We value META at $611 a share. The report's $641.76 price (Oct 26, 2026) was 5% above fair value: Hold.
Restated Oct 26, 2026 (method case-spread-2026-10-24); the report's original value was $611, Hold.
- Five-year return
- 8.7%
- a year, at the report's price
What would change the view
- To Buy: ad growth 0.5 point a year above our path, or evidence that Meta's capital per revenue dollar falls after the build (a long-run return on new capital of 22.6%; ours is 16.5%)
- To Avoid: ad growth 1.15 points a year below our path (2027 at 18% instead of 20% is $596.39), apps cash costs 2.5 points of revenue higher, a long-run return below about 10.3% (the build's own 10.7% gives $549.26), or a first 2027 capex guide above about $269B
- The return on Meta's new capital (the year's rise in after-tax operating income before legal charges over the net investment of the year before; 13% at Q3 2026) below 9.4% at Q4 2027 while Meta's 2028 spending guide is above the October path ($159.5B): the extra capital is not paying its way
Catalysts
- Pricing of the $22.5B forward sale and the $12.5B preferred, expected within days of the 2026-10-26 launch: the offering price, the preferred's dividend rate and conversion band, and whether the over-allotment options are taken. At $580 the central is $3.26 lower ($607.58), at $700 $2.32 higher; options -$0.40; the dividend and band move it by cents
- Meta's restated charter takes effect: each Class A and B share receives one Class C share, trading from late January 2027; every per-share figure halves. The planned, unsigned $60B at-the-market program is to start after that, in Q1 2027. Without it the central is $3.36 higher and central debt peaks at $222.6B instead of $160B
- Q4 2026 results, date unconfirmed, with Meta's first 2027 guides. We expect a first capital spending guide near $191.3B at the midpoint. A guide above about $269B (2027 spending above $308.9B at our 1.15 ratio, 2028 held there) is an Avoid even if the extra earns its cost; at $198B the central is $605.30 if it does and $555.28 if none of it does
This text argues the original call. The current figures are at the top.
Hold: Meta trades 5.1% above our $611 fair value, now well clear of the
$545.50 Avoid line (85% of the price). Nothing about the business or the
forecast changed. One input changed: the long-run return on new capital
that the long fade uses after 2031. Our four hyperscaler reports set it four
ways this morning; the owner approved one rule for all of them
(valuation.md, study
2026-10-26-hyperscaler-return-rule).
Each case now earns, on growth after 2031, its own year-five return on its
net capital: growth needs the same capital per revenue dollar the case
itself reaches in 2031. For Meta's central that is 16.5%, against the
10.7% the sixth report used (+$61.59). The rating does not change; the
distance to the Avoid line does.
1. What changed, in order
model.py, output.txt. META2_SETTINGS replays
the sixth report at $549.26; the last step sets Settings.held.
| Step | Pessimistic | Central | Optimistic | Weighted |
|---|---|---|---|---|
| The sixth report, replayed | 279.49 | 549.26 | 916.50 | 573.63 |
| Each case's long-run return holds its own 2031 capital per revenue dollar | −14.89 | +61.59 | +54.11 | 614.22 |
| This report | 264.60 | 610.85 | 970.60 | 614.22 |
The rule. The long-run return is year-five after-tax operating profit
over the case's net capital at the end of 2031: the 10-K's 2025 net
property of $176.4B (without $20.4B of operating-lease assets, which the
forecast never rolls forward), plus each year's capital spending less
depreciation through 2031. The central leaves out the $124.9B net book of
its spending above the October path, because the bridge already values
that capital apart and its profit is out of year-five profit. The returns
are 13.4%, 16.5% and 16.5% for the pessimistic, central and optimistic
cases (method.with_held_returns, printed in output.txt). The rule
floors the return at the 9.39% WACC and caps it at Meta's median five-year
return on the same measure, 41.3%; neither binds.
Why not the sixth report's cap. The sixth report capped the central at the 10.7% its 2027-31 build earns by 2031. That return is measured while Meta's capital per revenue dollar climbs: each new revenue dollar in 2027-31 takes about $2.45 of net capital, against $1.59 for the whole business in 2031. Holding 10.7% for ever means that ratio keeps climbing after the build ends, the same kind of unsupported permanence our capital reviews removed from the plant-per-revenue floor. The cap was also not one rule: the pessimistic case kept 17.2%, above its own 13.4% today, which is why it falls $14.89.
Why not Meta's history. Meta's five-year returns on the same volume-only measure ran 26% to 107% over 2013-25 (median 41.3%), and every history rule makes Meta a Buy ($652 to $685, the study's dry run). Those returns were earned on 0.3-1.0x net capital per revenue dollar; this forecast holds 1.6x. Using them would assume Meta returns to the old intensity, which nothing in the forecast shows. If it does, the place to show it is the capital path, not the long-run return.
| Long-run return (central) | Central | View at $641.76 |
|---|---|---|
| 41%, Meta 2015-25 | 678.98 | Buy |
| 27%, Meta 2018-23 (the fifth report's) | 655.24 | Buy |
| 22.6%, what the price needs | 641.76 | |
| 16.5%, the central's own 2031 capital held (this report) | 610.85 | Hold |
| 12.0%, Meta 2017-22 | 567.92 | Hold |
| 10.7%, the 2027-31 build's own (the sixth report) | 549.26 | Hold |
| 10.3%, the Avoid line | 545.50 | |
| 9.39%, the cost of capital | 536.44 | Avoid |
The gap between the exit's two halves narrows: the long fade gives $826 a share and the outside view, 20.4 times year-five operating profit, $1,123, 36% more (79% in the sixth report). The input behind the rest of the gap is still the return: peers earned their multiples on returns above 16.5%. We keep the method's mean.
2. Equity funding
Unchanged from the fifth report and our read of the filings: the $22.5B forward sale settles in 2027 (35.0M shares at $642.62), the $12.5B preferred pays 6.25% from February 2027 and converts in August 2029 (16.2M shares at our central value), and the planned $60B at-the-market program sells $30B in each of 2027 and 2028. Shares rise 5.25% by 2031.
The $160B debt peak assumes the unsigned $60B program sells in full. With the underwritten offerings only, central debt runs $121.2B, $189.7B and $222.6B in 2027-29 and stays there (market.txt); borrowing alone, $258.1B.
| If | Central | Change |
|---|---|---|
| Borrowing alone, no stock | 617.08 | +6.23 |
| Underwritten offerings only, no program | 614.21 | +3.36 |
| Offerings priced at $580 / $700 | 607.58 / 613.17 | −3.26 / +2.32 |
| Both options exercised | 610.45 | −0.40 |
| Program sold at our value path, not the market's | 609.89 | −0.95 |
| Preferred dividend 6.5%; band 17.5% or 22.5% | 610.83; 610.77 / 610.92 | under ±0.10 |
Stock sold near $642 now costs existing holders $6.23 a share against borrowing ($3.28 in the sixth report), because our value per share is closer to the issue price.
3. Forecast and value
Unchanged apart from the long-run return: ad growth 20, 13, 10, 9 and 8% in 2027-31; central capital spending $220B, $220B, $211B, $191B, $161B; revenue $524B in 2031 with a 29.5% operating margin.
| Pessimistic | Central | Optimistic | |
|---|---|---|---|
| Capital spending 2027-31 ($B) | 150, 110, 98, 98, 103 | 220, 220, 211, 191, 161 | 240, 240, 220, 204, 192 |
| Net capital at 2031 ($B) | 538.9 | 746.1 | 942.2 |
| Long-run return on new capital | 13.4% | 16.5% | 16.5% |
| Year-five after-tax operating profit as valued ($B) | 72.0 | 122.9 | 155.3 |
| Debt at 2031 ($B) | 85.2 | 160.0 | 171.5 |
| Exit price, 2031 | $413 | $974 | $1,557 |
| Fair value today | $265 | $611 | $971 |
4. What the price expects, and our view against it
The price needs any one of: ad growth 0.5 point a year above our path, apps cash costs 1.2 points of revenue lower, or a 22.6% long-run return (odds.json, market.txt). 39% of 4,000 simulated values are at or above the price; the middle one is $549. The market, in our reading, expects slightly faster ad growth or a build that needs less capital per revenue dollar after 2031 than ours does. Either is plausible; neither is in the forecast we wrote.
Capfolio's ranking model scores Meta 15.6 of 100, 589th of 697, "Weak" (cache as of 2026-10-24, read 2026-10-26; rerating pending).
5. Why the other side could be right
For Buy: Meta's AI ranking work is lifting price per ad 12-14% a year, and Meta has never had a five-year window below 26% on the measure the fade uses. Ad growth half a point a year above our path, or capital needs after 2031 falling toward Meta's history (a 22.6% long-run return), makes it a Buy.
For Avoid: ad growth 1.15 points a year below our path, apps cash costs 2.5 points of revenue higher, a long-run return below about 10.3% (the build's own 10.7% gives $549.26; the cost of capital, $536.44), or 2027 spending above about $309B with 2028 held there. Meta selling stock while rated Aa3/AA- fits a management that expects the newest capital to earn less than its past builds. The Hold is fragile to the forecast itself: the ad-growth margins above sit inside the report's own driver ranges.
6. Next results: Q4 2026, expected 2027-01-27
The fourth report's seven Q4 2026 forecasts stay the graded set; none is changed by this revision. What a surprise is worth now: each $1B of Q4 ad revenue above ours, kept as a level, about $3.7 a share; 2027 ad growth of 18% instead of 20% is −$14.46; each $20B of 2027 spending, with 2028 held at it, about −$14.
7. Dates, and what would change the view
| Date | Event | What it tests |
|---|---|---|
| Within days of 2026-10-26 | Pricing of the two offerings | Under $4 either way; priced at $580 the central is $607.58 |
| 2027-01-19, then late January | Class C dividend and trading; the program may start | Per-share figures halve; without the program debt peaks at $222.6B |
| 2027-01-27 (unconfirmed) | Q4 2026 results and the first 2027 guides | A first capital spending guide above about $269B (2027 spending above $308.9B at our 1.15 ratio) is an Avoid even if the extra earns its cost; at $198B, $605.30 if it does and $555.28 if none of it does |
| Q4 2027 | Return on new capital (13% at Q3 2026) | Below 9.4% with a 2028 guide above the October path ($159.5B) |
To Buy: ad growth 0.5 point a year above our path; evidence that capital per revenue dollar falls after the build (a 22.6% long-run return). To Avoid (below $545.50): ad growth 1.15 points a year below our path; apps cash costs 2.5 points higher; a first 2027 capital spending guide above about $269B.
8. Reviews
The rule was reviewed once, before the owner approved it: an independent Codex review of the study argued Meta's Buy and Avoid cases against it (proposal.md).
| Finding (Codex, 2026-10-26, on the rule) | Effect | Answer |
|---|---|---|
| Meta's capital base included $20.4B of operating-lease assets the forecast never rolls forward | +$3.11 central | Fixed: the 10-K's $176.4B net property |
| Holding 2031 capital per revenue dollar for good is still a judgment | none | Accepted; it is the case's own level, neither rising (the cap) nor falling (history) |
| The Hold is fragile to the forecast: about +0.6 point of ad growth reaches Buy, −1.1 Avoid | none | Accepted and stated above |
| An average return on depreciated book need not equal the return on new AI capacity; at the WACC the central is $536.46, an Avoid | none | The pessimistic case carries that risk; measured consistently the build earns 12.1%, not 10.7% |
The sixth report's reviews, unchanged:
| Finding (first review: independent Codex, 2026-10-26) | Effect | Answer |
|---|---|---|
| 27% is not Meta's worst five-year return (2017-22 earned 12%); cap the mature return at the build's own return | −76.57 central | Fixed: capped at the build's own return |
| Avoid line used 85% of fair value; the rule is 85% of the price ($545.50) | view thresholds | Fixed in market.py and throughout |
| Optimistic case skipped the 2028 hold | −71.87 optimistic | Fixed |
| Payout counted future shares in earlier dividends (about $0.05) | +0.03 | Changed: the 2026 dividend now uses the exit share count like later years; on our check this raised the value |
| Stale citations in drivers.json | none | Fixed |
| The $160B debt peak assumes the unsigned program sells in full | none | Stated: $222.6B without it |
| Finding (second review: independent Opus, 2026-10-26, scripts probe1-3.py) | Before | After | Effect | Answer |
|---|---|---|---|---|
| Year-five depreciation normalised to 25.2% long-run spending (+$20.3B NOPAT) while the fade reinvests at a return that implies 37.1%; the cap measured on the raised profit | depreciation at 17.3% of revenue; cap 12.5% | the forecast's own 22.5%; cap 10.7% | central −46.28, pessimistic +9.67, optimistic −80.42, weighted −40.82 | Fixed, reproduced to the cent ($549.26, weighted $573.63). The 25.2% had no source independent of the uncapped return, so it was removed rather than kept |
| The report said the peer models also normalise depreciation; the record said the Alphabet and Amazon reports apply the cap | wording | wording | none | Fixed: only the Microsoft report caps the return, and none of the three normalises depreciation |
| drivers.json cited the cash-cost sensitivity at ±$26 when the draft gave ±$21.5 | ±$21.5 | ±$26 | none | Fixed by the rerun: the sensitivity is now −$25.87 / +$25.80 |
| odds.json drew with the draft's 12.5% return and 25.2% spending | P 39%, median $548 | P 33%, median $496 | none | Rerun after the fix |
After the fix, drivers.json's mature-return range was widened at the bottom (6% to 3.9%) because the odds tool requires the central (now 10.7%) to sit between 29% and 71% of the range; it moves no value. The second review's fix did not change the view (Hold before and after), so no second round was run.
Revision history
uv run alpha research history META (2026-10-26):
| 2026-10-15-meta | avoid | 112 | 362 | 997 | first | 732.51 | 2026-10-15T16:38:19Z | partly unblinded | | 2026-10-15-meta-2 | hold | 166 | 656 | 1,178 | +294 | 740.26 | 2026-10-15T18:58:20Z | partly unblinded | | 2026-10-20-meta | hold | 250 | 734 | 1,198 | +78 | 694.44 | 2026-10-20T19:59:59Z | unblinded | | 2026-10-21-meta | hold | 232 | 686 | 1,133 | −48 | 698.43 | 2026-10-21T19:59:59Z | unblinded | | 2026-10-22-meta | hold | 265 | 686 | 1,139 | 0 | 658.50 | 2026-10-22T14:13:18Z | unblinded | | 2026-10-22-meta-2 | buy | 325 | 686 | 1,139 | 0 | 646.09 | 2026-10-22T16:49:46Z | unblinded | | 2026-10-22-meta-3 | hold | 320 | 557 | 985 | −129 | 648.87 | 2026-10-22T19:59:59Z | unblinded | | 2026-10-22-meta-4 | buy | 320 | 655 | 985 | +98 | 648.85 | 2026-10-22T20:00:00Z | unblinded | | 2026-10-26-meta | buy | 270 | 672 | 1,167 | +17 | 641.76 | 2026-10-26T16:24:39Z | unblinded | | 2026-10-26-meta-2 | hold | 279 | 549 | 916 | −123 | 641.76 | 2026-10-26T16:24:39Z | unblinded | | 2026-10-26-meta-3 | hold | 265 | 611 | 971 | +62 | 641.76 | 2026-10-26T16:24:39Z | unblinded |
- 2026-10-15-meta to 2026-10-22-meta-4: as given in the fourth and fifth reports (error fixes, the shared method, Q3 data, the Buy rule, the spending-and-revenue rule).
- 2026-10-26-meta, $655 to $672: equity funding (−$9.48 against borrowing), 2028 capex held (+$1.08) and the long-run floor replaced by a 27% return (+$25.21).
- 2026-10-26-meta-2, $672 to $549, errors fixed after two independent reviews: the long-run return capped at the build's own return (−$76.57), year-five depreciation no longer normalised to a 25.2% long-run spending share (−$46.28), the optimistic case's 2028 hold, the Avoid line at 85% of the price.
- 2026-10-26-meta-3 (this report), $549 to $611: one long-run-return rule for the four hyperscaler reports, approved by the owner; each case holds its own 2031 capital per revenue dollar (+$61.59). Made after the price was seen.
Sources
As in the sixth report, plus Meta's 10-K
2025 balance sheet (net property $176.4B, operating-lease assets $20.4B;
data/ of 2026-10-22-meta-4), valuation.md's
mature-return rule and the study
2026-10-26-hyperscaler-return-rule.
Details
- Report date
- Oct 26, 2026
- Price source
- Last trade $641.7642 at 2026-10-26 16:24:39 UTC (official close 2026-10-23 $642.62), Oct 26, 2026, 9:24 AM PDT
- Bear / bull
- $265 / $971
- Exit price
- $974 (on Dec 31, 2031)
- Confidence
- 26%
- Price against our scenarios
- The price is justified in 20-50% of our scenarios4,000 simulated outcomes, each drawing the uncertain drivers from ranges we chose: our judgement, not a measured chance
- Worlds
- 3 stories about the businessEach outcome first picks one of these stories by its weight, then draws that story's own driver ranges; the answer above is their mix.
- World: base (50%)
- the base drivers. The price is justified in 20-50% of its scenarios; middle outcome $581, 10th to 90th $282 to $1,216Weight: Our judgment, matching the central case's 50% weight.
- World: bear_lean (25%)
- an ad slowdown, spending above the October path that does not pay, long-run returns below the central's own. The price is justified in under 20% of its scenarios; middle outcome $362, 10th to 90th $185 to $741Weight: Our judgment, matching the bear case's 25% weight.
- World: bull_lean (25%)
- AI ranking keeps ads near 20% into 2028 and costs leverage. The price is justified in over 50% of its scenarios; middle outcome $722, 10th to 90th $372 to $1,429Weight: Our judgment, matching the bull case's 25% weight.
- What moves the answer
- the exit price (market outcome) 69%; advertising revenue growth 2027-31, points a year added to the central path (20, 13, 10, 9, 8%) 18%; Family of Apps cash costs 2027-31, points of FoA revenue added to the central path (46% falling to 41%) 13%; return on new capital after the build (the fade's mature return); year-five depreciation as the path has it 8%How much of the spread in value each driver explains on its own; drivers that move together share credit, so the shares need not add to 100%.
- Leaning cases
- 2 other views we ranThe same simulation with some ranges moved to another view, to show how much the answer depends on our ranges.
- Case: bear_lean
- an ad slowdown, spending above the October path that does not pay, long-run returns below the central's own. The price is justified in under 20% of this case's scenarios; middle outcome $362, 10th to 90th $185 to $741
- Case: bull_lean
- AI ranking keeps ads near 20% into 2028 and costs leverage. The price is justified in over 50% of this case's scenarios; middle outcome $722, 10th to 90th $372 to $1,429
- Cost of equity (original)
- 9.8%
- Original figures
- fair value $611 (bear $265, bull $971), exit $974; Hold
- Street
- average $834, median $862, low $595, high $1,000; 31 targets in the last quarter; Street consensus targets, Oct 26, 2026, 12:17 PM PDT
- Restated
- Oct 26, 2026, method case-spread-2026-10-24; forecast.json valued by method.py at market inputs 2026-10-01
- Cost of equity calibration
- +0.0 points
- Calibration source
- none: the 1-point cut set 2026-10-16 to pull the batch toward the Street was removed on 2026-10-18 (Dong); the cost of equity is risk-free plus beta times the premium
- View rule
- strong buy when the price is at or below 0.80x fair value; buy when at or below 0.90x fair value; avoid when fair value is below 0.85x the price; otherwise hold
- Blind step
- Unblinded; the model was committed in 86e20b4 before the record
Original call, Oct 26, 2026 (before the Oct 26, 2026 restatement)
We keep Meta a Hold and raise fair value to $611 from $549, against $641.76: the price is 5.1% above our value, now well clear of the $545.50 Avoid line. Nothing about the business or the forecast changed. One input did: the long-run return on new capital after 2031. Our four hyperscaler reports set it four ways this morning, and the owner approved one rule for all of them: each case earns, on growth after 2031, its own year-five return on its net capital, so growth needs the capital per revenue dollar the case itself reaches in 2031. For Meta's central that is 16.5%, against the 10.7% its 2027-31 build earns by 2031 (the sixth report's cap, which assumed capital per revenue dollar keeps rising for ever) and Meta's historical 26-107%, earned on far lighter capital. Buy needs ad growth half a point a year above our path or a 22.6% long-run return; Avoid needs ad growth 1.15 points a year below it or a return under about 10.3%. Pessimistic $265, optimistic $971. Per-share figures halve after the January 2027 Class C dividend.
As of 2026-10-26 · unblinded: a same-day revision of the sixth report under one long-run-return rule the owner approved for MSFT, GOOGL, AMZN and META, written after the price was seen (assumptions.md). Model committed before this report (86e20b4). Revises 2026-10-26-meta-2 ($549, Hold), which stays on the record.
- Fair value today
- $611 per share, discounting at a 9.77% cost of equity. Weighting the pessimistic, central and optimistic cases 25/50/25 (our weights) gives $614. After the January 2027 Class C dividend every per-share figure halves: $305
- Price
- $641.76, last trade 2026-10-26 16:24 UTC (official close 2026-10-23 $642.62) → 5.1% above fair value
- Five-year exit price
- $974 a share on 2031-12-31: the average of a long fade of Meta's returns toward its cost of capital ($826) and what similar companies have traded at ($1,123). Fading over 10 or 20 years gives $942 or $1,007
- Five-year return
- 8.7% a year from $641.76 to the central exit, dividends included, against a 9.8% cost of equity
- Bear / bull today
- $265 / $971: our pessimistic and optimistic cases, each one coherent business outcome
- Confidence
- About 26%: the share of our simulated values within 20% of $611, from odds.json's percentiles (subjective, not calibrated)
- Moat
- 3.64 billion people use a Meta app each day (September 2026). The best ad-targeting system outside Google, in our judgment; no moat yet in AI models or in selling computing power
- Market expects
- A little more than our central: ad growth half a point a year above our path, or a 22.6% long-run return on new capital where ours is 16.5%
- Price needs
- Any one of: ad growth 0.5 point a year above our path; apps cash costs 1.2 points of revenue below it; a 22.6% long-run return on new capital
- Odds
- P(price justified): 20-50%, the share of our simulated values at or above the price (39%). Subjective, not calibrated (odds.json)
- Key driver
- Ad growth in 2027-31: each point a year adds $62 or takes $57. Then the long-run return on new capital: Avoid below about 10.3%, Buy above about 22.6%
- Street
- Average 12-month target $834 (median $863, range $595-1,000), 31 analysts in the last quarter; 14 in the last month averaged $886 (market data, read 2026-10-26). An expectations reference, not validation