Earnings Preview | Meta (META.US) Ushers in an “AI Monetization Test”: Can the Muse Explosion Ease “Money-Burning” Concerns?

The Zhitong Finance App learned that Meta (META.US) will release financial results for the third quarter of 2026 after the US stock market on October 21. The market has repriced Meta from an “AI money burner” to an “AI monetizer,” and what this financial report wants to do is prove that this transformation is real. Since the July earnings report, Meta has accumulated an increase of about 37%, clearly outperforming the Nasdaq 100 Index by 6.7% and the S&P 500 by 4.4% — while analysts lowered earnings forecasts by about 4% during the same period. After a recent surge, the valuation seems to have fully reflected optimistic expectations.

Three numbers to look at in the third quarter

The first is growth rate. The official guide is revenue of $61 billion to $64 billion, corresponding to an increase of 19% to 25% of $512 billion in the same period last year, with a median value of about 22% — still high growth, but lower than 28% in the second quarter. Analyst median revenue expectations of $63.29 billion, adjusted earnings per share of $6.85, operating profit of $18.53 billion, EBITDA of $26.99 billion, and gross profit margin of 80.16%. The implied third-quarter operating margin was approximately 29.3%, compared to 40.1% in the same period last year. The market's revenue forecast for the fourth quarter was $73.56 billion, so guidance for the next quarter was more important than “exceeding expectations” for the current quarter.

The second is that legal accusation. On August 26, Meta and state attorneys general reached a settlement on youth safety issues, paying about $18 billion over ten years (of which about $5.3 billion depends on whether TikTok and YouTube take similar measures). Investors were relieved that the main litigation risks were cleared, and the stock price rose, but this cost was not included in the July spending forecast. Meta expects legal expenses of about $10 billion for the third quarter, which is equivalent to about $3.9 billion per share before taxes — GAAP earnings per share will be unsightly, but the market has already anticipated this.

The third is capital expenditure in 2027. This is the one that is most likely to trigger a sharp reaction in stock prices. Management declined to give a 2027 outlook during the second-quarter conference call, and the stock price fell by about 10% the next day. Currently, sellers are extremely divided on this figure: FactSet agreed to expect about $193.88 billion; Guggenheim's Michael Morris raised the estimate from $203 billion to $225 billion; and Truist's Youssef Squali raised the estimate from $18011 billion to $190 billion (up 32% year over year). Morris' judgment is that Meta will continue to aggressively expand AI infrastructure, and “we expect free cash flow to remain under pressure until 2028 and begin to resume positive growth in 2029.” Squali went even further: “Whether large-scale AI infrastructure investments can be converted into actual paid service revenue will be the key to a further rise in stock prices.”

Key Factors

The advertising engine is still the bottom line, but the cost side is already alerting

The second quarter was one of the best quarters in Meta's advertising business history: revenue of $60.8 billion, up 28% year over year (fixed exchange rate 27%), number of ad impressions increased 14%, average ad price increased 12% — volume and price rose sharply; Family App (FoA) operating profit of $23.4 billion, while Reality Labs lost $4.6 billion.

AI's contribution to this improvement is already verifiable: Instagram achieved double-digit growth in usage time, the Advantage+ ad solution reached an annual revenue scale of over $75 billion, and Facebook ad conversion rates increased 15.7%. This is the strongest argument from the bulls — AI has improved the way ads work, and advertisers are willing to pay more for it.

However, in the same financial report, the cost side has changed dramatically. Total costs increased 55% year over year to $42 billion, including $2.4 billion in legal fees and $1.2 billion in severance payments (corresponding to about 8,000 layoffs in May 2026). As a result, operating profit fell 8% to $18.8 billion, operating margin fell from 43% to 31%, and diluted earnings per share fell 13% to $6.18.

Muse: Subscriptions performed well

Muse is the immediate catalyst for this upturn. Meta launched this personal AI smart device on September 8, and its stock price rose by about 27% in September; last week, it also launched a new model, Muse Wonder, which focuses on stronger memory, concentration, and ability to handle complex tasks. The stock price rose more than 2% within five trading days.

The download data is impressive, but the caliber of each source varies greatly: according to Sensor Tower data, Muse received 5 million downloads in the US and Canada within 22 days, “far faster than ChatGPT”; another article said it had nearly 9 million downloads; other reports say it surpassed 6.6 million downloads in 16 days. Sensor Tower's other caliber: Muse had 2.8 million global downloads in the first two weeks of launch, with 3.7 million ChatGPT over the same period — according to this caliber, Muse did not outperform ChatGPT.

The more critical metric is retention. According to Sensor Tower data, Muse's 14-day retention rate was 22%, which is slightly lower than the same period ChatGPT; there is also data that more than 30% of users will return 7 days after their first visit. Guggenheim analyst Michael Morris quoted Apptopia data as pointing out that Muse reached 1.3 million daily active users in five weeks — less than 0.05% of Meta's 3.6 billion daily active user base. He expects Meta to prioritize scale-up usage rather than pursue near-term monetization.

However, monetization has only just begun: Meta began charging Muse on October 12, and along with the launch of $20 and $100 monthly plans with the new model, there is no revenue data yet. Additionally, there are reports that Muse was blocked when visiting some websites, including Amazon — which means it is quite difficult to build an intelligent body that can reliably complete tasks.

Free cash flow: The real point of contention in this financial report

If advertising is a bottom card and Muse is an option, then free cash flow is the worst page in that earnings report.

Meta's operating cash flow for the second quarter was $31.9 billion, but capital expenses (including financial leases) reached $31.1 billion, and free cash flow was only $784 million, compared to $8.549 billion in the same period last year. The picture for the third quarter is likely to be worse. Meta's annual capital expenditure guide is US$130 billion to US$145 billion, and has already spent about US$50.9 billion in the first half of the year. Based on this, it is estimated that capital expenditure for the second half of the year will reach US$79 billion to US$94 billion. If operating cash flow remains at about $32 billion per quarter in the second quarter, free cash flow for the second half of the year will be between minus $15 billion and minus $30 billion. The consensus forecast from the agency survey is more straightforward: free cash flow for the third quarter was negative $9.14 billion, and capital expenditure was $409.7 billion.

Where did the money come from? Rely on loan. Meta issued $24.9 billion in long-term debt in the second quarter, with long-term debt obligations of $83.7 billion at the end of the quarter, corresponding to $90.3 billion in cash and securities; there were no share repurchases in the second quarter ($10.2 billion in the same period last year). Chief Financial Officer Susan Li made it clear during the second quarter call that Meta intends to use a “larger percentage of debt” to finance long-term infrastructure. Depreciation is also accelerating: depreciation in the second quarter was $6.4 billion, up 46% year over year, and capital expenditure was about 5 times that — depreciation will continue to rise as new data centers are put into operation. According to forward-looking authors' estimates (16% tax rate, approximately 2.57 billion diluted share capital), earnings per share decreased by approximately $3.3 billion for every $10 billion increase in annual depreciation.

competition

A new variable has emerged at the competitive level. Meta stopped accepting ByteDance ads on Facebook, Instagram and other platforms on October 8, involving the United States, Canada, Egypt, Indonesia, Japan, Thailand, and Vietnam; the restrictions also apply to advertisements placed by third party advertisers and linked to TikTok and other ByteDance products. A Meta spokesperson was straightforward: “We are under no obligation to advertise for a competitor that aims to steal our users.” TikTok has previously removed the ability to link exclusively to Instagram from profiles.

Valuation and transaction aspects

According to StockAnalysis data, based on 2027 earnings per share of approximately $34.55 (consistent expectations), the three scenarios are: 18 times the bear market (the market begins to price Meta as a capital-intensive company), 22 times the benchmark (about the current price), and 26 times the bull market (Muse and the planned cloud business become a real source of revenue). Furthermore, Meta's second-quarter free cash flow corresponded to a yield of only about 0.2%, and the dividend ratio was only about 0.3% — the valuation doesn't seem expensive, but shareholders are currently not receiving much cash return.

Wall Street's differences are concentrated on target prices. The bearish side: Goldman Sachs lowered its target price from $860 to $762 on the grounds that operating expenses faced upward pressure in 2027 — but it also raised its revenue forecast for the next 6 to 18 months and maintained purchases; Cantor Fitzgerald maintained gains and lowered its target price from $860 to $840.

The side that is looking to monetize more: UBS raised its target price from $715 to $884. The logic is the launch of Business Agents, the increase in Muse usage, and two generative AI monetization paths to increase WhatsApp and Messenger monetization and search ad share; Guggenheim raised from $700 to $850; Nick Jones of BNP Paribas offered $885; Pivotal Research It was raised from $790 to $930; Truist and Wells Fargo both gave $1,000 (Wells Fargo also warned that AI costs could weigh down 2027 profits). Bank of America analyst Justin Post maintained a buy with a target price of $810.

In terms of stock price trends, Meta rushed all the way from the $550 range in August to around $780 in September, then entered a narrow consolidation range of about $60. On the technical side, the upper resistance is around $780—$786 (which may open space above $800 after a breakout), and the lower support is around $710-732. You can refer to the recent price path: $718.67 on October 9 → $747.72 on October 12 (up 4.04%) → $763 on October 13 → $763 on October 13 → Goldman Sachs lowered the target price on October 14, and other mixed ratings fell about 3%. On Thursday, Meta shares closed down 0.11% to $736.72.

The options market suggests that the current earnings report's event fluctuation was about ± 4.1%, while Meta's average daily absolute fluctuation for the past eight quarters was 7.4% — the two are not of the same caliber, but pricing seems relatively restrained. Another statistic worth remembering is that Meta has only closed up after earnings reports four times in the past eight quarters — exceeding expectations alone does not guarantee a rise. Pepperstone's Chris Weston therefore gave a harsher threshold: revenue reaching or exceeding $64 billion, profit margins were better than expected, fourth quarter guidance was positive, and there was concrete evidence of AI monetization to reinforce bullish logic; conversely, weak guidance or another sharp increase in capital expenditure could trigger a sell-off even if the third quarter exceeded expectations.