deeply about people's preferences and interests in organic content and recommendations. We expect we will also be able to unlock benefits from this research for our ranking and recommendation engines. Third, we expect that generative AI tools, which are already very valuable to advertisers, will continue to grow. We mainly provide capabilities to advertisers through the Advantage Plus suite, making it easier for advertisers to quickly and seamlessly generate creative, or take an image and generate a video or generate voice, all of which serve to make ads increasingly personalized for the recipient. So I would say we feel there is meaningful room across all three areas. Your second question was about next year's revenue growth.

So, as I mentioned, I think we all feel good about the budget investments in next year's core advertising and engagement initiatives. We are also building several new AI businesses that we expect can begin contributing incremental revenue in 2027, but more importantly, can scale meaningfully over time. These will include monetizing our AI products through subscriptions, transactions, and advertising, and developing enterprise products, as we talked about in an earlier question. So I would say many of these opportunities are still early, and our near-term focus is really on building high-quality products and achieving product-market fit and scaled product-market fit. I think our track record in reaching billions of people and building businesses at scale makes us feel good that AI will become a meaningful source of revenue diversification over time.

Operator: The next question comes from Mark Shmulik of Bernstein. Please go ahead.

Mark Shmulik: Yes. Thank you for taking the questions. Mark, after Connect, it feels like we are entering a new era for the company, with the business repositioned around Muse as the center of everything. Is that a fair assessment of how things are progressing internally? And if so, can you talk more about some of the business mandates and how the results might change? Then Susan, interestingly, on the release cadence, the enterprise announcement came before some of the SMB announcements. I appreciate some of your comments on the SMB agent opportunity, but can you share more details on the work required to service and scale the SMB product? Thank you.

Mark Zuckerberg, Founder, Chairman and CEO: Yes, I can briefly talk about the first one. You see, more than a year ago, I think I wrote a letter about personal superintelligence as the north star for our work here. It basically fits with the line of work our company has done for more than 20 years, focused on empowering individuals, giving people a voice, and giving people the tools to connect in the ways they want.

I think it will become more pronounced in the coming years. So these are long-term bets. I think that's what we're here to do: make long-term technology bets and design the best end-to-end experiences for billions of people around the world. I think as long as we keep doing that, and as long as we keep executing well, the business will take care of itself over time, and I'm very proud of how the team has performed on this front.

Susan Li, Chief Financial Officer: Thank you, Mark. Your second question was about Muse for small businesses. So we launched it in late September, and you can think of it as an extension of Muse. It's not a standalone product, but rather a new set of skills and connectors within the existing Muse app, designed to help SMB owners run their businesses and find new customers.

With just a few clicks, it can connect to a business's Instagram account insights, Facebook Page or Meta ad account, as well as dozens of third-party connectors the business already uses, such as Shopify, QuickBooks, Canva, and more. It already understands what the business sells, its brand tone, and what customers are asking. Muse can take actions, including analyzing ad and content performance, drafting campaigns and customer replies, and flagging issues. We are actively listening to SMB users to understand their main use cases.

We see them using it for ad optimization, organizing direct messages, expanding their sales footprint, keeping their websites up to date based on inventory and new products, and managing employee workload balancing. So we're watching fairly closely how businesses use Muse to make sure we're thoughtful about the new features we release in Muse, so that it's maximally useful and efficient for SMBs. I'd also say we have unique touchpoints — again, I mentioned in my opening remarks the 500 million self-identified small businesses. We think many of the investments we've made in consumer products will carry over to help empower these businesses.

Operator: Our next question comes from Ross Sandler of Barclays. Please go ahead.

Ross Sandler: Great. Mark, you mentioned this earlier, but could you elaborate a bit on some of the high-level metrics around the per-user cost structure of Muse today, and how you see the cost curve coming down in the future? What are you doing to bring that cost curve down? Thank you.

Susan Li, Chief Financial Officer: Mark, feel free to jump in if you think I miss anything. You know, we don't have specific numbers to share on the cost structure of Muse today, but of course, as Mark said, it's quite compute-intensive, and making sure we focus on compute efficiency will be one of the key vectors for both scaling Muse and growing it into the business we want it to be. I think there are several dimensions to achieving efficiency. I think some of the nearer-term pieces include using off-peak or idle compute capacity to run certain jobs, and using smaller models.

There's a lot of work to do across different inference efficiency initiatives, really all aimed at improving GPU utilization and other product efficiency efforts. I think over the longer term, we expect that as LLMs scale, we'll be able to accomplish more with fewer tokens, and also rely on AI models themselves to help us improve the efficiency of the software we use to run the models. So I think we both feel there's a lot of work on the efficiency roadmap that will help us bend this curve, and expect that this arc of AI-driven technological progress will help us achieve greater savings over time.

Operator: Our next question comes from Mark Mahaney of Evercore ISI. Please go ahead.

Mark Mahaney: Great. Thank you. Two questions. Over the years, Meta has had a playbook for launching new interfaces: first build user growth and engagement, then layer in monetization.

Do you think it's a similar playbook here, or does it have to be faster because the costs associated with Muse are significantly higher than those for interfaces launched in the past? And then just to clarify the commerce data you provided, the roughly 25% of users and the average selling price of over $200 — does that include commercial travel purchases, or is it just retail data points? Thank you.

Mark Zuckerberg, Founder, Chairman and CEO: My guess is we'll start monetizing earlier than the normal playbook, for a few reasons. One is the scale of the numbers we're talking about, including both the costs and the opportunity, which I think makes it important that we make progress in order to scale this. The other is that a lot of what people are doing is commerce. To answer your question, yes, that does include travel. I think we might view travel as a different case, but travel is actually one of the other top use cases for Muse — basically people planning trips and planning all sorts of things they might get through our family of apps.

So I think it's important for the consumer value we provide that people want. And in doing so, I think we'll be able to build a model where we capture a percentage of the value we help people create. So I think it's very important to make all of this math work so that we can continue the compute buildout we want in order to bring this to everyone at scale. In addition to everything Susan said in the previous answer, I'd add that we're only seeing this trend: year over year, I expect that a year from now we'll be able to deliver equivalent intelligence in smaller models at a fraction of the cost of the intelligence of the current Muse Wonder or, frankly, any current frontier model.

So, if this represents some kind of baseline level that people seem to think is sufficient for product-market fit for the Muse product, I think we'll make progress on efficiency through continued model development. What's less clear, of course, is that I don't think most people will want to use the current level of model intelligence a year from now, because that will progress too. So I think we need to work hard on all of these things, including the revenue side and scaling different aspects, as well as advancing model development and all the efficiencies, so that we can both fund and support the compute buildout and reach billions of people, which, again, we want to do as quickly as possible.

Part of the reason we're not charging everyone from the start is that we want to build this Muse network. We think Muse is more valuable to people when many people and businesses are using it. That way they can collaborate and work together. I think that's really an important state, and we're very excited about it — again, I don't think there are many companies in the world as well positioned as we are to do this. I think we're very well suited to do it. We plan to move fairly aggressively.

Operator: That concludes the Q&A session. Mr. Bevegni, I'd like to turn the call back over to you.

Rob Bevegni, Director of Investor Relations: Thank you for joining today. We look forward to speaking with you again soon.

Operator: Ladies and gentlemen, that concludes today's conference. Thank you for your participation, and you may now disconnect. The event has ended. Risk Disclosure and Disclaimer

Markets carry risk, and investing requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. They invest at their own risk.