Analysis of the Form 10-K | Fiscal year ended December 31, 2025
What does Netflix sell, to whom, and how does a subscription fee become money that belongs to the shareholder?
Netflix sells access to a catalogue of series, films, games and live programming for a monthly fee. Members are billed in advance, and the fee is recognized evenly across the month. In December 2025 the price of a paid plan ranged from the U.S. dollar equivalent of $1 to $37 a month, depending on country and plan. Advertising, consumer products and live experiences also produce revenue, but the 10-K says that revenue from sources other than membership fees "was not a material component" in any of the three years. The company reports a single operating segment and discloses revenue only by region.
| Region (US$ mn) | FY2025 | Share | FY2024 | FY2023 | Change 25/24 |
|---|---|---|---|---|---|
| United States and Canada | 19,957.2 | 44.2% | 17,359.4 | 14,873.8 | +15.0% |
| Europe, Middle East and Africa | 14,514.6 | 32.1% | 12,387.0 | 10,556.5 | +17.2% |
| Latin America | 5,357.5 | 11.9% | 4,839.8 | 4,446.5 | +10.7% |
| Asia-Pacific | 5,353.7 | 11.8% | 4,414.7 | 3,763.7 | +21.3% |
| Total | 45,183.0 | 100.0% | 39,001.0 | 33,640.5 | +15.9% |
The 2023 total covers streaming only. DVD revenue of $82.8 million that year brings the reported total to $33,723.3 million. The United States and Canada supplied 42.0% of the year's growth and Europe 34.4%. Latin America grew 10.7% in dollars against 23% at constant currency. Netflix had about 16,000 full-time employees at year end, up from 14,000. At board level, Reed Hastings moved from Executive Chairman to non-executive Chairman in April 2025, according to the 2026 proxy.
A single monthly fee travels like this. The member pays in advance, so the cash arrives before the service is delivered and creates a deferred revenue balance of about two weeks of revenue. Where a partner bills the member, for example in a telecom bundle with no standalone Netflix price, Netflix books only the net amount it receives from the partner. Payment processing fees, including those paid to billing partners, are charged to cost of revenues. Before Netflix keeps anything, the fee must cover the amortization of the content library, which absorbed 36.3% of revenue in FY2025, and other cost of revenues (production overhead, music rights, delivery through the company's own Open Connect network, customer service, payments), which absorbed another 15.2%. What had to exist before the sale is the library: $32.8 billion of capitalized content, 59.0% of total assets.
The evidence on pricing power is limited to what the 10-K shows. The top of the price range rose from $32 a month at the end of 2024 to $37 at the end of 2025. Management attributes revenue growth to "growth in memberships, price increases, and increased advertising revenue", but does not quantify the three parts. Technology and development spending held at 7.5% of revenue in both years. The company names its own sources of advantage ("winning moments of truth", its recommendation technology, its delivery network), but those are the company's claims. Partner agreements are "typically between one and three years in duration", and members "can change their plans at any time". There is no switching cost written into the contract. Whether the advantage is widening or narrowing cannot be established from the document, and from FY2025 it is harder to test because the membership and revenue-per-member data are no longer published.
A newcomer would read the regional table as a map of where the profit comes from. It is not: 95.9% of income before taxes is booked in the United States, while U.S. revenue is 40.9% of the total.
What kind of company is this, and where is the case decided?
Netflix is classified as solid. Revenue grew 15.85% in FY2025 and 15.65% in FY2024. Over the three years shown, operating margin rose from 20.6% to 29.5%. It is not a rapid-growth company in the sense used here: growth is mid-teens rather than high double digits, and the margin is not still forming. It is also not a cyclical company: no price, volume or replacement cycle appears in the document. A pending acquisition could change the classification. The WBD transaction, at $82.7 billion of enterprise value, is 1.49 times Netflix's total assets, so the balance sheet analysed here would not describe the company after a closing.
Three questions decide the case.
The first is where the margin comes from and whether it lasts. Operating margin rose 2.78 points, and nearly all of the gain comes from content amortization growing 7.3% against revenue growing 15.9%. The document allows this to be measured precisely. What it no longer allows is splitting revenue into members and price, so it cannot show whether the operating leverage comes from more members, higher prices or advertising. The question can only partly be answered.
The second is whether the profit is cash. Content is paid for faster than it is expensed, operating cash flow was 92.4% of net income, and a $619 million Brazilian tax charge was accrued but not yet paid. The document answers this one fully.
The third is what management is doing with the capital: $9.1 billion of share repurchases at an average of $105.79, above the year-end price of $93.76, and an agreement to buy WBD's studio and streaming businesses for about $72.0 billion in cash. The financing commitments reached $67.2 billion in January 2026, and none had been drawn at year end. This question can also only partly be answered, for two reasons. The 10-K gives no financial information on the business being acquired. And the 2026 proxy statement, the only later document attached, calls the transaction "previously proposed" without explaining what happened to it.
There is no segment and no volume metric to pair with a price, because Netflix stopped reporting memberships during FY2025. The pair that governs the model is therefore revenue per dollar of content amortization, together with the ratio of cash paid for content to the amortization booked. Both are built from published lines. Measures that do not apply here: inventory days, segment margins, dividend cover, goodwill impairment headroom (goodwill is not disclosed) and tangible book value (not calculable for the same reason).
The sections that go deep are section 3 (the margin) and section 8 (capital allocation and the WBD commitment).
Each of the three questions has a line in next year's filing that will answer it, listed in the closing block.
Which lines moved the result, and what moved them?
Management gives three explanations. (1) Revenue rose 16% "primarily due to the growth in memberships, price increases, and increased advertising revenue, partially offset by unfavorable changes in foreign exchange rates, net of hedging". (2) Operating margin rose about three points "primarily driven by the growth in revenues outpacing the growth in cost of revenues, sales and marketing, and general and administrative expenses". (3) Cost of revenues rose $2,237 million: $1,121 million from content amortization and $1,116 million from other cost of revenues, the latter "primarily driven by non-income tax assessments in Brazil".
| US$ mn | FY2025 | FY2024 | FY2023 | Change 25/24 | % of revenue 25 | % of revenue 24 |
|---|---|---|---|---|---|---|
| Revenues | 45,183.0 | 39,001.0 | 33,723.3 | +15.9% | 100.0% | 100.0% |
| Content amortization | 16,422.2 | 15,301.5 | 14,197.4 | +7.3% | 36.3% | 39.2% |
| Other cost of revenues | 6,853.2 | 5,736.9 | 5,517.9 | +19.5% | 15.2% | 14.7% |
| Sales and marketing | 3,301.3 | 2,917.6 | 2,657.9 | +13.2% | 7.3% | 7.5% |
| Technology and development | 3,391.4 | 2,925.3 | 2,675.8 | +15.9% | 7.5% | 7.5% |
| General and administrative | 1,888.4 | 1,702.0 | 1,720.3 | +11.0% | 4.2% | 4.4% |
| Operating income | 13,326.6 | 10,417.6 | 6,954.0 | +27.9% | 29.5% | 26.7% |
| Provision for income taxes | (1,741.4) | (1,254.0) | (797.4) | +38.9% | ||
| Net income | 10,981.2 | 8,711.6 | 5,408.0 | +26.1% | 24.3% | 22.3% |
The complete statement, every line and every year with its source, is in Financials → Income statement on this company's page.
The statements confirm the second and third explanations, and the third needs a qualification. The first cannot be verified, because the 10-K no longer gives members, price or advertising revenue. Of the 2.78-point margin gain, content amortization contributed +2.89 points. Sales and marketing and general and administrative added about 0.2 points each, and other cost of revenues subtracted 0.46. Excluding the Brazil charge, other cost of revenues grew 8.7% and fell to 13.8% of revenue, which would have added about 0.9 points of margin instead of subtracting 0.46. Operating margin excluding the charge would have been 30.9%. Note 9 calls the amount the "cumulative loss" recognized in the third quarter but does not say how much of it relates to earlier years. In the MD&A, management does not expect Brazilian non-income taxes to "materially impact our results of operations in future periods", while Note 9 says the company continues "to accrue incremental non-income taxes".
Revenue per dollar of content amortization is how many dollars of subscription and advertising revenue the company collects for each dollar of its content library that it writes off in the year. Content amortization is the share of the cost of licensed and produced titles charged to the income statement, front-loaded according to how the company expects members to watch each title. The ratio therefore measures how hard the library is being worked: a rising ratio means each dollar of content supports more revenue. Revenue comes from the statement of operations. Content amortization comes from Note 5 and the cash flow statement, and Note 13 shows it as a separate line. On the same denominators, cash paid for content is the additions to content assets plus the decrease in content liabilities, both from the cash flow statement.
| Pair | FY2025 | FY2024 | FY2023 |
|---|---|---|---|
| Revenue / content amortization | 2.751 | 2.549 | 2.375 |
| Cash paid for content | 17,707.5 | 17,002.8 | 13,140.3 |
| Cash paid for content / content amortization | 1.078 | 1.111 | 0.926 |
| Revenue / cash paid for content | 2.552 | 2.294 | 2.566 |
The ratio rose by 0.20 in FY2025 and by 0.17 in FY2024. It can be split along the two kinds of content that Note 5 separates. Licensed content amortization grew 13.3%, close to revenue. Produced content amortization grew 1.3%, or 2.3% before tax incentives. Those incentives reduced produced amortization by about $1,000 million in FY2025, against $899 million in FY2024. Of the 2.89 points by which content amortization fell as a share of revenue, produced content accounts for 2.46 and licensed content for 0.43, so about 85% of the leverage comes from the produced library. That is the part Netflix owns, pays for during production and amortizes from release. In FY2024 cash paid for content jumped 29.4% while amortization grew 7.8%, and the FY2024 10-K attributes this only to higher payments for content assets, without giving a cause. In FY2025 cash grew 4.1% and the ratio fell back to 1.078.
The mechanism is the one the 10-K describes in its risk factors: content payments "are not tied to member usage or the size of our membership base". A content bill set years in advance, with $24.0 billion of obligations at year end, is spread over whatever revenue the year produces. When revenue rises 15.9% while the content bill grows 7.3%, much of the difference reaches operating income: revenue rose $6,182.1 million and operating income $2,909.0 million. The same mechanism works in reverse, which is why the risk factors warn that margins "may be adversely impacted" if growth falls short. Nothing in the amortization schedules suggests the leverage comes from stretching the write-off period. The share of the released produced library expected to be amortized within twelve months rose from 38.9% to 40.3%, and the three-year share was stable at about 83%.
The library grew more slowly than the revenue it supports. Content assets rose 1.0% in FY2025, so they fell from 83.2% to 72.5% of annual revenue. Titles in production also fell slightly, from $9,317.4 million to $9,210.7 million. This is the same fact as the ratio seen from the balance sheet: additions to content ran at 1.04 times amortization, so the library barely grew while revenue rose. Whether a library that is no longer growing on the balance sheet can keep supporting revenue growth is not something the document allows anyone to determine. It shows the balance, not the viewing.
Netflix does not publish the parts that would turn this ratio into a full explanation: memberships, average revenue per membership, advertising revenue, the split between price and volume, and content spending by region or by type of title. The last published figures, for FY2024, were 301.6 million paid memberships and an average monthly revenue per membership of $11.70. The 10-K says the company now focuses "on revenue and operating margin as the primary financial metrics". The consequence for the reader is concrete: the FY2025 margin gain cannot be traced to price, volume or advertising, the three causes management lists.
The year was organic. The only business combination was a $28 million cash acquisition in December 2025. Currency took about one point off growth, since constant-currency growth was 17%. Hedging moved the other way between years: a $124 million gain in FY2024 became a $91 million loss in FY2025. Scalability shows up line by line against revenue growth of 15.9%. Cost of revenues grew 10.6% and general and administrative 11.0%. Sales and marketing grew 13.2%, with advertising spend up 12.5% and more advertising-sales staff. Technology and development grew 15.9%, in line with revenue. Other cost of revenues grew 19.5%, or 8.7% without the Brazil charge.
Below operating income, interest expense rose 8.0% to $776.5 million, including about $60 million of amortized fees on the WBD financing commitments. Interest and other income fell to $172.5 million because of $123 million of foreign exchange losses. The effective tax rate rose from 12.6% to 13.7%, which is the main reason net income grew 26.1%, less than operating income at 27.9%.
The year's margin story is the content line: a library growing slower than the revenue it supports. The only cost line that grew well ahead of revenue did so because of a tax assessment that management does not expect to weigh materially in future periods.
Is the profit cash, and does it belong to the company?
All of it belongs to Netflix's shareholders: the group is wholly owned and has no non-controlling interests. Operating cash flow was $10,149.3 million against net income of $10,981.2 million, 92.4% of profit. The gap comes mainly from the content line, which is classified as operating. Netflix paid $17,707.5 million for content and expensed $16,422.2 million, so $1,285.3 million of profit went back into the library. Other non-cash items, mainly unspecified "other non-cash items", stock compensation and depreciation, added $1,351.6 million. Deferred taxes subtracted $442.1 million and working capital absorbed $456.2 million. Within working capital, other current assets used $790.7 million and other non-current items $793.7 million. Accrued expenses supplied $881.2 million and deferred revenue $254.9 million.
Three questions follow.
How much of operating cash flow is other people's money held temporarily? Accrued expenses, deferred revenue and accounts payable together supplied $1,128.1 million, 11.9% of free cash flow. Netting the $610.8 million used to reduce content liabilities leaves $517.3 million, 5.5%. Accrued expenses deserve a closer look. Excluding leases and derivatives, they rose $912.2 million. The 10-K does not say in which line the Brazil accrual sits, but it says the company expects to pay deposits of "approximately $700 million" on that matter. Part of FY2025's cash flow is therefore a tax that has been expensed but not yet paid.
How far are taxes paid from taxes expensed? Cash taxes were $2,220.4 million against a provision of $1,741.4 million, 27.5% higher. The difference is almost entirely the deferred tax benefit of $449.1 million, since the current provision of $2,190.5 million is within $29.9 million of the cash paid. The company also paid about $200 million of deposits on direct taxes beyond its normal obligations.
Does reported capex cover the investment? It does not. Purchases of property and equipment were $688.2 million, 1.5% of revenue. The real investment is the content payment described above, which runs through operating cash flow. Leases added $465.4 million of right-of-use assets without any cash outflow. A reader who takes free cash flow at face value is already reading a figure net of content investment, and that is correct for this business.
Owner earnings, defined as net income plus non-cash items minus capex and minus working capital absorbed, with content amortization replaced by cash paid for content, come to $9,548.8 million before working capital and $9,092.6 million after. The judgement here uses the figure after working capital. It is lower than free cash flow of $9,461.1 million by exactly the $368.4 million of stock-based compensation, which this definition treats as a real cost rather than adding it back. Before working capital, owner earnings exceed free cash flow by $87.8 million, because the working capital absorbed ($456.2 million) was larger than stock compensation.
Accounting choices that move profit are concentrated in content. The amortization pattern rests on estimated viewing; tax incentives reduce the cost of titles; and no impairment indicator has been identified "to date" for the library, which is reviewed only as a group, although titles that are abandoned are written off. Non-GAAP reporting is limited to constant-currency revenue: the 10-K presents no adjusted earnings and no list of exclusions. The 13.7% effective tax rate sits below the 21% statutory rate. The foreign-derived intangible income deduction accounts for 5.2 points of the gap, and share-based payments for 3.1 points. Foreign tax credits subtract 2.3 points, while Brazilian withholding tax adds 1.9.
| Early-deterioration comparison | Growth of the item | Growth of the reference | Gap (pts) |
|---|---|---|---|
| Trade receivables vs revenue | +52.1% | +15.9% | +36.3 |
| Allowance vs gross receivables | not disclosed | ||
| Accounts payable vs cost of revenues | +0.1% | +10.6% | −10.6 |
| Additions to content vs content amortization | +5.4% | +7.3% | −1.9 |
| Capex vs depreciation | +56.6% | +1.4% | +55.2 |
| Operating cash flow vs revenue | +37.9% | +15.9% | +22.0 |
Four gaps exceed five points. Trade receivables are the one that needs explaining. They rose from 12.5 to 16.4 days of revenue. The note says receivables "primarily consist of membership and advertising fees" but does not split them and does not disclose an allowance, so whether the increase comes from advertising customers paying on longer terms, from billing partners or from timing cannot be determined. The payables gap points the other way from the usual warning: suppliers were paid faster, which used cash rather than borrowing it. The capex gap is large in percentage but small in dollars ($248.7 million more capex). Inventory is not applicable. Operating cash flow grew faster than revenue for three reasons: net income (+$2,269.6 million), a smaller excess of content cash over amortization (+$415.9 million) and the unpaid Brazil accrual, which the document does not quantify, while working capital as a whole was $423.1 million less favourable than in FY2024. Working capital and reserves therefore do not explain the growth in free cash flow.
The profit is cash, provided free cash flow is read as already net of the library investment. The two items to watch are a receivables balance growing more than three times as fast as revenue and a tax accrual that turns into cash outflow next year.
What stands in front of the shareholder, and what is it worth?
Content assets of $32,778.4 million are 59.0% of total assets. Licensed titles account for $12,138.6 million and produced titles for $20,639.8 million, of which $9,952.4 million (30.4% of all content) is still in production or development and has not yet earned revenue. Cash and short-term investments of $9,062.4 million (16.3%) are almost all available: restricted cash is $5.5 million. Other current assets of $3,957.8 million (7.1%) are half trade receivables, at $2,031.5 million. The rest includes prepaid expenses and $552 million of tax-incentive receivables.
Other non-current assets of $7,794.1 million (14.0%) can be identified only in part. Right-of-use assets, net deferred tax assets and derivatives account for $4,358.4 million. The remaining $3,435.7 million, 6.2% of total assets, is not broken down by any note. It may include goodwill and acquired intangibles, which the 10-K never discloses as separate lines; the only trace is a $261.5 million deferred tax liability on "acquired intangibles". There is therefore no goodwill, no impairment test and no tangible equity figure to report. Net deferred tax assets of $1,954.8 million depend on future taxable income. Their largest component is capitalized research ($969.2 million), and the valuation allowance of $617.6 million relates mainly to California R&D credits, state loss carryforwards and U.S. foreign tax credits the company does not expect to use.
On the liability side, measured against total liabilities of $28,981.5 million, long-term debt is 46.5% and content liabilities 19.5% in total. Accrued expenses are 11.1%, other non-current liabilities 10.2% and deferred revenue 6.1%. Debt consists of twelve fixed-rate senior unsecured notes with $14,519 million of principal and a principal-weighted coupon of 4.75%. Maturities run from November 2026 to 2054, and 71.8% of principal is due by 2029. Euro notes total €4,700 million, and the schedule by year is in the data tables. The notes carry covenants on liens and mergers but no financial-ratio covenant. The revolving facilities require consolidated EBITDA of at least three times interest expense, and the company states it is in compliance. Undrawn at year end were the $3.0 billion revolving credit facility and the $3.0 billion commercial paper programme. Also undrawn were $59 billion of WBD-related facilities and bridge commitments, which rose to $67.2 billion in January 2026. Operating lease liabilities of $2,513.0 million are kept outside debt in this report; including them, gross debt is $16,975.8 million. Off the balance sheet sit $18.4 billion of content commitments not yet recognized, plus an estimated $1 to $4 billion for unknown future titles over three years.
Working capital is measured in days on year-end balances. Receivables went from 13.9 to 12.5 to 16.4 days of revenue. Payables went from 13.8 to 15.6 to 14.1 days of cost of revenues. Deferred revenue held between 14 and 16 days of revenue. Content liabilities fell from 195.7 to 116.8 days of cash paid for content over the three years: Netflix carries less unpaid content on its balance sheet each year.
Equity of $26,615.5 million is net of $22,372.7 million of treasury stock (40.2% of total assets) and a $580.4 million accumulated comprehensive loss. Return on average equity is 42.8%. Adding treasury stock back, it is 25.3%, and FY2024 was 26.6%. The difference is the repurchase programme: the 42.8% reflects the shrinking of the equity base as much as the profit. Retained earnings rose $10,981.2 million, all of net income, because Netflix pays no dividend. Here that line describes the earnings faithfully, but retained earnings are never read alone: repurchases are booked separately in treasury stock.
Interest expense is 5.8% of operating income. Long-term debt equals 1.23 years of current net income. Net debt is 0.40 times operating income plus depreciation.
About one-sixth of total assets (cash, money market funds, time deposits, derivatives) is measured at an observable price. The 59.0% in content is carried at cost less amortization driven by management's viewing estimates.
If growth stopped, what would protect the shareholder? Equity is $6.30 per year-end share, and net working capital is $2,039.3 million, a current ratio of 1.19. Cash is $2.15 per share, or −$1.28 per share net of gross debt. Protection would come from the library and the cash flow it produces, not from net cash or tangible assets. Cash already does not cover gross debt, and the $24.0 billion of content commitments would remain payable whatever revenue did.
Does the capital earn a return, and does the shareholder receive it?
Per-share values use diluted weighted-average shares.
| US$ per diluted share | FY2025 | FY2024 | FY2023 |
|---|---|---|---|
| Revenue | 10.40 | 8.88 | 7.50 |
| Net income | 2.53 | 1.98 | 1.20 |
| Free cash flow | 2.18 | 1.58 | 1.54 |
| Stockholders' equity | 6.13 | 5.63 | 4.58 |
The diluted share count fell 2.28% and then 1.11%, so every per-share line grew faster than the corresponding total. Revenue per share rose 17.2% against total revenue growth of 15.9%. Free cash flow per share barely moved in FY2024 and rose 38.2% in FY2025. The shareholder moved forward faster than the company.
Return on average invested capital, defined as operating income after the effective tax rate divided by equity plus gross debt minus cash and short-term investments, was 36.7% in FY2025 and 31.0% in FY2024. This is our construction, not a figure the company publishes. Its limit is that invested capital takes the content library at book value, after an amortization pattern the company sets. The incremental return from FY2023 to FY2025 was 135.3%, because NOPAT rose by more than the capital added. With only two increments (110.8% and 188.2%), this is two data points, not a trend. An incremental return far above the average return describes a business whose added profit comes mostly from capital already in place, which is the content leverage measured in section 3.
The reinvestment rate, combining net capex, the excess of content cash over amortization and working capital absorbed, was 18.2% of NOPAT. At a 36.7% return, that implies 6.7% growth, against realized operating income growth of 27.9%. The difference came from operating leverage on a content base growing more slowly than revenue, not from reinvestment. A business earning 36.7% on its capital and reinvesting 18.2% of what it earns does not compound that return inside the company: it collects it, and over three years it handed the rest back through repurchases.
The company is self-financed. Cumulative free cash flow over FY2023-FY2025 was $23,308.6 million. Repurchases took $21,436.3 million (92.0%), net debt repaid was $439.0 million, and employees paid in $1,669.8 million on option exercises. No equity was issued except to employees.
Capital intensity depends on where the spending is counted. Capex over three years was 5.9% of cumulative net income, and capex was 2.06 times depreciation in FY2025. Calling Netflix capital-light on that basis would be misleading, because growth spending runs through content: $47,850.5 million paid over three years, and in FY2025 alone an amount equal to 39.2% of revenue. The document does not separate maintenance from expansion. Using content amortization as an approximation of the cost of maintaining the library (an approximation, not a disclosed figure), the expansion portion was $1,285.3 million in FY2025.
The multiples use the closing price of $93.76 on December 31, 2025, the last trading day of the fiscal year, and the 4,222,162,150 shares outstanding on the same date. Both were supplied with the request and have not been updated, so numerator and denominator describe the same moment. The resulting market capitalization is $395.9 billion and the enterprise value, with leases excluded, $401.3 billion. At this price the buyer pays 37.1 times diluted earnings and 41.8 times free cash flow. A buyer of the whole company would earn 3.32% on enterprise value in operating income before financing and taxes, or about 3.5% with the Brazil charge added back, since management does not expect it to weigh materially in future periods. Book value is 14.9 times. Any return beyond that operating yield would have to come from growth, which the document does not forecast.
The return figures above describe a business earning high returns on existing capital and returning most of its cash. The year-end balance sheet, and therefore these multiples, do not contain the capital the WBD transaction would require.
Which notes change the reading, and by how much?
The statements say how much. The notes say how much of that figure is management's opinion. Four notes qualify: the only Critical Audit Matter, two notes that weigh more than 5% of income before taxes or of assets (income taxes and the WBD commitments), and one note whose amount rests on a discretionary judgement (Brazil).
Content amortization, the Critical Audit Matter (Note 1, Note 5). What it is worth: $16,422.2 million of expense, 36.3% of revenue, on a $32,778.4 million asset. How much discretion it contains: amortization follows "historical and estimated viewing patterns", over the shorter of the licence window, the estimated period of use or ten years, on an accelerated basis, with "over 90%" expected to be amortized within four years of release. Ernst & Young chose it as the matter involving "especially challenging, subjective or complex judgments" and tested it by comparing past estimates of viewing patterns with actual results. What would move it: viewing that falls off faster than estimated would bring expense forward and lower margins. The company discloses no sensitivity, so the size of that effect cannot be calculated. The expected-amortization schedules show no lengthening from FY2024 to FY2025, and no group impairment has been recorded "to date".
Brazil non-income taxes (Note 9). What it is worth: about $619 million cumulative, recognized in the third quarter, 4.6% of operating income, plus about $700 million of expected deposits. How much discretion it contains: the trigger was "developments in another taxpayer's judicial proceedings", which led the company to judge a loss "probable". The FY2024 10-K had described "meritorious defenses" and an exposure of "approximately $400 million". What would move it: further assessments, which the company says it continues to accrue, or a final ruling in either direction. No range is given.
Income taxes (Note 11). What it is worth: a provision of $1,741.4 million and unrecognized tax benefits of $566.4 million, up from $432.3 million. Of those benefits, $336 million would lower the effective rate if recognized. How much discretion it contains: the IRS is examining 2016 through 2022, and the rate depends on the foreign-derived intangible income deduction, which is worth 5.2 points. What would move it: a change in that regime, or an adverse audit. Uncertain tax benefits grew 31.0% in the year, and accrued interest and penalties rose from $44 million to $82 million.
The WBD agreement and its financing (Note 6, Note 7, Note 9, Note 14). What it is worth: $72.0 billion of equity value and $82.7 billion of enterprise value, paid in cash at $27.75 per WBD share. A $5.8 billion termination fee may be payable by Netflix under specified circumstances, equal to 64.0% of year-end cash and short-term investments. How much discretion it contains: apart from about $60 million of financing fees expensed in interest, the FY2025 statements contain nothing related to it, and the price may change with WBD's capitalization at closing. What would move it: regulatory approval, the WBD shareholder vote, the separation of Discovery Global, and the terms of the debt actually raised, which the risk factors say "may include secured debt and/or debt with restrictive covenants".
There are no defined-benefit pension plans. Multiemployer contributions of $220.3 million carry the usual exposure to other employers' underfunding, which the note describes without quantifying.
The notes move the reading in two directions. The content estimate is the largest judgement and shows no sign of having been stretched. The tax and WBD notes add obligations that the year's operating figures do not yet carry.
What did management do with the money, and do they say so when things go badly?
Netflix has one class of shares, one vote per share, and no controlling shareholder. According to the 2026 proxy, three asset managers report holdings above 5%, based on positions dated December 2023. The board has twelve nominees for the June 2026 meeting. Reed Hastings, founder and Chairman, will not stand for re-election. Ted Sarandos and Greg Peters are co-chief executives and directors. Ernst & Young has audited the company since 2012, its fees were $11.9 million, of which $3.0 million was for tax work, and it issued an unqualified opinion on internal control. The proxy reports two related-party arrangements. A co-CEO's son worked as an editor on productions, with pay above $120,000. A co-CEO's wife was engaged as producer on two productions, with up to $1.3 million contingent on milestones, of which $6,250 had been paid.
Directors and executive officers as a group beneficially owned 52.4 million shares at April 6, 2026, 1.24% of the class. That includes 27.3 million options and 0.1 million RSUs vesting within 60 days, so about 24.9 million shares are held outright. Most of the holding comes from Hastings: 37.8 million, of which 16.6 million are options. For the co-CEOs, options are most of the holding: 5.38 million of Sarandos's 5.72 million and 2.57 million of Peters's 2.75 million. The proxy does not say how the shares held outright were acquired, and neither document reports open-market purchases. According to the 10-K, Peters adopted a 10b5-1 plan in October 2025 covering the possible sale of up to 2.95 million shares through 2029.
Where the money went is set out in section 6: almost all free cash flow into repurchases, no dividends, and $17.2 million of acquisitions in cash. The $28 million December acquisition is too small to have left anything on the balance sheet that can be tested.
Repurchases in FY2025 covered 86.5 million shares at a cost of $9,154.9 million, the equity-statement basis used here, which includes direct costs. The cash outflow was $9,127.2 million, and the document does not reconcile the difference. Dividing gives an average price of $105.79, 12.8% above the year-end close of $93.76. In every month of the fourth quarter the average price paid was above the close: $117.27 in October, $108.53 in November and $103.89 in December. Measured against free cash flow per diluted share of the same year, Netflix bought its own shares at 27.3 times in the first year shown and 48.6 times in the last, with 40.2 times in between. Spending on repurchases rose from $6.0 billion to $9.1 billion while the multiple paid rose from 27 to 49 times free cash flow, so spending did not fall as the shares became more expensive relative to the cash they produce. All of it was paid from free cash flow. No debt was raised in FY2025, and $1,833.5 million of notes was repaid.
On net dilution, 31.6 million shares were issued to employees, mainly on exercise of options at an average strike of $21.73, against 86.5 million repurchased. Net shares outstanding fell by 55.4 million, or 1.30%. Repurchases are 24.8 times the stock compensation expense, so they return capital rather than merely offset dilution. Still, 36.5% of the shares bought replaced shares issued to employees in the same year. The expense line understates the economics of those options: their intrinsic value at exercise was $2,560.9 million, against $368.4 million of expense. The 127.7 million options still outstanding, all vested, equal 3.02% of shares outstanding.
On accounting aggressiveness, Netflix reports no adjusted earnings, so there are no recurring "non-recurring" exclusions to add up. The only judgement-driven adjustment visible in the filings is in pay: the annual bonus is measured on F/X-neutral revenue (35%) and F/X-neutral operating margin (65%), and the proxy certifies the margin at 28.10% against a reported 29.5%. The proxy does not say whether the Brazil charge is inside or outside that figure, so it cannot be determined whether management's bonus bore the cost of the charge.
Governance signals in the documents: internal control effective, no auditor change, no restatement, and no pledge of shares disclosed. One item is unusual. Lead independent director Jay Hoag did not receive a majority of votes at the 2025 meeting, which the company attributes to his attending fewer than 75% of meetings in 2024. The board rejected his resignation in June 2025.
Taken together, the year's decisions say more than any single one. In October 2025 the board approved the 10-for-1 split. In November it amended the executive severance plan, effective January 2026: termination without cause outside a change of control now pays two years of salary and target bonus, and retirement allows continued vesting. On December 4 Netflix agreed to buy WBD's studios and streaming businesses for about $72.0 billion in cash and signed $59 billion of bridge commitments the same day. On December 19 it replaced $25 billion of the bridge with a revolver and term loans. In January it amended the agreement and raised the bridge to $42.2 billion. In the same quarter it kept buying back 18.9 million shares at prices above the year-end close. Netflix closed FY2025 with $9.1 billion of cash and $14.5 billion of notes. Within weeks it had commitments to borrow up to $67.2 billion more, 4.6 times its existing note principal. The 2026 proxy refers to the "previously proposed acquisition" without explaining the change. The attached documents do not establish whether the transaction was completed, amended or abandoned, or whether the $5.8 billion fee became payable.
Compensation (proxy, step 1): the co-CEOs each received about $53 million in reported 2025 pay, $41.4 million of it in RSUs and PSUs. PSUs depend on total shareholder return relative to the S&P 500 over three years, and the cash bonus on F/X-neutral revenue and margin. Neither metric measures return on capital, per-share cash flow or the price paid for repurchases or acquisitions, which are the decisions examined in this section.
On candour, the 10-K contains one revised judgement stated as such: the Brazil matter, where the company moved from "meritorious defenses" to "probable" loss and named the trigger. The discontinuation of membership reporting is presented as a change of focus, not as an admission. There is no impairment, no project described as closed, and no target described as missed, which is an observation rather than an accusation.
The decisions read together point in one direction: from returning nearly all cash through repurchases to committing it, and a large amount of borrowed money, to an acquisition. The year-end balance sheet does not yet show that change.
What would have to happen for capital to be permanently lost here?
Compared with the FY2024 10-K: a new section on the WBD transaction adds two risks, and the debt risk now describes the bridge and term financing. The labour risk now names the 2026 expiry dates, the reputation risk now covers advertisements, and the advertising risk is no longer described as "new". The regulatory risk records the July 2025 EU statement on network fees. No risk-factor heading present in the FY2024 10-K is missing from the FY2025 one.
Pre-mortem: three years from now the investment has lost capital permanently. What happened?
First, the acquisition was completed and did not earn its cost. Today's measure: $82.7 billion of enterprise value against Netflix's operating income of $13.3 billion. The signal would be goodwill and debt appearing on the balance sheet while revenue per dollar of content amortization (2.751 today) falls as the two libraries are combined.
Second, revenue growth slowed while the content bill stayed fixed. Today's measure: content amortization growing 7.3% against revenue at 15.9%. The signal would be content amortization as a share of revenue rising from 36.3%, with cash paid for content staying above amortization.
Third, taxes took back part of the margin. Today's measure: an effective rate of 13.7% against 21% statutory, and $566.4 million of uncertain positions. The signal would be growth in uncertain tax benefits, accrued expenses rising without a matching cost explanation, or a jump in the effective rate.
The first scenario depends on a transaction whose status the attached documents leave open. The other two can be read directly from lines the company will publish next year.
What did the company promise, and what came of it?
The catalysts the document names are these. (1) The WBD transaction, which the company expected to close "12-18 months from December 4, 2025", subject to regulators, WBD shareholders and the separation of Discovery Global. (2) The renewal of the WGA, SAG-AFTRA and DGA contracts in 2026. (3) The Brazil deposits of about $700 million. (4) The $8.0 billion left on the repurchase authorization. (5) Unknown content obligations of $1 to $4 billion over three years. (6) The new expense-disaggregation standard (ASU 2024-03), effective for fiscal years beginning after December 15, 2026, which will require more detail on costs. None of these catalysts controls revenue per member, which the company no longer discloses.
The declared expectations of a year ago can be checked line by line against what happened. The company's own forecasts of revenue or margin are not in the 10-K, so the comparison is limited to amounts the FY2024 filing committed to in writing.
| What the FY2024 10-K said | What happened in FY2025 | Gap |
|---|---|---|
| Brazil non-income tax exposure "approximately $400 million", "expected to increase" | Cumulative loss of about $619 million recognized | Above the stated exposure by about $219 million |
| Possible deposits of "approximately $800 million" on direct and indirect taxes within twelve months | About $200 million of direct-tax deposits paid; about $700 million on Brazil still expected | Not reconcilable line by line; most of the amount is still described as expected |
| Debt principal and interest due within twelve months: $2,486.9 million | Repaid $1,833.5 million plus interest paid of $718.6 million | In line |
| Content obligations not on the balance sheet: $17.0 billion | $18.4 billion | +8.2% |
| Unknown future titles: $1 to $4 billion over three years | Same range repeated | No change |
| Repurchase authorization remaining: $17.1 billion | $9.1 billion used; $8.0 billion remaining | Consistent |
This is the first year of coverage, so the comparison is with the company's previous filing. From next year, the closing block below will be the baseline.
The three questions in section 2 can be answered as far as the document allows, and the third remains open.
On the margin: the gain is real and comes from one source. Revenue per dollar of content amortization rose from 2.375 to 2.751 over three years, and the produced library generated about 85% of the FY2025 leverage. The amortization schedules show no stretching. The reading holds as long as content amortization keeps growing more slowly than revenue. It fails if the ratio stalls while cash paid for content stays above amortization, and the ratio line next year will show which. What the reading cannot provide, because the company stopped disclosing it, is the cause of the revenue growth.
On cash: operating cash flow is 92.4% of net income after paying for content, and owner earnings of $9.1 billion are close to free cash flow. Two items qualify that: a $619 million tax accrued but not yet paid, and trade receivables growing 52.1%. The reading holds if the receivables note explains the increase and the Brazil deposits are paid without a new charge. It fails if receivables keep outgrowing revenue without an explanation in the note.
On capital: until December 2025, the answer was a self-financed company returning 92.0% of its free cash flow through repurchases, at a price rising faster than cash flow per share. The WBD agreement replaces that pattern with a borrowed purchase larger than the company's balance sheet, whose outcome the attached documents do not establish. The reading of section 6, a high-return business that distributes its cash, holds only if the transaction did not complete or completes without the debt and goodwill overwhelming the content leverage. The debt line, goodwill and interest expense in next year's filing will settle it.
Each strength rests on a figure already shown in the body.
The weaknesses are the same figures read from the other side.
The document leaves five things unsaid that would change the judgement.
Two of the five open questions concern the transaction and the incentives around it, and none concerns the core economics of the library, which the document measures well. The case is therefore decided less by what Netflix earns on its content than by what it does with the money that content produces.
What next year's filing will settle
- Revenue / content amortization: 2.751 today. Whether the margin leverage continues or reverses (question 1).
- Cash paid for content / content amortization: 1.078 today. Whether the library is growing faster than it is being expensed (questions 1 and 2).
- Gross debt: $14,462.8 million today, with bridge commitments of $42.2 billion undrawn. Whether the WBD transaction was financed and completed (question 3).
- Trade receivables: $2,031.5 million today, 16.4 days of revenue. Whether the 52.1% increase was timing or a change in who pays and when (question 2).
- Accrued expenses and other liabilities: $3,220.9 million today. Whether the Brazil accrual turns into cash deposits without a new charge (question 2).
This report is based on the Netflix, Inc. Form 10-K for the fiscal year ended December 31, 2025, signed January 23, 2026, with the Form 10-K for FY2024 and the 2026 Proxy Statement used for comparison and governance information. Multiples use the closing price of US$93.76 on December 31, 2025 and 4,222,162,150 shares outstanding at the same date, both supplied with the request and not updated. This report is for analytical purposes only and does not constitute investment advice.
The full data table behind this report — every line item with its source in the filing and the arithmetic checks — is published separately: Netflix, Inc. FY2025 data table.