Meta Stock Price Prediction: $850 Bull vs $500 Bear (META)

Our Meta stock price prediction starts by correcting the obvious reading of the past fortnight. META has rallied 16.3% since 31 August, and the popular version of the story is that Muse, the personal AI agent Meta launched on 8 September, has finally given $130-145bn of annual capital spending a revenue line. The tape says something narrower. At the 14 September close of $665.60, Meta Platforms is up 0.8% for 2026 and still 14.7% below its record close of $780.25 on 18 September 2025, according to stockanalysis.com daily data. The rally has returned the stock to where it started the year, not to where it stood before the AI spending cycle began. Our new 12-month levels are $850 bull, $720 base and $500 bear. That range sits above the $700 / $585 / $450 we published on 1 September, and this piece sets out which assumptions moved and which did not.

Here is the number that no coverage of the Muse rally has put on the table. Between the 8 September close of $613.48 and the 14 September close of $665.60, Meta added $52.12 a share, or roughly $132.8bn of market value on the 2.548bn shares outstanding at 30 June. At the stock’s multiple of 19.0x the Street’s FY2027 earnings estimate, that value capitalises about $7.0bn of annual net income. Assume, generously, that a subscription agent keeps half of each dollar after inference costs, then apply the 15-17% tax rate Meta has guided, and you need about $16.6bn of revenue a year. At $20 a month, that is 69m people paying every month; at $100 a month, 13.8m. Meta has disclosed no Muse subscriber figure, and the product is live in the US only. Anyone crediting the whole move to Muse is implicitly forecasting one of those numbers within a year or two. That is the bar the rally set, and it is our FinanceFeeds calculation, not company guidance.

What actually moved Meta stock in September

Muse went live on 8 September as a free tier plus paid plans at $20 or $100 a month, CNBC reported, and Meta says it works in the Muse app or directly in WhatsApp. It is available to users in the US and is designed to handle tasks such as online shopping, booking travel and organising calendars on a user’s behalf. The market’s verdict came the next session: META closed at $653.69 on 9 September, up 6.55% on 35.2m shares, about 98% above its three-month average volume, The Motley Fool recorded. QQQ fell 0.29% that day.

The second leg came on 14 September. Meta rose 2.71% to $665.60 while QQQ slipped 0.80%, a gap of 3.5 percentage points on a down day for the Nasdaq-100. Goldman Sachs analyst Eric Sheridan reiterated a Buy rating and a $725 target on 14 September, according to the stockanalysis.com ratings feed. FinanceFeeds covered why the AI trade had turned nervous that week in our analysis of what a deliberate slowdown would do to $700bn of capex; Meta trading against that tape is itself information.

Having tracked every Meta scenario page FinanceFeeds has published since June, the pattern in the price history is unusually clean. The three worst sessions of the past 12 months all landed the day after a results filing: 30 October 2025 (−11.33%), 30 April 2026 (−8.55%) and 30 July 2026 (−7.95%). The best session, 29 January 2026 (+10.40%), was also a results day. By contrast, the day the multistate child-safety settlement was announced produced a 1.07% up-close. This stock moves on guidance, not on lawyers. The 9 September session was its biggest one-day gain since 1 July, and the first non-earnings move of that size since the Q2 print.

Management had flagged the product in July. “We’re developing new personal agents that will be the foundation for our next wave of products and revenue lines in the months and years ahead,” Mark Zuckerberg, Chief Executive Officer at Meta Platforms, said on the Q2 2026 earnings call.

How Meta, Wall Street and insiders responded

Meta’s pitch for Muse leans hard on trust, which is the right emphasis for a company that agreed to pay nearly $17bn to a coalition of state attorneys general less than two weeks before the launch. In a video message, Zuckerberg addressed the obvious objection directly: “We built a Secure Credential Store for your passwords and credit cards so Muse can’t read this information. Muse also checks directly with you before taking sensitive actions like making payments or sending messages,” he said, as reported by Yahoo Finance. Alexandr Wang, Meta’s AI chief, said the app runs within “its own isolated environment” inside Meta’s computing infrastructure, according to CNBC. TechCrunch framed the launch as Meta’s biggest consumer AI bet to date, and one that requires more user trust than social media ever did.

The sell side moved quickly. Mizuho, which rates the stock Outperform with a $750 target, wrote: “We believe Meta’s Muse consumer AI agent marks the beginning of a substantial product cycle for Meta that is not priced into shares.” KeyBanc, Overweight with a $780 target, said it continues “to believe the market underestimates Meta’s AI positioning and product cycle.” Both notes were quoted in the same Yahoo Finance report. The aggregate picture from 40 analysts is an average target of $767.20, which is 15.3% above the 14 September close; the lowest published target, $580, is 12.9% below it.

Insiders sold into the move, though on pre-set schedules. Chief Product Officer Chris Cox sold 20,000 shares on 9 September for about $13.0m, Chief Operating Officer Javier Olivan sold 5,354 shares on 8 September for about $3.3m, and Chief Accounting Officer Aaron Anderson sold 3,240 shares on 3 September for about $2.0m. Each Form 4, including Cox’s filing, states the sales were made under Rule 10b5-1 plans adopted between November 2025 and May 2026, so they tell you about scheduling, not conviction.

The rest of the organisation has been cutting while it builds. The May 2026 reduction of roughly 8,000 roles, which FinanceFeeds detailed in our coverage of the Meta layoffs, cost $1.18bn in Q2 severance and is one of the levers management has to keep 2026 operating income above 2025’s, as the company has promised.

The numbers under the Meta stock price prediction

Start with what is not in dispute. Q2 revenue grew 28% to $60.80bn, ad impressions rose 14% and the average price per ad rose 12%, per the Q2 release. Meta guided Q3 revenue to $61-64bn, which is 19% to 25% growth on the $51.24bn it reported for Q3 2025. The advertising engine is compounding.

The cost base is compounding faster. Total costs and expenses rose 55% to $42.03bn in Q2, including $2.40bn of legal charges. Depreciation on property and equipment reached $6.00bn in the quarter against $4.28bn a year earlier, a 40% jump, per the 10-Q. With $50.92bn of capex spent in the first half, Meta must spend $79bn to $94bn in the second half to land inside its $130-145bn range. That spending arrives in the 2027 income statement as depreciation, which is why our bear case is built on 2027 rather than 2026.

The valuation is not stretched on consensus numbers. At $665.60, META trades on 21.3x the Street’s FY2026 EPS estimate of $31.20 and 19.0x the FY2027 estimate of $34.96, per the stockanalysis.com consensus. The same page shows analysts expecting revenue of $254.2bn this year and $305.9bn next. Our FY2027 base EPS on 1 September was $27.40, which was 21.6% below that consensus. The Q3 guide did not support a discount that wide, and that gap, more than Muse, is why the base case rises.

Here is how the call has changed in two weeks, side by side.

For context on how the market treats the same trade-off elsewhere, compare our Microsoft MSFT prediction at $675 bull and $400 bear, and the July episode FinanceFeeds described as Microsoft’s $175bn capex rewarded, Meta’s $145bn punished. The difference then was free-cash-flow conversion, and it still is.

Short-dated money is more cautious than the analysts. On Polymarket, the contract asking whether META will trade at $700 at any point in September was priced at 43.2% on 15 September, and $720 at 21.5%, on the “What will META hit in September 2026” market. Spreads were wide, so read those as rough odds rather than precise ones.

The structural tension: a pipeline balance sheet with a software multiple

The regulatory overhang has narrowed but not closed. The multistate settlement that CNBC put at nearly $17bn covers the attorneys general coalition. New Mexico sat outside it and won its own case: the state’s Department of Justice says a court ordered Meta to pay $942m and overhaul its protections for children. A second matter is larger on paper. Meta’s own 10-Q says trial in the New Mexico Attorney General’s case, “which has expanded to include various claims related to content moderation issues, is scheduled to begin on September 8, 2026,” and that the New Mexico Attorney General “has indicated that they intend to seek up to $62.85 billion in penalties.” We could not confirm from a primary source whether that trial opened on schedule, and Meta still faces personal-injury and school-district suits.

The deeper tension is financial. Meta repurchased no shares in the first half of 2026, leaving $25.03bn of authorisation untouched, while it raised $24.91bn from a May bond issue. Its 10-Q discloses residual value guarantees of about $28bn on a Louisiana data-centre venture and up to about $13bn on a planned El Paso venture, plus about $68bn of new data-centre leases signed in July on 18-20 year terms. That is the liability profile of a pipeline operator, and pipelines do not trade on 19x forward earnings. Either the AI spend produces software-like returns, or the multiple drifts towards the balance sheet.

Management has declined to say how big 2027 gets. “We aren’t providing a specific outlook for 2027 CapEx at this time. Infrastructure planning remains highly dynamic and even this year, there are a range of outcomes embedded in our outlook,” Susan Li, Chief Financial Officer at Meta Platforms, said on the Q2 call. The power side of that build is already contracted in places, as FinanceFeeds reported in our piece on Vistra’s 4GW of nuclear under 20-year contracts to Amazon and Meta.

The call: $850 bull, $720 base, $500 bear

Horizon is 12 months, to September 2027. Each level is FY2027 EPS multiplied by an explicit multiple, so the assumptions can be checked line by line. Probabilities are 25% bull, 50% base and 25% bear, which weights to $697.50, or 4.8% above the 14 September close.

Bull, $850 (+27.7%, 25% probability). FY2027 EPS of $38.00, 8.7% above consensus, at 22.4x. This requires 2027 revenue growth near the Street’s 20%, operating-expense growth decelerating as the May headcount cut flows through, and early evidence that Muse and business agents are paid products at scale. Meta disclosing a subscriber base in the tens of millions within a year would put this case in play. Note what it does not need: a return to the 2025 record multiple.

Base, $720 (+8.2%, 50% probability). FY2027 EPS of $34.00, a touch under consensus, at 21.2x. Advertising keeps compounding, Muse contributes little to reported revenue in the window, and the 2027 capex figure lands high but without a shock. The stock re-tests the $720-750 area it held in the weeks after the September 2025 record.

Bear, $500 (−24.9%, 25% probability). FY2027 EPS of $27.50 at 18.2x. Depreciation from the 2026 build hits the 2027 income statement in full, expense growth stays above 25%, Q2’s $784m of free cash flow turns negative for the year, and a 2027 capex outlook above this year’s range compresses the multiple. $500 is below the $525.72 closing low of 27 March 2026. The Street is already partway there on cash: the stockanalysis.com consensus models 2026 free cash flow at about −$7.5bn.

What would change our mind. A Q3 print above the top of the $61-64bn range with a 2027 capex framing close to 2026 levels moves us towards the bull case. A Q3 print below $61bn, or a 2027 capex outlook materially above $145bn with no matching revenue line, moves us towards $500. Any adverse outcome in the second New Mexico matter would also shift the probabilities. Our earlier pieces on Nvidia’s $365 bull and $145 bear setup cover the supplier side of the same capex cycle.

FAQ

What is the Meta stock price prediction for the next 12 months?

FinanceFeeds’ Meta stock price prediction sets a $720 base case, an $850 bull case and a $500 bear case over 12 months, against a $665.60 close on 14 September 2026. Each level multiplies an FY2027 EPS estimate ($34.00, $38.00, $27.50) by an explicit multiple. The probability-weighted value is $697.50, about 4.8% above spot.

Why did Meta stock jump in September 2026?

Meta rose 6.55% on 9 September after launching Muse, a personal AI agent with $20 and $100 monthly plans, and another 2.71% on 14 September while QQQ fell 0.80%. The late-August multistate child-safety settlement had already removed one legal overhang. Even so, META is up only 0.8% for 2026.

How much is Meta spending on AI infrastructure in 2026?

Meta guides 2026 capital expenditures, including finance-lease principal payments, of $130-145bn, against $72.2bn in 2025. It spent $50.92bn in the first half, so $79-94bn remains for the second half. The Q2 10-Q also lists $349.31bn of non-cancelable commitments, with $81.65bn due in 2027.

What could push Meta stock down to $500?

The bear case assumes depreciation from the 2026 data-centre build lands in full in 2027, expense growth stays above 25% and full-year free cash flow turns negative. A 2027 capex outlook above this year’s range could compress the multiple to about 18x on $27.50 of EPS, which is roughly $500.

What do Wall Street analysts expect for Meta stock?

Forty analysts tracked by stockanalysis.com had an average Meta price target of $767.20 on 14 September 2026, with a median of $763, a low of $580 and a high of $1,000. Goldman Sachs reiterated a $725 target on 14 September; Mizuho sits at $750 and KeyBanc at $780.

This article is analysis and commentary, not investment advice. Scenario levels and probabilities are FinanceFeeds estimates and may be wrong. Share prices can fall as well as rise, and your capital is at risk. Figures are as of the 14 September 2026 close unless stated otherwise.