Meta’s Friday close is not a verdict that the advertising engine failed. Nasdaq’s official daily close on Friday 23 October 2026 was $642.62, after Thursday 22 October closed at $648.85, down $48.67, or 6.98 percent, from Wednesday’s $697.52. Friday fell another $6.23, or 0.96 percent, and did not repeat the drop. The obvious reading is that investors punished a higher capital-spending floor, now $135 billion to $145 billion for 2026. The missing fact is that the new non-voting Class C share raises no cash on the day it is created. It is a stock dividend, one share for each Class A share and each Class B share, and Meta expects the Class A price to reflect a two-for-one split after payment. Chief financial officer Susan Li said the company has no specific financing plans to announce. From this Friday close, 47 sessions through 31 December 2026 put the bull at $910.82, the base at $665.95, and the bear at $540.47.
Those three prices are the distribution around Friday’s print. They are not a broker target, and they are not the $1,158.56 bull and $676.09 bear that used the 14 October close of $737.54. The cash question sits in the commitments note and in a free-cash-flow line that is already negative, not in the $5 billion lift at the bottom of the capex range.
Key facts
- Friday 23 October 2026 official close: $642.62. Thursday 22 October: $648.85, down 6.98 percent from Wednesday’s $697.52. Friday then fell 0.96 percent. Source: Nasdaq official daily closes, fetched 24 October 2026.
- The 20-day average is $717.65 and the 50-day average is $663.81. Friday sits 10.45 percent under the shorter average and 3.19 percent under the longer one. Same Nasdaq file.
- Third-quarter revenue was $64.60 billion, up 26 percent. Diluted EPS was $3.83, or $7.40 excluding a $9.89 billion legal charge. Source: Meta earnings release, 21 October 2026, exhibit 99.1.
- 2026 capital expenditures, including finance-lease principal, narrowed to $135 billion to $145 billion from $130 billion to $145 billion. Quarterly capex was $34.91 billion. Free cash flow was negative $2.12 billion. Same release.
- The board approved non-voting Class C stock effective 19 January 2027, plus a one-for-one stock dividend. Source: Form 8-K, filed 21 October 2026.
- Non-cancelable contractual commitments were $454.52 billion at 30 September 2026. Leases not yet commenced were about $355.73 billion. Source: Form 10-Q.
- Family daily active people were 3.64 billion in September, up 3 percent. Reality Labs lost $4.13 billion on $443 million of revenue. Earnings release, 21 October 2026.
- The bull, base, and bear use 47 sessions and the last 90 log-returns. TipRanks on 22 October listed broker targets from $845 to $995.
The drop is not a vote on the ad print
The quarter reported on 21 October was an advertising quarter. Family of Apps revenue was $64.2 billion, up 26 percent, and advertising was $63.06 billion of the $64.60 billion total, about 97.6 percent. Ad impressions rose 10 percent and the average price per ad rose 14 percent. Fourth-quarter revenue guidance is $73 billion to $76 billion. The midpoint, $74.5 billion, is about 15.3 percent above third-quarter revenue, with foreign exchange guided as about a 1 percent headwind.
Operating income was $12.99 billion, down 37 percent, and the margin was 20 percent against 40 percent a year earlier. A $9.89 billion charge, tied to an August agreement with 52 state attorneys general, sits inside costs. Payments can reach about $18.59 billion. Excluding the charge, operating income would have been $22.88 billion and diluted EPS $7.40. Net income of $9.86 billion rose 264 percent because a year earlier the tax rate was 87 percent. This quarter’s rate was 24 percent.
Thursday fell from $697.52 to $648.85, and Friday closed at $642.62. Wednesday to Friday was down $54.90, or 7.87 percent. Volume and the lows are in the table below. The largest down day in the last 90 sessions remains 30 July 2026, from $585.61 to $539.03, a 7.95 percent decline, with a $524.49 low. Thursday stays in that sample.
In late July the floor moved from $125 billion to $130 billion and the shares were sold while a larger Microsoft figure was not. FinanceFeeds laid out that split then. October repeats it: the floor moved from $130 billion to $135 billion, and the $145 billion ceiling did not move.
Who has to live with the new share
Written consent, not the listed holders, approved the charter. The preliminary information statement says CZI Holdings and six Chan Zuckerberg entities consented on 20 October 2026 with 315,826,951 Class B shares, about 56.1 percent of the vote. Class A holders are not being asked for a proxy. The 10-Q counted 2,214,472,367 Class A shares and 341,768,886 Class B shares as of 16 October. The record date is unset, so a dividend on that count is arithmetic.
As of 1 October, Mark Zuckerberg held about 60.6 percent of the vote. The committee wrote that he would not support a large Class A sale that cut that power, and that the debt under discussion could hurt credit ratings. The alignment agreement makes 20 percent of his Class C freely saleable, including for a potential California Proposition 40 tax if enacted, and staples the other 80 percent. For eight years in a full-time leadership role, he agreed to reside in California or wherever the headquarters sits.
Advertisers remain the cash engine. Li said nearly 10 million used a generative AI tool in the quarter, video-generation adoption rose 25 percent, and partnership ads grew more than 97 percent year over year past a $15 billion run-rate. Shopping is the most common Muse use, above a quarter of users, with purchases averaging over $200. Zuckerberg said Plaid had not seen this connector activity, and Stripe limits were raised in the first week. Li named Sierra, Walmart, Stripe, and Shopify. Headcount was 66,775, down 15 percent, and CJ Desai leads inference sales. No priced Class C sale attaches to those names.
What the tape and the filings actually show
Friday’s $642.62 is $95.87 under the prior five-session high of $738.49 on 16 October and $2.93 above Thursday’s $639.69 low. Friday’s low of $633.27 went under that mark. The close is 17.36 percent, or $134.97, under the 24 September close of $777.59. That day’s high of $779.82 is the top of the fetched 52-week range, and $520.26 on 27 March is the low. The 20-day average is $717.65 and the 50-day is $663.81, so Friday is under both and $21.19 under the longer one.
| Session | Close | Change | Low | Volume |
|---|---|---|---|---|
| Wednesday 21 October | $697.52 | Earnings released that day | $689.68 | 29,203,900 |
| Thursday 22 October | $648.85 | Down $48.67, or 6.98 percent | $639.69 | 40,571,020 |
| Friday 23 October | $642.62 | Down $6.23, or 0.96 percent | $633.27 | 14,791,600 |
Free cash flow in the release is operating cash flow minus equipment purchases minus finance-lease principal. Both readings below are negative. A year earlier the same definition was positive $10.63 billion. Nine-month free cash flow was still positive at $11.05 billion.
| Line, third quarter 2026 | Amount |
|---|---|
| Operating cash flow | $32,786 million |
| Purchases of property and equipment | $33,942 million |
| Finance-lease principal | $964 million |
| Free cash flow, company definition | Negative $2,120 million |
| Operating cash flow minus equipment only | Negative $1,156 million |
| Capex including finance-lease principal | $34.91 billion |
Nine-month equipment purchases of $83,055 million plus $2,769 million of lease principal leave about $49.18 billion to $59.18 billion of the $135 billion to $145 billion range still ahead, above the $34.91 billion just reported. The 10-Q, signed 21 October by Li and chief accounting officer Aaron Anderson, lists $454.52 billion of non-cancelable commitments, about $45.85 billion due in the rest of 2026 and $121.41 billion in 2027, plus about $355.73 billion of leases not yet started and up to $14.72 billion of contingent cloud. Those lines stay separate. Cash and marketable securities were $82.45 billion against $80.92 billion of long-term debt, including $24.91 billion from May notes. Restricted funds of $10.80 billion release from 2028 to 2030. Nine-month buybacks were zero, against $26.25 billion a year earlier, and the quarter’s dividend line was $1.36 billion.
What Zuckerberg and Li said, and what they would not date
Founder, chairman, and chief executive Mark Zuckerberg used the call for Muse. “In just a month and a half, more than 10 million people in the US and Canada have set up Muse personal agents,” he said. About 25 percent of people who set it up use it on a given day. On cost: “In the near term, Muse is very compute-intensive to run. But we expect to improve its cost efficiency by more than 10x in the next year, which will enable it to be profitable once the revenue lines are at scale.” The source is the 21 October earnings call. He did not give a 2027 capex number.
Chief financial officer Susan Li said Class C is a publicly traded non-voting stock for capital raising, compensation, acquisitions, and general purposes, preserving Class A and Class B votes, with trading expected in late January 2027. Answering Eric Sheridan of Goldman Sachs, she said: “We don’t have any specific financing plans to announce at this time but we thought it was prudent to establish the appropriate capital structure framework to ensure that we have access to a diverse pool of funding sources.”
On the follow-up call, Li said the hundreds of billions “was really a theoretical maximum” for the equity raise that would cost founder control. “It is not actually a description of our present equity financing needs or plans. And we don’t have anything to announce on that front.” She told Ken Gawrelski of Wells Fargo that compute “remains pretty constrained.” Reality Labs losses are guided at $18 billion to $19 billion for 2026 and around $15 billion next year. Fourth-quarter tax is guided at 15 percent to 17 percent.
Muse is not a segment line. Other Family of Apps revenue was $1.10 billion. FinanceFeeds has set out how wide the long-dated Muse range still is, including Deutsche Bank’s spread by 2030. That range is not an input to the year-end prices. The September note on Muse payments covers the card path. It does not fund the fourth-quarter residual.
The financing tension is the charter, not the $5 billion
The committee says AI investment is unprecedented, that raises by Alphabet, SpaceX, Anthropic, and OpenAI have made capital harder to find, and that Class C should be ready before more of it is absorbed. Management wanted it by the end of 2026. The charter is effective 19 January 2027. Offerings are expected from time to time, subject to market conditions. That is a tool. Non-voting stock also, the committee says, allows equity from sovereign institutions under US foreign-investment rules. The dividend is meant to build a market so later sales avoid a steep discount to Class A.
Eighty percent of his dividend shares sell only if, within two business days, he sells or converts matching voting stock. A breach converts Class B into Class A. While he leads, 2.5 percent of that Class C releases each quarter, and the rest after 32 quarters. Further dividends need a doubling from the 10 September 2026 Class A price of $580.60 and a 10 percent compound return. He can end the agreement if the charter misses 19 January 2027 or the dividend misses the fifth business day after. Moelis advised the committee and Morgan Stanley advised Meta. Neither is an underwriter of a launched sale.
A separate path is already open. The 10-Q describes a Louisiana venture from October 2025 and a Texas venture from July 2026, with Meta at 20 percent of each. Maximum exposure was $68.40 billion at 30 September, including $41.49 billion of residual-value guarantees. Texas formation contributed $2.26 billion of net assets held for sale and returned a $1.04 billion distribution.
Bull, base, and bear through 31 December
Count the sessions after Friday. From 26 October through 30 October is 5. November has 20 weekdays once Thanksgiving, 26 November, is removed, and the 27 November early close still counts. December has 22 weekdays once Christmas, 25 December, is removed, and the 24 December early close still counts. Five plus 20 plus 22 is 47.
The last 90 daily log-returns have a mean, mu, of 0.00075860 and a sample standard deviation, sigma, of 0.03045088, divisor 89. Base is $642.62 times e to the power of 47 times 0.00075860, which is $665.95. Bull adds 1.5 times 0.03045088 times the square root of 47 inside that exponent, which is $910.82. Bear subtracts 1.0 times sigma times the square root of 47, which is $540.47. The 30 July session stays inside sigma.
Versus Friday, the bull is up 41.74 percent, or $268.20, the base is up 3.63 percent, or $23.33, and the bear is down 15.90 percent, or $102.15. Neither rounded price sits within a dollar of a bull or bear already printed in an English FinanceFeeds title, including the $1,158.56 and $676.09 pair, so the horizon stays 31 December.
| Case | Price | Versus Friday’s $642.62 | What has to be true |
|---|---|---|---|
| Bull | $910.82 | Up 41.74 percent, or $268.20 | Forty-seven sessions realize the recent drift plus 1.5 standard deviations, with no discounted Class C sale in the window. |
| Base | $665.95 | Up 3.63 percent, or $23.33 | Only the drift. The year ends a little above Friday because mu is slightly positive. |
| Bear | $540.47 | Down 15.90 percent, or $102.15 | The same drift minus one standard deviation. A 30 July-sized session would do it. The split alone would not. |
TipRanks wrote on 22 October that Bank of America moved from $810 to $845, TD Cowen from $865 to $850, Barclays from $780 to $880, Citi from $895 to $905, and Wells Fargo from $1,000 to $995. All five sit above this base. The lowest, $845, is $202.38 above Friday. This $910.82 bull sits between Citi’s $905 and Wells Fargo’s $995 by coincidence. The title is the distribution, not those notes.
What happens next
Through 31 December 2026 the table is the call. Forty-seven sessions of this drift compound to roughly 3.6 percent before volatility. If the $73 billion to $76 billion fourth-quarter range holds and no Class C offering is filed, a finish near $665.95 is the drift path. A finish near $910.82 needs the upside shock. A finish near $540.47 needs a downside shock already inside sigma, or a priced sale.
On 19 January 2027 the charter is scheduled to take effect, and the dividend is due by the fifth business day after, unless the date slips and the agreement ends. Meta says Class A should then reflect a two-for-one split, with votes unchanged. A print near half the pre-dividend price is the unit change, not cash raised and not, by itself, the bear case.
Late January 2027 is when Li expects Class C to trade. If no primary sale is on file, the hundreds-of-billions line stays the control math she described and the year-end band still reads the stock. If a sale is filed, a new share count changes the claim and the band stops being sufficient. The dated item is 19 January.
Frequently asked questions
Where do the sessions through 31 December 2026 put Meta after the drop?
The sessions through 31 December 2026 put the bull at $910.82, the base at $665.95, and the bear at $540.47, from Friday’s $642.62 Nasdaq close over 47 sessions. The path uses the last 90 daily log-returns, not a broker target. The base is only 3.63 percent above the close because daily drift is small. The bull adds 1.5 standard deviations and the bear subtracts one. The older start was the 14 October close of $737.54.
Did the shares fall 7 percent, and did Friday repeat it?
Thursday’s official close was $648.85, down $48.67, or 6.98 percent, from Wednesday’s $697.52. That is the session behind the about-7-percent headlines. Friday’s official close was $642.62, down a further $6.23, or 0.96 percent, on about 14.8 million shares against about 40.6 million on Thursday. Friday did not repeat the drop. This calculation uses Friday, the last completed daily bar.
Does the Class C dividend raise cash?
No. The board declared one Class C share for each Class A share and each Class B share, after a charter effective 19 January 2027. Meta expects Class A to reflect a two-for-one split, and voting power does not change. Susan Li said on 21 October there are no specific financing plans to announce. A later sale of new shares would be a separate event.
What is 2026 capex, and what is free cash flow?
The 21 October release narrowed 2026 capital expenditures, including finance-lease principal, to $135 billion to $145 billion from $130 billion to $145 billion. Third-quarter capex on that definition was $34.91 billion. Free cash flow was negative $2.12 billion. Leaving out $964 million of lease principal still leaves a negative $1,156 million. Both readings are in the same reconciliation.
How do these bull and bear numbers differ from analyst targets?
They measure a different object. The $910.82 bull and the $540.47 bear are the 47-session distribution around Friday’s close. TipRanks posted on 22 October that Bank of America moved to $845, TD Cowen to $850, Barclays to $880, Citi to $905, and Wells Fargo to $995. Each sits above the $665.95 base. This bull is not Citi’s $905.
What happens to the Class A price when the dividend is paid?
Meta says that after the payment date the Class A price should reflect a two-for-one split, and that voting power does not change. A print near half the pre-dividend price would be the unit change, not by itself the bear case. The record date is not set. Li expects Class C trading in late January 2027. If the charter misses 19 January, Zuckerberg can end the alignment agreement.
This is not financial advice.
