NFLX Deep Dive: The margin story is real, and the price already pays for most of it

2. Since the valuation date

The valuation in this report is dated 11-02-2026, when Netflix closed at $62.34, and nothing in it has been recalculated since. Two things in the market have moved. The stock closed at $63.95 on 11-09-2026, the settled close of that session, which is $1.61 or 2.6% above the valuation close and $0.54 or 0.9% above the $63.41 close of 11-06-2026, and the Treasury 10-year par yield fell from 4.91% to 4.86% by 11-06-2026, five basis points, having ranged between 4.84% and 4.95% on the days in between. We captured no yield after 11-06-2026. The 11-09-2026 close is the only new observation in this report: the filing index we retrieved on 11-07-2026 lists only insider filings, Forms 4 and 144, for the dates from 11-03-2026 to 11-05-2026, with no 10-K, 10-Q or 8-K among them, and we have not retrieved it again, so we say nothing about filings after that retrieval. The date of the fourth-quarter report is not confirmed in any document we captured, so we do not give one. Everywhere else in this report, "the price" means the $62.34 valuation close.

The intrinsic values are those of 11-02-2026. They have not been rolled forward to 11-09-2026, and the model's rate is still the 11-02-2026 build, so the only thing that differs in the table is the price each value is set against. Against the 11-09-2026 close the reference path is 6.8% above the price where it was 9.5% above at the valuation close and 7.7% above on 11-06-2026, the skeptic's path is 1.3% above where it was 3.9% and 2.2%, the downside is 24.5% below where it was 22.6% and 23.9%, and the upside is 32.9% above where it was 36.3% and 34.0%. Both earnings cross-checks, $47.33 and $42.80, are still below the price. None of the four values in the table, and neither cross-check, has crossed the close in either direction: the close sits 43% of the way from the downside value to the upside value, a position in an unweighted range and not a probability, with the reference $4.34 above it and the skeptic's path $0.84 above it.

The same reference path can also be read against each close by asking what the model needs in order to match it. The table solves the model's own reverse valuation twice, once at the valuation close and once at the latest settled close, changing one input at a time and holding the rest of the reference fixed, including the market weights and the share count of the valuation date. It is a statement about arithmetic and says nothing about what investors believe or which combination is likelier.

The higher close moves each of these break-evens toward the assumptions we hold. At the valuation close the model matched the price at a cost of equity of 10.49%; at $63.95 it matches at 10.31%, 0.45 points above the built 9.86% where the gap was 0.63 points, which is a beta of 1.20 where it was 1.24 at the same premium. Margins would have to be 2.0 points lower, not 2.7, or growth scaled by 0.83, not 0.76, so the latest close is matched by a smaller miss on either axis, and the reference's lead, though still positive, is that much thinner against the rate we judged.

The later sections read their stored values against the $62.34 valuation close, as they say, and four of those values lie between that close and the latest one. The upside at a beta of 1.30 and a premium of 5.5%, $62.56, the reference with the low growth path, $62.77, the reference with growth scaled by 0.8, $63.21, and the reference with margins one point lower and growth scaled by 0.9, $63.62, are each above $62.34 and below $63.95. The reference with margins two points lower, $63.96, is above the latest close by less than a cent. None of these values has been revised; they sit on the other side of a higher price.

The comparison does not change our reading. The stored sensitivities give the scale of the yield's move that we can see: the reference is $81.10 with the discount rate one point lower and $58.88 with it one point higher, so five basis points on the rate is worth about $0.56 a share. That is a scale and not a revaluation, a lower yield would add to the values rather than take from them, and it applies only to the move to 11-06-2026, since we captured no yield after that date. The reference path and the skeptic's path were above the price on the valuation date and are above the latest close, and the BALANCED reading rests on how small the reference's lead is against the two judged inputs to the discount rate, which a lead of 6.8% still is: at the 4.5% premium a beta of 1.20 would now remove it, where 1.24 would have at the valuation close. The rating stays as proposed at the 11-02-2026 valuation close, and the comparison with the 11-09-2026 close adds no argument for a different level.

We rate Netflix 3 of 5, BALANCED, as a proposal. The two levels either side were serious alternatives, and the case for the middle one rests on three facts that a reader can check against the sections above. The first is how small the lead is relative to the two inputs we judged. A beta of 1.20 with a premium of 5.0% puts the reference at $58.95, below the price, and the skeptic's path at $55.96; a beta of 1.00 with a premium of 4.0% puts the reference at $79.57. The rate is a judgment we have not measured, so a lead that a plausible change in it removes cannot support a FAVORABLE rating, and a lead that a plausible change in it more than doubles cannot support a CAUTIOUS one. The second is that the margin evidence and the growth evidence point in different directions, as sections 11 and 12 argue: the bull is stronger on the record of margin and the bear is stronger on the record of growth, and the reference sits between them without being a compromise. The third is that the price does not need strong execution, since it is matched by margins 2.7 points below the reference's with its growth, or by growth at 0.76 of its path with its margins, or by a cost of equity 0.63 points higher.

A move to 2 of 5 would follow from evidence that the reference's lead has gone, and the rate is the easiest way for that to happen without any operating news: at a cost of equity of 10.49%, which a beta of 1.24 at our premium gives, the reference equals the price. Operating evidence that would point that way is a 2027 revenue guide below 8.4%, a margin guide at or below the 31.5% of 2026, content cash held at 1.17 times amortization through 2027, or a further cut to the advertising guide. They do not reach equally far. Growth at 0.76 of the path, which cuts every year's growth by nearly a quarter, or margins 2.7 points lower in every year each brings the reference to the price on its own. Content cash works on a different scale: held at 1.17 times in 2027 alone it takes $0.30 off the reference, which is then $67.99, and held in all five explicit years $1.49, to $66.80, so what would remove the lead is 1.17 times becoming the lasting ratio, which the terminal year carries forever: held in every year including the terminal year, the reference is $58.48. A 2027 at 1.17 times matters to the view as possible evidence of that, not for its own cash. A move to 4 of 5 would follow from the opposite evidence together, a 2027 guide at or above 11.0% growth and 32.5% margin with content cash converging toward 1.05 times, or from a lower price: the rating is price-sensitive, and at the reference's value a price of $56.91 would put it 20% above the price, a level we offer as an illustration and not a threshold, because the methodology does not set one. The proposal does not change for a price move of a few percent, such as the 2.6% between the valuation close of 11-02-2026 and the settled close of 11-09-2026.

Why BALANCED, weighed both ways. Several facts support a higher reading. At the built central rate the reference is 9.5% above the price, and the skeptic's path, with slower growth and margin gains, is still 3.9% above it. The break-even cost of equity of 10.49% is 0.63 points above the built 9.86%, so a buyer at the price earns slightly more than the cost of equity if the reference path happens. The upside is 36.3% above the price, and net debt is only $6.9 billion. Other facts support a lower reading. The downside is 22.6% below the price, so the range is wide on both sides. Both earnings cross-checks, $47.33 and $42.80, sit below the price, 77% of the reference's value is the terminal value, and the reference's lead depends on two judged inputs: at a beta of 1.20 and a premium of 5.0% the reference is $58.95, which is 5.4% below the price, and the skeptic's path is 10.2% below it. On the house scale, BALANCED describes a base case near the relevant hurdle, or upside that uncertainty and downside substantially offset. A 0.63-point margin over the cost of equity, on a rate whose plausible band moves the reference from $51.20 to $85.46, is not a useful margin, which keeps the reading from FAVORABLE. A price that is matched by margins 2.7 points below the reference, or growth at 0.76 of it, does not need strong execution, which keeps it from CAUTIOUS.

The first row is the one most likely to move the explicit forecast. The fourth-quarter figures themselves say little. Management describes its guidance as its actual internal forecast, and its letters describe revenue and operating margin in each of the first three quarters of 2026 as in line with or above that forecast (revenue slightly above it in the first quarter, in line in the second and consistent with it in the third; margin slightly above, in line and ahead), which are the company's own descriptions and three quarters, not a record. For the fourth quarter it guides a margin of 26.0% against 24.5% a year earlier, a gain of 1.5 points that carries about 120 basis points of headcount cost, so the margin before that cost would be 27.2%, 2.7 points above the year before. The number that moves the reference is the 2027 guide, and the decision table below says what each result would do to the view: a guide at or above 11.0% growth and 32.5% margin strengthens the reference, and a guide near 8.4% or a margin that stays at 31.5% moves it toward the price or below.

The second thing the fourth quarter and the FY2026 Form 10-K will show is whether content cash comes back toward the guided 1.1 times. Over the first nine months it ran at 1.17 times, so a full year at 1.1 times needs the fourth quarter to spend about 0.92 times its amortization in cash, below it, on the amortization that management's guided growth of about 10% implies, and a full year at the nine-month 1.17 times needs the fourth quarter to match it. The first figure rests on two approximate guides and is a scale and not a forecast, but it says what the test is: whether Netflix lets cash fall below amortization in a quarter that carries three NFL games and a boxing match, by the October letter's account. The model prices the difference. The reference is $58.48 if 1.17 times holds in every year including the terminal year, and $68.72 to $66.80 across the explicit-year range of 1.08 to 1.17 times.

Advertising and competition are the slower tests. The 2026 advertising forecast has been cut from about $3.0 billion to about $2.7 billion, a cut of about $0.3 billion that is 0.6% of FY2026E revenue, and the launches in March 2027 are the first with the revised plan; a guide for 2027 that restores the earlier pace would support the upside's margin path, which is worth $6.82 of its value, and a further cut would not move revenue much but would weaken the case for margin beyond the reference. The merger of Paramount Skydance and Warner Bros. Discovery has no closing date in the filing and carries a $7.0 billion regulatory termination fee, with $46.7 billion of equity from the Ellison Parties named as its funding. We would watch it for evidence in Netflix's own numbers, not in the rivals': content cash held above amortization for longer, or a United States and Canada growth rate that keeps falling after the 2026 price increases have passed through.

The main risks are the ones already in the model: slower growth than the path (the price matches at about 0.76 of it), margin that stalls at this year's 31.5% (the downside), a higher cost of capital than we assumed, and content cash that stays above amortization. Outside the model are a rival that materially changes the cost of content or engagement, regulation and tax outcomes in Brazil and elsewhere, and the loss of the buyback tailwind if cash generation slows.

What would change the view is evidence on the same few drivers. Each row below is a test we can score later, not a forecast.

The cost-of-equity row has no weakening band of its own. The rate is our judgment and no report will state it, so a rate between the built 9.86% and the 10.49% break-even, which narrows the reference's lead without removing it, is counted once, as leaving the view unresolved; splitting it at the 10.2% break-even discount rate would borrow a figure measured on the whole firm and not on the equity.

Each row can also leave the view unresolved, and we say where, and what the model says about it.

Two cautions apply to the whole table. A single guide is one observation and is not a trend, so a result that strengthens or weakens the view in one quarter has not by itself changed the evidence. And the decision rows are written to be scored later against what the company reports, so a result that falls between two columns stays unresolved until a further result moves it.