NFLX

As of 2026-10-20 · partly unblinded · revised after review · blind commit ddd43ce7219 · market pass after it

We rate Netflix a Buy at $68.66. Our central fair value is $88.72 a share, 29.2% above the price, and the central case returns 16.2% a year over five years, above the 12% a year we require (the price is under $82.50, the price at which the central case earns exactly 12%). The call hinges on one input: what buyers will pay for Netflix's profit in 2031, the exit multiple. At the 16.1x EV/EBIT (enterprise value over operating profit) of the cheapest media peers instead of our 26.3x, the price would be right. The stock fell 5.8% after hours on the Q3 2026 results, to $64.66.

How to read the cases. Each report values a company in three cases. The central case is our best estimate. The bear case ($61.00) is a severe but plausible bad outcome, here an engagement-share loss to YouTube and short video. The bull case ($115.89) is the matching good outcome, here the ads tier and price rises both working. The exit price is our estimate of the stock's value on 2031-12-31, the mean of two: the "long fade" (which fades growth from 7% to 3% over 15 years, so value follows cash flows) and the "outside view" (what buyers have paid for similar companies). Cost of equity (10.37%) is the yearly return investors require; we discount future cash to today at that rate.

What "blind" means here. We first value a company without seeing its market price. The model, forecast and assumptions were committed blind at ddd43ce7219 (fair value $108.23) before any quote, multiple, target or estimate was read. The report is partly unblinded because research-mask.py (the script that hides price-revealing passages from the filings we read) missed one repurchase sentence in the Q4 2025 letter that implies an average Q4 2025 buyback price near US$110; because the method's anchor history (the multiples it uses to set the outside view) shows Netflix's year-end 2025 multiple, so its year-end price level; and because the analyst remembers 2025 trading levels (assumptions.md lists each). None is a model input.

This is a revision. After the market pass (d30d4a6a5, $108.23 at a $68.66 price) two outside reviewers read the report with the price in hand, and our chief investment officer (CIO) accepted four changes to the model (section 10). A revision after the market is read cannot be blind. Each change rests on dated evidence that is not the price (the Q3 letter's growth and ad figures, the 2018-2021 history behind the multiple, the tax history), with its before, after and dollar effect in section 10; all four lower the value, so they move it toward the price, and the reader should weigh them with that in mind. Written the evening of the Q3 2026 results; the earnings interview was not read (section 8).

1. Netflix earns from subscriptions and ads, and its content budget grows slower than revenue

Netflix sells a monthly subscription, in 2026 with a cheaper ad-supported plan that takes over 60% of sign-ups where it is offered (Q1 2026 letter, claim). Revenue is members times price plus ads: about 334M average paid memberships in 2026 (our assumption from 301.63M at end-2024 and "over 325M" at end-2025; Netflix stopped reporting the count in 2025) at about US$12.79 a month gives US$51.2B (model; guide ~US$51.1B), of which ads are about US$2.7B (Q3 2026 letter, claim).

The cost that matters is content. Amortization was US$16.42B in 2025, 36.3% of revenue, and grows about 10% in 2026 (10-K 2025; Q4 2025 letter). It is a fixed budget for a global audience: in 2024-2025 it grew 7-8% a year against revenue growth near 16%, which is how the operating margin went from 17.8% (2022, market data) to a guided 31.5% (2026). Cash spent on content runs about 1.1x amortization (guide), so reinvestment is small: the model's 2027-2031 net reinvestment is US$10.5B while NOPAT rises US$8.7B from 2026 to 2031, a volume-only return on new capital of 70% (model). The 2031 central margin of 35.8% rests on one sentence: content grows 6-7% a year while revenue grows 7-10%.

2. Q3 2026 matched the revenue guide, beat on profit and cut the ad forecast

Q3 2026, released today: revenue US$12,874M against the US$12,860M guide (+0.1%); operating income US$4,459M against US$4,268M (+4.5%, "timing of expenses"); EPS US$0.87 against US$0.82 (Q2 and Q3 2026 letters). The vendor's EPS 0.87 and revenue US$12.874B agree with the letter; the vendor's Q3 tax line (-653) and balance-sheet receivables (34,605, which is content assets) are mislabelled.

What changed in the letter:

• 2026 revenue ~US$51.1B, 0.2% under the old range midpoint, because ads come in at ~US$2.7B against ~US$3B; subscriptions beat the plan (claim).
• Revenue growth by quarter in 2026: 16.2%, 13.4%, 11.8%, and the Q4 guide 11.5% (letters; model arithmetic). The slowdown is the fact behind the 2027 re-base in section 10.
• Q4 margin 26.0% includes ~120bp for headcount cuts; 2026 margin still 31.5%.
• 2026 free cash flow (FCF) ~US$12.0B (was 12.5B) for a Q4 real-estate purchase.
• Record buybacks: US$4.80B in Q3 after US$4.71B in Q2 (cash-flow statements); about US$22.3B of authorization left (inference, 8-K 2026-04-23).
• The ads tier reaches 15 more countries from March 2027.

The guide record over the last eight quarters (model): revenue median +0.58% (range -0.14% to +1.22%), operating income +3.06%, EPS +2.66%. The two misses are explained in the filings: Q3 2025's US$619M Brazil non-income-tax charge, and the ads shortfall now. Netflix guides to its internal forecast and lands on it; the measured bias is a small, steady beat on operating income from expense timing, not on revenue.

Since the last 10-Q: Warner Bros. Discovery (WBD) terminated the Warner Bros. deal on 2026-02-27 and Paramount paid Netflix the US$2.8B fee (8-K), already in cash; Netflix raised US$1B of 5.250% 2036 notes to repay the November 2026 maturity (8-K 2026-07-22).

3. Our forecast: $76.4B of revenue in 2031 at a 35.8% margin is worth $88.72 a share

model.py, output.txt, assumptions.md.

Where we differ from guidance and history: we take the 2026 guide as given plus the median beat. After it, our margin rises about 0.8 points a year against 2-3 points a year in 2023-2026, because content amortization growth stays near 6-7% rather than falling further, and member growth fades from 6.5% (2026) to 3.2% (2031) as view hours grow only 1-2% a year. Revenue growth falls from the Q4 guide's 11.5% to 9.9% in 2027 (section 10, the 2027 growth change).

The three cases are each a coherent world (assumptions.md, Cases):

• Bear, $61.00: an engagement-share loss to YouTube and short video. Members grow 3% falling to 1.5%, price 2-2.5%, ads reach only US$4.9B, but content still grows 5-6.5%, so the margin peaks at 31.7% in 2027 and drifts to 29.9%. Its 2027 revenue growth is 6.6%.
• Central, $88.72: above.
• Bull, $115.89: the ads tier and price both work: members 6.2% to 4.4%, price 4.3% to 3.7%, ads US$9.5B; content grows 6.5-7.5% and the margin reaches 40.3%. Its 2027 revenue growth is 12.4%, at the Q4 exit rate plus the median beat, not above it.

Exit. The change from the method's default is recorded in Forecast.changes: the outside view is 26.3x EV/EBIT, "anchor (f)" of the method's exit-anchor study (which tested how well each candidate formula predicts a company's multiple five years on), the mean of Netflix's own 2021-2025 median (32.3x) and the peers' ten-year median (20.2x; Disney, Comcast, WBD, Alphabet, Meta and Electronic Arts; Paramount Skydance and Spotify lack enough years), in place of the default 30.8x ("anchor (d)": own ten-year median 41.4x with the peers' 20.2x). Why: the 41.4x is set by the 2018-2021 year-ends of 45-77x, when revenue grew 20-30% on 10-18% operating margins. A 2031 Netflix grows 7% at a 36% margin. The checks (output.txt, "Exit multiple checks"):

• Value-driver formula. At 7% growth after 2031, a 19.0% mature return and a 10.37% WACC it gives 18.7x year-six NOPAT, 16.2x year-five EBIT. The default 30.8x outside view (38.0x NOPAT, 35.5x year-six NOPAT) needs perpetual growth of 8.9% at the mature return, above the path's own 7%; the judged 26.3x needs 8.6%, still above it. An outside view is what buyers pay, not a perpetuity, so it is averaged with the fade. The exit that results, the mean of the two, is 19.3x EBIT (23.8x NOPAT, 22.3x year-six NOPAT), which implies 7.7% perpetual growth at the mature return against the path's 7.0% in year six: close, and 19% above the formula's 16.2x. The formula prices one mature return forever, so it understates a business still growing 7% in year six.
• Sales multiple. Netflix's year-end EV over sales was 4.4x to 10.1x in 2016-2025 (median 7.7x; 8.9x in 2021-2025), the peers' median 3.8x (their median margins 13-26% against Netflix's 36% in 2031). The exit is 6.9x sales, inside Netflix's own range and under its median; the judged outside view alone is 9.4x (top of the range); the default 30.8x would be 11.0x, above every year of the ten.
• Where it sits. 19.3x EBIT is also within 4% of the peers' median 20.2x; the bull's 20.1x and bear's 18.2x differ because the fade's half moves with each case.

The fade prices 2031 at 12.3x EBIT and the outside view at 26.3x, a gap of more than 2x. The input behind it is again the multiple, not the forecast: the fade fades growth from 7% to 3% over 15 years at a 10.37% WACC. We believe a 2031 Netflix growing 7% is worth nearer the fade on cash flows than the outside view; the outside view is the record of what buyers have paid, and the study found the mean the better predictor of the multiple five years on. Anchor (f) is one of the study's tested alternatives (error 0.435 against (d)'s 0.449 over 266 company-years; the study set no adoption rule for it). We chose it on the 2018-2021 reasoning above, not on the price. Alternatives, all in output.txt: the peers' 20.2x alone -$11.93; the default 30.8x +$8.86; Netflix's own 41.4x +$29.65.

One multiple for every case (a limit of our method). method.Change is forecast-wide, so the bear, central and bull all take the 26.3x; the method cannot give them coherent multiples of their own, and we did not invent a mode. The bear's 26.3x is therefore generous, as it was before. The bear at the peers' 20.2x is $52.52 and at 15x $45.22; the bull at 25x is $112.64 and at the default 30.8x $127.16 (output.txt). The simulated odds do use coherent multiples (bear-leaning 13-26x, bull-leaning 26-41x), so they are wider than the three cases.

Bridge (counted today): cash and short-term investments US$8,223M less US$2,000M operating cash, less US$15,337M of debt at principal (Q3 2026 letter). Free cash flow is paid out as earned (Q4 2026 US$2.64B, then US$12.9-20.2B a year), at or below the US$22.3B authorization left in total; full payout is the method's convention. Shares 4,165.7M diluted (inference, assumptions.md), held flat: stock pay is inside GAAP operating income, so it is charged once.

Tax. 19% for 2027-31 and the fade, 17% in 2026 (the Q4 guide). Effective rates were 9.5-14.7% in 2019-2025 and 15-16% in Q2 and Q3 2026 (market data; Q3 letter), helped by option-exercise benefits that shrink as the stock falls, against a US federal rate of 21% before state tax and credits. 19% is the midpoint of the 18-21% band; 21% is -$1.09, 17% +$1.09.

Cost of equity 10.37% = 5.29% + 1.210 x 4.20% (Damodaran 2026-10-01; beta the Blume-adjusted median of the eight peers' market data betas, raw 1.314). Beta 1.0 is +$7.16, 1.4 is -$5.45; Netflix's own beta is in section 8.

Drivers and their worth (central fair value, output.txt)

Fair value on the fade alone is $61.26; on the outside view alone $116.17.

4. Where we differ from the common view: ads are not the engine; pricing and content spend are

Without the price, the variant is against the company's own story and the common one. Ads are not the engine: even our central has them at 9% of 2031 revenue, and 2026 already fell 10% short of management's own number. The value rests on two quieter things: price rises that members keep accepting (the company calls the 2026 rises "consistent with prior changes") and content spend that grows slower than revenue. Engagement is the weak link: view hours grew 1.5% in 2025 and 2% in H1 2026, and the Q3 letter answers "investor concern" about it. A business that raises price faster than it adds viewing is spending down member goodwill or proving pricing power; the 2027 member and price split, which Netflix no longer reports, is what we cannot see directly. The WBD walk-away with a US$2.8B fee and record buybacks say capital discipline holds.

5. The bear case ($61.00) is not a collapse

The bear is not a collapse: it is a world in which YouTube, short video and the media owners' bundles take TV time, Netflix keeps paying more for sports and films to hold members, and pricing slows. That world is worth $61.00, 31% under the central, and it needs no event, only engagement staying flat for five more years. The bull is the ads tier finally scaling and price rises continuing at 4%: $115.89. Bear and bull are roughly symmetric around the central (weighted $88.58), so the cases do not lean. But the bear is also generous, because it keeps the central's 26.3x exit multiple; at the peers' 20.2x it is $52.52.

6. What to watch: dates and results that would change our view

• 2027-01-19 (inference from the Tuesday pattern of the last three Q4 letters): Q4 2026 results and the 2027 revenue and margin outlook.
• By the end of Q1 2027: Netflix's first annual What We Watched engagement report (full-year 2026 view hours). From 2027 it is published once a year in the first quarter, separately from earnings (Q2 2026 letter, 2026-07-16); the date is not announced and it does not come with the Q4 results.
• March 2027: ads tier launches in 12 of the 15 new countries (Q3 2026 letter).

Falsifiers:

1. A 2027 revenue outlook midpoint under US$54.8B (+7.0% on our 2026) says our members-and-price path is still too high; each US$1B is about $3.01 a share.
2. A 2027 operating margin outlook under 32.0% says content growth is not slowing; each point held through 2031 is about $2.63 a share.
3. Full-year 2026 view hours growing under 1.5%, the 2025 rate, in the annual What We Watched report moves the central toward the bear's member path.
4. A 2027 cash content to amortization ratio above 1.15x, or a large acquisition (another WBD-scale bid), changes the reinvestment case.
5. 2027 ad revenue under US$3.1B (if disclosed) moves the ad path toward the bear's US$4.9B for 2031 (about -$5 a share).

7. Next results preview: Q4 2026 (due about 2027-01-19)

From output.txt, "Next results preview"; every point uses the company's guide and its measured beat record (last eight quarters), and the 2027 lines use the model's cases. Also in the forecasts list of record.json, due 2027-01-19 (inferred from the pattern of the last three Q4 letters, filed 2024-01-23, 2025-01-21 and 2026-01-20, EDGAR; the company has not announced it).

The most important is the FY2027 revenue outlook: a US$1B surprise against our US$56.3B, carried through the path, is worth about $3.01 a share (output.txt). The Q3 2026 actuals are inputs here and are not graded.

8. What the price implies, the Street, and the market's side

All figures from market.txt and odds.json unless named.

The stock closed at $68.66 on 2026-10-20, 45% under its 52-week high of $124.86 and $3.58 above the low of $65.08 (market data quote). After the Q3 letter it traded at $64.66 (market data get_equity_quotes, 20:59:02Z, -5.8% on the close; it was $65.12 at 20:38:08Z), below the 52-week low. The record price stays the regular-session close. By rule the view is buy: the price is under the $82.50 purchase price, and the central return is +16.2% a year (+17.6% from after hours). On Netflix's own market data beta (1.613, Blume 1.411, cost of equity 11.21%) the fair value is $82.98 and the purchase price $80.14; the view holds (section 10).

The earnings interview. We looked for a transcript or summary of the Q3 earnings interview through news feed (get-news, 60 headlines on 2026-10-20) and found none: the call entry is a live notice, and the Dow Jones, Reuters and WSJ items that touch it are paywalled headlines ("Netflix Says Engagement is Healthy Despite 2026 Price Hikes", "Netflix Says Layoffs Will Impact 4Q Margins", "Netflix Doubles Down on Third-Party Content", WSJ "Revenue Growth Slows as It Overhauls Programming Units"). The two free summaries (Seeking Alpha, StockStory) repeat the letter. We did not read the interview, so nothing from it is in the model, and its content is not claimed here. Whoever reads it should check the ad rollout and the 2027 pricing.

What the price says (our report check's low-odds question was not raised: P 74%). The price is between our bear case ($61.00) and the fade alone ($61.26 at the central path), and 22% under the central. The world in which $68.66 is right, built as hard as we can:

• The exit is a media multiple, not Netflix's own. The outside view at 16.1x EV/EBIT, instead of the judged 26.3x, gives the price. That is the value-driver formula's multiple and the cheapest peers' (CMCSA 13.6x, WBD 11.4x). If streaming is now a mature utility competing with YouTube for TV time, the history of a growth company's multiple is the wrong anchor, and this is the single input that decides the call.
• Growth slows sooner. Or the multiple holds and member and price growth each run 1.4 points a year under the central: 2031 revenue $67.3B and EBIT $21.2B, with the margin 31.4%, under 2026's. Odds find the same thing one driver at a time: members growing 1.2% a year, or price per member 0.3%, or content cost growing 11.4% a year.
• The evidence for it is real. View hours grew 1.5% in 2025 and 2% in H1 2026; 2026 ad revenue was cut 10% to about US$2.7B; revenue growth has fallen from 16.2% to a guided 11.5% in three quarters; the letter answers "investor concern" on engagement; and the stock lost 5.8% after hours on these results.

We keep the central, because the operating path the price needs is not what the company reports or guides. 2026 revenue and margin track the guide (our 31.7% against a guided 31.5%), and the 2026 price rises went "consistent with prior changes" (Q3 letter, claim). The multiple is the open question, and we have already taken $9.68 off for it. The 30% confidence says the rest. We do not widen the ranges toward the price; the simulated range ($50-$149, median $95) is wider than before because the exit multiple is now drawn (our report check's narrow-range question was not raised).

Against the Street (our report check's question about a large gap to the Street target was not raised). The average target of $88.26, discounted one year at 10.37%, is $79.96; our $88.72 is 1.11x it, inside the 0.6x-1.5x band. The Street's operating forecasts are close to ours: 2027-2030 revenue within 1-3% (Street $73.2B in 2030 against our $71.5B). Its EBIT, scaled to GAAP by its 2025 miss of reported operating income (0.965), sits close to ours in 2027-28 and over it in 2029-30 (42.4% margin against our 35.5%). Run through this model, the Street's revenue alone gives $92.79, and its revenue and EBIT together give $107.08. Its $88 target therefore holds a lower exit multiple, not a weaker business: the disagreement is the same one the price shows. The 2029-30 Street rows are thin (29 and 15 analysts) and the 2026 EBIT of $18.7B sits $2.5B above ours, near the US$2.8B termination fee (inference: a fee recorded outside operating income in an adjusted figure).

Net move from the blind value (-$19.51, -18.0%). Our report check asks us to name the largest change. The largest change is the exit multiple, -$9.68 of it. It rests on the 2018-2021 history in the company's own filings-based multiples (market data year-end EV over operating income) and on the formula and sales checks, not on a management claim or a Street figure; the Street's own target did not enter. The 2027 growth re-base (-$8.47 with ads) rests on the Q3 letter's quarterly revenue and ad figures, a management claim in its dated form (the guide and the ad cut), and the tax change (-$1.36) on the market data tax history.

Two further questions. Value 29.2% above the price is the exit multiple again, not a risk charged twice: the cost of equity uses peers' beta, and the bear case carries the engagement risk once. The outside view at $172.64 against the fade's $80.94 comes from one input, the 26.3x multiple.

9. Sources

All dated; each is pinned in data-manifest.json.

• Netflix shareholder letters, Q3 2024 to Q3 2026 (8-K exhibit 99.1, EDGAR), 2024-10-17 to 2026-10-20.
• 10-K 2025 (2026-01-23); 10-Qs Q1 and Q2 2026 (2026-04-17, 2026-07-17).
• 8-Ks: 2025-10-30 (split), 2025-12-22 (bridge loan), 2026-01-20 (all-cash amendment), 2026-02-27 (termination, US$2.8B fee), 2026-04-23 (US$25B authorization), 2026-07-22 (2036 notes), 2026-07-30 (director).
• Market data income, balance-sheet and cash-flow statements, annual and quarterly, fetched 2026-10-20; market data profile betas for the eight peers only; annual revenue for Netflix and the eight peers (data/peer-sales.json, fetched by sales_data.py, for the sales-multiple check).
• data/exit-anchor.json (market data key-metrics and income, fetched 2026-10-20).
• Damodaran risk-free rate and equity risk premium, 2026-10-01 (method).
• Market pass, read after the blind commit: market data quote, profile, price-target consensus and analyst estimates (data/market/, 2026-10-20T20:37Z); market data get_equity_quotes (2026-10-20T20:38Z and 20:59Z, read-only); news feed headlines and two free story summaries (2026-10-20); target counts by period from market data as research-check.py printed them on 2026-10-20.

10. Revision after the adversarial review: four changes lowered value from $108.23 to $88.72

Output: revision.py, revision.txt. It starts from the committed inputs (reproducing $108.23) and applies one accepted finding at a time, in this order, each on top of the one before.

Exit price $156.74 to $126.79; bear $68.23 to $61.00; bull $144.80 to $115.89; weighted $107.37 to $88.58; purchase price $100.62 to $82.50; five-year return +21.0% to +16.2%. The step order matters for the split, not the total.

Accepted

• The exit multiple (a blocking finding: a reviewer said it must be fixed). The default formula's 41.4x own median is set by 2018-2021 year-ends of 45-77x at 20-30% growth and 10-18% margins; a perpetual 30.8x needs growth above 8% at the mature return and the formula gives 18.7x year-six NOPAT (16.2x year-five EBIT). The Forecast records the change. We chose anchor (f) 26.3x over the peers' 20.2x because 20.2x is a median that includes media and holding-company multiples (DIS, CMCSA, WBD) for a 36%-margin business, and the exit (the mean of the fade and the outside view) at 26.3x is already 19.3x EBIT, within 4% of the peers' 20.2x, and 7.7% perpetual growth against the path's 7.0%. At the peers' 20.2x the exit is 16.6x EBIT, closer to the formula, at -$11.93 more; section 3 shows both. It is a judgment, and it is the largest open input.
• Sales-multiple check. Added to the model (EV over sales against Netflix's and the peers' history): the exit is 6.9x sales against Netflix's 4.4x-10.1x. See section 3.
• One multiple for every case. A method limit, stated in section 3; the bear at 20.2x and 15x and the bull at 25x are sensitivity lines, no new mode.
• Odds. exit_multiple is a drawn driver (triangular 15-26.3-41.4x; bear-leaning 13-18-26.3x, bull-leaning 26.3-31-41.4x), correlated with member and post-2031 growth, and the odds rerun: P 74% (was 90%), 10th-90th $50-$149.
• 2027 growth. Quarterly growth fell from 16.2% to 11.8% in Q3 with an 11.5% Q4 guide (letters). Nothing in the letters shows growth stopping at the exit rate, so we let it keep easing about 0.5 point a quarter: 2027 is +9.9% (was +11.3%), with members 5.0% (was 5.5%) and price 3.5% (was 4.0%), and later years eased by about the same. Ads: Q2 letter about US$3B, Q3 letter about US$2.7B; +40% off that base is not supported, so +32% in 2027, +27%, +21%, +16%, +13% to US$7.2B (was US$8.0B). The bull is trimmed to +12.4% in 2027 and the bear stays at +6.6%. Content amortization is unchanged, so the margin falls with revenue (2031 35.8%, was 37.3%): content is a fixed budget.
• Tax. 19%, see section 3.
• Engagement report. The catalyst and falsifier were wrong: the report is annual, published in Q1, separate from results (Q2 letter, 2026-07-16). Fixed in section 6 and record.json.
• After hours and interview. After-hours price refreshed ($64.66 at 20:59:02Z); interview not found in readable form (section 8). The record price stays the close.

Sensitivity or note only (contested)

• Content cash. Holding the ratio at 1.10x every year, instead of 1.09 falling to 1.07: -$0.53. The 2026 guide and 2024 are 1.10-1.11 and 2025 was 1.08; not changed.
• Netflix's own beta. 1.613 raw, 1.411 Blume, cost of equity 11.21%: -$5.74 to $82.98, purchase price $80.14. The peers' median beta stays the method's rule; the view holds on either.

Not done. We did not give the bear and bull exit multiples of their own (the method does not allow it). The earnings interview was not read.