Meta's partnership ads, in which advertisers run ads with creators, brands and other businesses, grew more than 97% year over year and passed $15 billion in annual revenue run rate, chief financial officer Susan Li told analysts today on the company's third-quarter 2026 earnings call. The call added figures that Meta's press release and presentation left out: nearly 10 million advertisers used at least one generative AI creative tool in the quarter, more than 10 million people in the US and Canada have set up Muse personal agents, and Meta intends to charge businesses a fee on transactions that Muse helps complete.

In Short

Meta's bosses spent an hour after the results explaining where the money came from and where it is going. They said ads made with creators have become a $15 billion-a-year business and that almost 10 million advertisers now use Meta's AI tools to make ads, while the new Muse assistant is being set up by millions of people who use it mostly to shop. Meta plans to make money from Muse by charging businesses a cut of the sales it helps make, but nobody on the call mentioned the $25 billion New Mexico penalty ordered the same day.

The call and the transcript

The conference call followed the results that Meta Platforms published on October 21, 2026. PPC Land covered those documents earlier the same day in Meta takes $9.89bn teen settlement charge as ad revenue gains 26%, which reported that revenue rose 26% to $64.60 billion and that "Average price per ad rose 14% worldwide and 23% in the US and Canada, while ad impressions grew 10%, the slowest rate in the five quarters Meta charted."

This article is based on a third-party transcript of the call published by Investing.com, which states at the end that it "was generated with the support of AI and reviewed by an editor." It is not Meta's own transcript. Where the text appears to contain transcription errors, they are flagged below rather than corrected. The transcript carries no publication date; the call itself refers to the results announced "today," and an analyst refers to the share class announced "tonight with earnings."

Four people from Meta's side speak. Krista, the conference operator, ran the queue. The director of investor relations opened and closed the call; the transcript labels him "Rob Bhavani" in each speaker heading, while the operator addresses him as "Mr. Bhavnani," so at least one of the two spellings is wrong. Mark Zuckerberg, chief executive officer, and Susan Li, chief financial officer, gave prepared remarks and answered questions. Seven analysts asked questions: Brian Nowak of Morgan Stanley, Eric Sheridan of Goldman Sachs, Doug Anmuth of J.P. Morgan, Justin Post of Bank of America, Mark Shmulik of Bernstein, Ross Sandler of Barclays and Mark Mahaney of Evercore ISI.

Investing.com's summary above the transcript adds market data that does not come from Meta. According to Investing.com, earnings per share of $3.83 beat a consensus estimate of $3.57, revenue beat a $63.34 billion forecast, and shares rose 3.79% in after-hours trading to $723.93 after closing at $697.43. The same summary says Meta "raised its full-year 2026 capital spending outlook to $135 billion to $145 billion, narrowing the range from a prior estimate of $130 billion to $145 billion." On the call, Li described the change only as "narrowed from our prior outlook," which matches the release: the low end went up, the top did not.

What Li said about ad prices and impressions

The release gave the growth rates. The call gave Meta's explanation of them.

"In Q3, the total number of ad impressions served across our services increased 10%," Li said. "Impression growth was healthy across all regions, driven by growth in engagement and users, as well as ad load optimizations." Ad load is the share of content slots given to ads, so the second half of that sentence means some of the extra inventory came from showing more ads per session rather than from more usage alone.

On price, she said: "The global average price per ad increased 14% year-over-year, driven by ad performance gains and improvements in macro conditions relative to Q3 of last year. This was partially offset by strong impression growth, particularly from lower monetizing surfaces and regions."

That description sits uneasily with the trend in Meta's own charts. Li called impression growth "strong," while PPC Land's results article described the same 10% as the slowest in the five quarters Meta displayed. The phrase "lower monetizing surfaces and regions" is not broken out further on the call. The presentation's regional data, as reported by PPC Land, showed Asia-Pacific impressions up 14% with price per ad up 1%, which fits that description, but Li did not name a region or a surface.

Neither executive discussed the US and Canada figure, where price per ad rose 23% on 8% more impressions. No analyst asked about ad pricing, impression growth, regional performance or the cost to advertisers. Nor did anyone on the call discuss advertiser verticals, such as retail, gaming or Chinese export advertisers, or demand from any industry group.

Other revenue: the WhatsApp answer

One gap in the documents was closed. PPC Land's results article noted that "Neither document says what drove other revenue," which rose about 59% to $1.1 billion. On the call, Li said Family of Apps other revenue "grew 59%, driven primarily by WhatsApp paid messaging and subscriptions revenue." In the second quarter the same line grew 73%, driven mainly by "WhatsApp paid messaging and subscription revenue," according to PPC Land's second-quarter coverage.

The ranking and recommendation numbers

Most of Li's prepared remarks on the business were a list of percentage gains from specific models. Taken together, they are the most detailed account Meta gave today of how it says AI translates into ad revenue. All of them are Meta's own measurements, and none came with a test design or a baseline.

Organic engagement

On Instagram, Li said global time spent per user reached an all-time high and "grew double digits year-over-year in Q3, with strength across both Reels and Feed." On Facebook, Reels time spent "reached all-time highs both globally and in the U.S., with both growing double digits year-over-year." Threads time spent per user grew "over 20% year-over-year."

The specific launches she cited were:

  • Instagram Reels ranking: relevance launches in the quarter "increased Reels time spent by over 4%," driven in part by a model architecture "that ranks sequences of Reels at a time instead of evaluating each Reel independently," and richer watch time signals.
  • Facebook generative retrieval: Meta completed the global launch of a retrieval system that, "inspired by LLMs, labels pieces of content with semantic representations and generates potential recommendations," so that new posts can surface before they accumulate engagement. Li said it now covers "roughly one-quarter of unconnected feed views and half of Reels views, and increased Reels watch time by 1%."
  • Freshness: "Nearly 45% of recommended Instagram Reels are now less than one day old, up more than 11 percentage points from Q2."
  • Facebook Reels history: Meta "doubled the amount of history our ranking model can consider, driving a 1% increase in Reels watch time in the U.S."
  • Instagram search: LLM-based content understanding "improved the engagement with Instagram search results by 1% through the end of this quarter."

The transcript reads "We continue to process public Instagram posts each day through LLM-based content understanding systems." The sentence has no quantity, and its construction suggests a number was spoken and lost in transcription.

Li added that Meta began moving its content understanding models onto its own Muse model family during the quarter. "In offline evaluations, these models are outperforming our prior production models and are reducing per-unit content processing costs by more than 80% across Instagram, Threads, Facebook, and ads," she said. The qualifier matters: offline evaluations are tests on stored data, not live results.

Ad systems

For advertising, Li named four systems:

  • Meta Generative Recommender: introduced in the second quarter, it "uses LLMs to reason across ad content and user preferences to predict the best ad for each person." Launched as generative retrieval on Facebook Feed in the third quarter, it drove "a nearly 2% increase in conversions across targeted segments."
  • GEM: Meta expanded its ads ranking model to more surfaces, "driving a 1.7% increase in conversions for off-site conversion campaigns and a 9.6% increase in app ad conversions."
  • Andromeda: improvements to the ad retrieval system, "such as scaling model complexity and throughput, increased conversions by 1.4% on Facebook."
  • Meta AI Business Assistant: "We estimate that advertisers who adopted its recommendations drove almost 5% more conversions at the same cost per conversion and increased their advertising investment by 9% compared with similar non-adopters."

The GEM numbers are far smaller than the figure Meta's engineers published eight days earlier. On October 13, Meta said its GEM ads model lifted conversions 18.9% across its apps in a technical post. The call figures describe incremental gains from expanding GEM to additional surfaces in one quarter, not the model's cumulative effect, so the two numbers measure different things, but neither the post nor the call reconciles them. In the second quarter, Li had credited GEM and user understanding models with "a 15.7% uplift in conversions on Facebook," according to PPC Land's second-quarter coverage.

The Business Assistant figure is a comparison between advertisers who chose to adopt recommendations and "similar non-adopters." That design measures a correlation among self-selected accounts. The 9% increase in advertising investment among adopters also means the tool is associated with higher spending on Meta, which is part of the commercial case for it.

Li gave one case study. Simply Organic Maple, described as a family-run maple syrup producer in Vermont, followed the assistant's recommendation to test longer, more descriptive ad copy. According to Li, "the business saw a 57% increase in purchases and a 36% reduction in cost per purchase. Its strongest revised ad delivered a 3x return on ad spend, a 52% improvement from the original." No time frame or spend level was given.

She also said the assistant is gaining "creative insights, stronger business context, and more agentic actions," which will let marketers "plan, create, and optimize campaigns as well as schedule tasks and alerts through a conversational experience with approvals and previews that keep them in control."

Creative tools, creators and Advantage+

PPC Land's results article observed that "Neither document names a specific advertising product" and that "There is no mention of Advantage+, of ads on Threads or WhatsApp, or of advertising inside Muse." The call changed only part of that.

Li said: "In Q3, nearly 10 million advertisers used at least one of these tools," referring to generative AI ad creative tools. The figure continues a series. PPC Land's first-quarter coverage reported "over 8 million advertisers." For the second quarter, PPC Land's report quoted Zuckerberg on "9 million small businesses" using at least one tool, a different unit from advertisers, so the three figures are not strictly comparable. Advertiser adoption of video generation features "grew 25% quarter-over-quarter," Li said, "driven in part by ready-made videos that reduce generation time to just a few seconds from a single image upload."

The partnership ads number was the largest new figure. "Our partnership ads product grew over 97% year-over-year and surpassed $15 billion in annual revenue run rate," Li said. Partnership ads are paid placements that run under a creator's or partner's identity with the brand alongside. PPC Land's first-quarter coverage reported a run rate of $10 billion in the first quarter, so by those two figures the run rate rose by about $5 billion in two quarters. A run rate annualises recent revenue; it is not a reported revenue line, and Meta does not report partnership ads separately in its financial statements.

Li also said Meta "completed the global rollout of Creator Marketing Hub to all advertisers," which brings "creator discovery, content organization, and partnership ad activation into one place with more than 10 million creators now discoverable." Meta is integrating its Business Assistant into the hub so advertisers can "conversationally search for creators."

Advantage+ came up once, in the question-and-answer session. Asked by Justin Post how much room remains to improve the ad business with the new model, Li said generative AI tools "will continue to grow in terms of the features that we make available to advertisers, primarily through the Advantage+ suite," including features to "take an image and generate video or generate sound, all, I think, in service of making the ads increasingly personalized for the person that they're being served to."

Ads on Threads and WhatsApp were not discussed. Zuckerberg said in his opening remarks that "Threads has now overtaken X in monthly actives and continues to grow quickly," without a number. PPC Land's second-quarter coverage reported 500 million monthly actives for Threads; X's figure was not given. Meta extended Threads ads to users in every market from the week of January 26, 2026, but neither executive mentioned Threads ad revenue. WhatsApp appeared as a messaging platform "hitting all-time highs in messaging" and "becoming an increasingly important platform for agents," and as the main driver of other revenue.

Meta One subscriptions

Li said Meta announced in September the global expansion of Meta One, which she described as "our unified subscription service for consumers, creators, and businesses." According to Li, it 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 figure combines paying subscriptions and trials without separating them. "The core experience across our apps and Meta AI will remain free, while subscriptions offer more specialized capabilities and expanded AI usage," she said.

Muse: users, uses and how Meta plans to charge

The results documents named Muse but, as PPC Land's article put it, "Neither document reports Muse users, revenue or downloads." Zuckerberg filled that in.

"In just a month and a half, more than 10 million people in the U.S. and Canada have set up Muse personal agents," he said. "About 25% of people who have set up Muse use it on any given day." He said this was "more than twice the level that Threads started at, or where we believe that other AI apps started at as well," without giving the comparison figures.

The count matches independent estimates in size, though the measures differ. Sensor Tower estimated on October 20 that Muse passed 10 million US downloads within 40 days and put US daily active users at 2.5 million. Meta's figure counts people who have set up an agent across two countries; Sensor Tower's counts US app downloads. Zuckerberg's 25% daily rate applied to 10 million also gives about 2.5 million, by PPC Land's calculation. When Muse went live on September 8, PPC Land reported a US-only release; Zuckerberg now describes availability in the US and Canada.

Shopping is the main use. "The most common use case for Muse is shopping, with more than a quarter of people who are using Muse engaging in this with an average purchase size of over $200," Zuckerberg said. Li later gave the base as "over 25% of Muse weekly actives." Asked by Mark Mahaney whether that commerce figure included travel, Zuckerberg said: "To your question, that does include travel too."

Zuckerberg also cited partners. "Plaid told us that they hadn't seen activity like they're seeing with their connector on any other service before. In the first week, we had to work with Stripe to raise credit limits since the volumes were above what any of us expected."

The business model

The revenue plan is the part of the call most directly relevant to advertisers and merchants. "For most people, the product should be free, and we will build commerce tools so we get paid a fee based on the transactions that we help facilitate. We expect this will be paid for by businesses," Zuckerberg said. Users who want the most advanced models or extra tokens will subscribe. Business owners and companies that want Muse for work will also be able to subscribe.

A fourth component connects Muse to advertising. "For those who want Muse to help improve the personalization and quality of their content and ads across Meta, they will be able to connect to those experiences as well," Zuckerberg said. He did not explain how that connection would work or whether Muse data would inform ad targeting. Li, answering Post, listed "subscriptions, transactions, advertising" among the ways Meta expects to monetise its AI products.

Zuckerberg corrected himself mid-sentence on revenue per user. The transcript reads: "the average revenue per American user for our existing Muse users, sorry, for our existing family of apps users, is several hundreds of dollars annually, and we expect that Muse will reach an even higher level over time." He then clarified: "the average revenue that I was talking about before is the family of apps revenue for our existing Muse users." The statement therefore concerns what Meta already earns from the Facebook and Instagram activity of people who use Muse, not Muse revenue.

On timing, Zuckerberg told Mahaney that Meta would "work on monetization somewhat sooner on this than our normal playbook," citing "the size of the numbers that we are talking about in terms of both the costs and the opportunity" and the share of commerce in Muse usage. "I think we will be able to build a model where we get a percent of the value that we are helping to create for people," he said. No fee level, start date or merchant terms were announced.

Small businesses and enterprise

Li said Meta estimates, "based on a recent user survey," that "at least 500 million self-identified small businesses use Meta's family of apps." In the question-and-answer session, the figure became broader: "Already today, we have 500 million businesses using our product." The first version is a survey estimate of self-identified small businesses; the second drops both qualifiers.

Answering Shmulik, Li described Muse for small business, launched in late September, as "a set of new skills and connectors inside the existing Muse app" that connects to Instagram analytics, Facebook Pages and Meta ad accounts, "along with dozens of third-party connectors that businesses already run on, like Shopify, QuickBooks, Canva, et cetera." When Meta announced the product on September 29, PPC Land reported that Muse gained access to ad accounts and 15 business tools. Li's "dozens" is higher than the 15 described at launch; the call did not say whether connectors were added since. According to Li, Muse "can take actions, including analyze ad and content performance, draft campaigns, customer responses, flag issues," and small business users rely on it "for ads optimization, for direct message organizing."

On enterprise, Zuckerberg named "the Meta Enterprise platform led by CJ Desai," which Meta set up in September; PPC Land reported at the time that Meta's new enterprise AI unit hired MongoDB's chief executive. He told Nowak that "the main enterprise offering in the near term is going to be the model API in Muse Code," alongside business agents. Li said Meta hopes enterprise becomes "a new diversified, recurring revenue stream" and that it will "scale our investment behind demonstrated customer demand and product market fit."

Agentic commerce and the Personal Agent Protocol

Eric Sheridan asked how Meta would handle "aggregators or suppliers that don't want agents to interact with their platforms." Neither executive named a company. Zuckerberg said "overwhelmingly businesses are very excited about enabling this" and "Not everyone wants to be an early adopter. That is fine." Li acknowledged that "There are also businesses that are understandably focused on protecting customer relationships." The most prominent case is Amazon, which in September blocked Muse from placing orders on its site, according to PPC Land's October 8 report.

Li said she expects the market "to evolve away from agents attempting to navigate every digital experience like a human" toward "authenticated APIs, approved agent partnerships, or direct communications between a consumer's personal agent and a business's own agent." She said Meta previewed "the Personal Agent Protocol, an open standard across Muse and Meta Business Agents that we are developing with industry partners like Sierra, Walmart, Stripe, Shopify, et cetera," earlier in October. PPC Land's searches found no earlier coverage of the protocol on the site.

One sentence in Li's answer appears garbled. The transcript reads: "I think agents are going to be an important part of how reality evolves." The context, a question about commerce, suggests a different word was spoken; the transcript does not make it possible to say which. Later in the same exchange, a sentence beginning "when more actions are taken, primarily when the latent demand expresses itself" also does not parse as transcribed.

Cost

Ross Sandler asked for per-user cost figures. "We do not have specific numbers to share around the cost structure of Muse today," Li said. Both executives repeated that Meta expects to improve cost efficiency "by more than 10x" or "more than tenfold" over the next year, through "better utilization of off-peak capacity, more efficient models, and inference optimization." Zuckerberg said compute is the main limit on growth: "because this is a compute-intensive service to run, we have not launched it globally yet."

Inference, the running of a trained model to answer requests, is the cost line in question. "Since training is a large fixed cost, now that we have leading models, we want to serve as much high-margin inference demand as possible," Zuckerberg said. Li said "we expect total serving costs to rise" as the user base grows.

Models and privacy

Zuckerberg called Muse Wonder Meta's "first frontier model" and "state-of-the-art across many dimensions that matter for personal superintelligence," without citing benchmarks. Muse launched on Muse Spark, "currently the default across the free tier," while Muse Wonder is available in the paid tier. He listed other releases from recent months: Muse Spark 1.2 and 1.3, Muse Code, Muse Voice, and two open-weights models, Muse Glimmer and Muse Spark Open. The latter was the model Meta and Apple ran on two Mac Studio computers, as PPC Land reported on October 15. He added that "engagement per person in Meta AI has more than doubled since April."

On privacy, Zuckerberg cited the "Muse secure VM" and said Meta is "looking forward to following up soon with a Muse confidential VM, which will make it so that even Meta cannot see your content." The call did not mention that, one day earlier, Meta halted Muse share links after 6,800 turned up in public DNS records.

Spending, headcount and 2027

The guidance numbers on the call match the release: fourth-quarter revenue of $73 billion to $76 billion with an approximately 1% currency headwind, 2026 total expenses of $175 billion to $179 billion, 2026 capital expenditures of $135 billion to $145 billion, and a fourth-quarter tax rate of 15% to 17%. The transcript runs some ranges together, printing "$73 billion$76 billion" and "$18 billion$19 billion," which appears to be a formatting error.

The call added a few details:

  • Tax rate: the third-quarter rate of 24% was "above our outlook range, primarily driven by the non-deductible portion of legal expenses."
  • Cost drivers: expense growth came from "legal-related costs, infrastructure costs, and third-party AI token costs," with infrastructure growth driven by "higher third-party cloud spend, depreciation, and data center operating costs."
  • Balance sheet: Li said Meta ended the quarter with "$82.4 billion in cash and marketable securities and $83.7 billion in debt." PPC Land's results article reported long-term debt of $80,922 million from the release; the call's figure is about $2.8 billion higher, and Li did not say what it includes.
  • Operating income: "Excluding the Q3 legal charges related to the aforementioned legal settlement, our third quarter operating income would have increased 11% year-over-year," which matches PPC Land's earlier calculation.

Headcount

PPC Land's results article noted that "The documents give no reason for the decline and do not mention severance." Li provided the reason: "We ended Q3 with over 66,000 employees, down 12% from Q2 following the May 2026 headcount reduction." PPC Land had calculated an 11.5% fall from 75,472 to 66,775, which rounds to Li's figure.

The timing is notable. The May reduction affected about 8,000 employees, according to PPC Land's earlier report. Yet headcount at June 30 was down only about 2,500 from 77,986 at the end of March, by PPC Land's calculation from its first- and second-quarter reports. Li's wording suggests that most departures from the May reduction were recorded in the third quarter, although she did not say so directly.

2027

Doug Anmuth asked how earlier comments, such as maximising compute in 2026 and 2027, should frame next year. "We are quite early in developing our budget and capacity plans for next year," Li said, adding that Meta is "evaluating a lot of investment proposals that have attractive return profiles over different time horizons." Specific guidance will come on the fourth-quarter call, "which is typically when we share guidance for the next year."

On revenue, Li repeated the release language that AI investments "position us to sustain strong overall revenue growth throughout 2027." Asked by Post whether that included new AI revenue, she said Meta expects new AI businesses "can begin contributing incremental revenue in 2027, but more importantly, can scale meaningfully over time." She described the ad pipeline in three parts: "business-as-usual machine learning work," "new research bets" in "a more LLM-oriented AI infrastructure," and generative AI tools for advertisers.

Zuckerberg added that Meta has "implemented an eval with Muse Wonder to improve how effectively it can autonomously improve our recommendations and ad systems," without results.

Reality Labs

Reality Labs revenue fell 6% to $443 million "due to lower Quest headset sales, which were partially offset by continued strong growth in AI glasses revenue," Li said. Zuckerberg said Muse comes to Meta Glasses "starting next month" and that Meta has "priced our devices not to make a large profit on the hardware." Meta will now "only ship new devices when we have major breakthroughs like we had with the Meta VR Glasses this year," he said. Reality Labs losses are expected at $18 billion to $19 billion in 2026 and "around $15 billion next year and trending towards profitability from there."

Zuckerberg also cited Edits, which "has already passed 50 million actives," new apps called Forum and Seller, and Muse Charm, which he said would go "from concept to a hardware device in people's pockets in less than a year." He gave no launch date for Muse Charm.

Class C shares

Eric Sheridan asked about the new non-voting share class. Li said it was created "with the purpose of creating a publicly traded non-voting stock usable for capital raising, employee compensation, acquisitions, other general corporate purposes while preserving the voting power of existing Class A and B shareholders." She added: "We do not have any specific financing plans to announce at this time." In her prepared remarks, she said each existing share will receive one non-voting share, "effectively resulting in a two-for-one stock split," with trading expected "in late January 2027." The release, as PPC Land reported, gives January 19, 2027 as the effective date of the amended certificate.

Li mentioned the settlement twice. The quarter's expenses "included a $9.9 billion charge related to the August 2026 multi-state attorneys general agreement," and Meta raised its expense outlook "to incorporate the $9.9 billion charge related to legal proceedings recognized in Q3." She repeated the release's general warning: "we continue to see scrutiny on youth-related issues in several markets and have a number of youth-related trials scheduled in the coming months in the U.S."

Nobody on the call mentioned New Mexico. Earlier on October 21, a Santa Fe judge entered a written judgment ordering Meta to pay $25 billion over Facebook privacy statements, and PPC Land reported that Meta faces 15% yearly interest on the $25bn New Mexico penalty, which amounts to about $3.75 billion a year until the sum is paid. No analyst asked about it, about the teen settlement charge, or about how either might affect cash flow already strained by capital spending. The European Union, the Digital Markets Act and European regulators also went unmentioned. Investing.com's summary lists "more youth-related trials" as a risk, and nothing more on litigation.

Why this matters for advertisers

The call was mostly about Muse. Six of the seven analysts asked about Muse, enterprise AI or agentic commerce, and only Justin Post asked about the ad business directly. That balance reflects where Meta's executives want attention, but the details they gave on advertising still matter for anyone buying on Meta's apps.

First, the ad performance figures are modest increments. Gains of 1.4%, 1.7% and nearly 2% in conversions, each on a specific surface or segment, are the building blocks Meta cites for a 14% rise in average price per ad. Li attributed that rise to "ad performance gains and improvements in macro conditions," so Meta's position is that advertisers are paying more because ads convert better. The call offered no aggregate measure to test that against.

Second, creator content is now a large, separately tracked line inside Meta's ad business. A $15 billion run rate growing at 97% means partnership ads are taking a growing share of budgets, and Creator Marketing Hub plus the Business Assistant are designed to make that easier to buy.

Third, Muse introduces a different commercial relationship with businesses. Instead of buying an ad, a merchant may pay a fee when a user's agent completes a purchase, and Meta intends that "this will be paid for by businesses." With no fee level, start date or terms disclosed, and with retailers such as Amazon already blocking the agent, how that model interacts with existing ad budgets remains open.

Timeline

Summary

Who: Meta Platforms chief executive Mark Zuckerberg and chief financial officer Susan Li, answering questions from analysts at Morgan Stanley, Goldman Sachs, J.P. Morgan, Bank of America, Bernstein, Barclays and Evercore ISI. The source is a third-party, AI-assisted transcript published by Investing.com.

What: Meta disclosed that partnership ads grew more than 97% to a run rate above $15 billion, that nearly 10 million advertisers used generative AI creative tools, and that specific models lifted conversions by between 1.4% and 9.6% on particular surfaces. Zuckerberg said more than 10 million people have set up Muse and that Meta will charge businesses fees on transactions Muse facilitates. Li attributed the headcount fall to the May reduction and other revenue growth to WhatsApp paid messaging and subscriptions.

When: The call took place on October 21, 2026, after the release of third-quarter results for the period ended September 30, 2026. Muse arrives on Meta Glasses in November, and trading in the new non-voting shares is expected in late January 2027.

Where: Muse is available in the US and Canada. Meta's results cover its apps worldwide, with its largest ad revenue in the US and Canada.

Why: The call explained the drivers behind a quarter in which ad prices rose faster than ad volume, and set out how Meta intends to pay for Muse and its AI spending. It did not address the New Mexico judgment ordered the same day or EU regulatory matters, and no analyst asked about them.