I. Transfer Pricing: Netflix India characterized as Limited Risk Distributor; TNMM upheld over TPO’s Royalty Model
Issue
Characterization: Whether the assessee (Netflix India) should be characterized as a Limited Risk Distributor (LRD) of access to streaming services or as an entrepreneurial service provider owning intangible assets.
Methodology: Whether the Transactional Net Margin Method (TNMM) using software distributors as comparables is the Most Appropriate Method (MAM), or if the “Other Method” (Royalty model) applied by the TPO is correct.
Facts
Business Model: The assessee, an Indian company, was appointed as a non-exclusive distributor for the ‘Netflix Service’ in India by its Associated Enterprises (AEs). It remitted distribution fees to its AEs.
Assessee’s Benchmarking: The assessee characterized itself as a distributor and applied TNMM, selecting distributors of software and related products as comparable companies.
TPO’s Re-characterization: The Transfer Pricing Officer (TPO) rejected this characterization. The TPO alleged that the assessee was an “entrepreneurial provider” because it:
Maintained “Open Connect Appliances” (OCAs – servers) in India, which the TPO equated to owning platform technology.
Allegedly controlled pricing, legal risks, and customer support.
TPO’s Adjustment: The TPO applied the “Other Method”, treating the transaction as a licensing of content/technology, and computed a “blended royalty” of 57.12% of the revenue, leading to a significant adjustment.
Decision
Functional Reality: The Tribunal held that the distribution agreement and terms of use clearly established the assessee as a distributor. It had no ownership rights to the content or the proprietary technology, which remained with the AEs. The OCAs (servers) were merely hardware to cache content for better bandwidth and did not constitute ownership of core technology.
Functions Performed: The assessee’s functions were limited to promotion, marketing, invoicing, customer support, and regulatory compliance. It did not bear significant entrepreneurial risks (like content failure risk) or develop Intangible Property (IP).
Rejection of “Other Method”: The TPO’s “Other Method” failed to meet the criteria of Rule 10B(2). The TPO did not demonstrate how the third-party royalty agreements used were comparable to the assessee’s transaction. There was a lack of functional, geographical, and market similarity.
Comparability Standard: The Tribunal ruled that functional similarity is the touchstone of comparability, not just the sectoral label. In the absence of direct media-content distributor comparables, software distributors serve as valid analogues because both distribute digital access/intangible products.
Final Ruling: The assessee is a Limited Risk Distributor. TNMM is the Most Appropriate Method. The TPO’s adjustment based on the royalty model was set aside in its entirety.
Key Takeaways
Functional Profile is Key: Re-characterization of an entity by the Revenue from a “Distributor” to an “Entrepreneur” requires evidence that the entity actually owns critical IP or assumes significant risks (DEMPE functions). Mere ownership of local infrastructure (like servers) does not equate to ownership of core technology.
Software Distributors as Comparables: In the digital economy, software distributors can be valid comparables for entities distributing access to digital content (like streaming services) due to functional similarities in distribution mechanics.
TDS & VsV Scheme: Settlement of TDS appeal cures Section 40(a)(i) disallowance
Issue
Whether a disallowance of expenditure made under Section 40(a)(i) (for failure to deduct TDS) survives if the assessee settles the underlying appeal regarding the TDS default (Section 201 order) under the Direct Tax Vivad Se Vishwas (VsV) Scheme.
Facts
The Disallowance: The Dispute Resolution Panel (DRP) enhanced the assessee’s assessment by disallowing certain foreign remittances under Section 40(a)(i) because tax was not deducted at source.
The Settlement: The assessee had a separate appeal pending regarding the Section 201/201(1A) order (which treated them as an ‘assessee in default’ for the same transaction). The assessee opted to settle this Section 201 appeal under the Vivad Se Vishwas Scheme, 2020.
Decision
CBDT Circular Reliance: The Tribunal relied on CBDT Circular No. 12 of 2024 (specifically relating to the VsV Scheme FAQs).
Consequential Relief: The scheme/circular clarifies that once the primary liability (TDS default) is settled under the VsV Scheme, the consequential disallowance under Section 40(a)(i) associated with that default is essentially cured.
Ruling: Since the assessee settled the dispute regarding the failure to deduct tax, the disallowance under Section 40(a)(i) (which exists to penalize non-deduction) no longer survives. The addition was deleted.
Key Takeaways
Settlement Cures Disallowance: Opting for the Vivad Se Vishwas Scheme to settle a TDS default (Section 201) has a domino effect; it automatically resolves the related disallowance of expenditure under Section 40(a)(i) for the same transaction.
IN THE ITAT MUMBAI BENCH ‘J’
Netflix Entertainment Services india LLP
v.
Deputy Commissioner of Income-tax*
Amit Shukla, Judicial Member
and Ms. RENU JAUHRI, Accountant Member
and Ms. RENU JAUHRI, Accountant Member
IT Appeal No. 6857 (Mum) of 2024
[Assessment year 2021-22]
[Assessment year 2021-22]
OCTOBER 17, 2025
Porus Kaka, Divesh Chawla, Harsh Shah, Advs., Ms. Karishma R. Phatarphekar and Rohan Vesuna, CAs for the Appellant. Pankaj Kumar, CIT DR for the Respondent.
ORDER
Amit Shukla, Judicial Member.- The aforesaid appeal has been filed by the assessee against final assessment order dated 25/10/2024 passed u/s.143(3) r.w.s. 144C (13) for the A.Y.2021-22 in pursuance of direction given by the DRP u/s. 144C(5) dated 28/09/2024.
2. In various grounds of appeal the assessee has challenged firstly-
| • | Transfer pricing adjustment of Rs.4,44,95,50,224/- in relation to the payment of distribution fee (ground No.5-13) |
| • | Assailing the enhancement by the DRP u/s.40(a)(i) (Ground no.14-16); |
| • | Assailing final assessment order has been without jurisdiction (Ground No.2-4); |
| • | Error in computation of assessed income (Ground No.17); |
| • | Levy of excess interest u/s.234A, 234B and 234D of the Act (Ground No.18) |
| • | Initiation of penalty proceedings u/s.270A of the Act (Ground No.19) |
The brief background of Transfer Pricing Adjustment
3. The brief facts are that Netflix Inc. (“Netflix US”), incorporated in the United States in 1997, is a globally renowned subscription-based entertainment enterprise that pioneered the digital streaming model enabling subscribers across the world to view movies, documentaries, and television series on any internet-enabled device. It operates on a subscription model whereby users, through the Netflix application or website, gain access to a curated library of video-on-demand (VOD) content. Netflix US has, over time, invested colossal sums in developing and maintaining the content library, service architecture, proprietary streaming technology, infrastructure, trademarks, and other intellectual property assets which form the backbone of its global operations.
4. Since inception, Netflix US has only ever granted to its subscribers a limited ability to view the content hosted on its platform. At no point has it transferred or granted any intellectual property rights, ownership, or exploitation rights in any content, technology, or know-how forming part of the Netflix Service to the subscribers. The subscribers merely receive a limited, personal, non-exclusive right of access.
5. For non-US territories, up to 31 December 2020, Netflix US granted a licence to its associated enterprise, Netflix International B.V. (NIBV), a company incorporated in the Netherlands. Under this licence, NIBV was authorised to use, exhibit, distribute, sub-distribute and premiere Netflix content and marketing intangibles, and was vested with rights to copy, reproduce, publicly perform and broadcast the Netflix Service in all media outside the United States.
6. In India, Netflix Entertainment Services India LLP (hereinafter “Netflix India” or “the Assessee”) was incorporated on 12 April 2017 to facilitate the Indian presence of the group. Its primary business is to distribute access to the global Netflix Service comprising a video-on-demand streaming subscription that enables subscribers to watch global content available on Netflix’s digital platform.
7. On 5 September 2017, Netflix International B.V. entered into a Distribution Agreement with Netflix India (given at page 102 paper book). Under the said agreement, NIBV appointed Netflix India as its non-exclusive distributor of access to the Netflix Service within India. Thereafter, effective 1 January 2021, Netflix US directly entered into an analogous agreement with Netflix India, appointing it as a non-exclusive distributor of the access to Netflix Service to end users in India. This agreement was further revised with effect from 1 March 2021 (paper book pages 122 and 142). Since all these agreements were materially identical in their terms, reference is generally made to the original distribution agreement with NIBV.
8. Under these agreements, Netflix India was authorised only to distribute access to the “Netflix Service”, which is defined as “access to a global video-on-demand service that is delivered through streaming over the internet for personal use.” Netflix India’s duties included soliciting and promoting distribution of the Netflix Service to customers in India, enabling them to purchase subscriptions, invoicing such subscribers, and collecting the subscription fees. It was, therefore, the primary distributor of access to Netflix Service in India. Prior to Netflix India’s incorporation, the distribution of access to the Netflix Service in India was undertaken directly by NIBV.
9. Netflix India was also required to enter into the “Terms of Use” directly with the Indian subscribers (PB page 394). These Terms of Use made it explicit that Netflix India did not acquire any intellectual property rights, nor was it entitled to any proprietary rights in respect of the service architecture, content, trademarks, or technology comprising the Netflix Service. There was no transfer of any know-how, model, invention, or other intellectual property, patented or otherwise, to Netflix India.
10. Based on these contractual terms, Netflix India’s sole and entire role was confined to the distribution of access to the Netflix Service. It neither performed any other function nor owned any assets or assumed any risks beyond those typically borne by a limited-risk distributor. It was not involved in content creation, technology development, or global decision-making. All significant functions were performed and controlled by the Associated Enterprises, while Netflix India merely facilitated distribution of access to the platform within India. Furthermore, responsibility for lastmile connectivity that is, actual internet access enabling streaming rested exclusively with the subscriber, who independently chose his or her Internet Service Provider (ISP), speed, and bandwidth, and bore the associated costs.
11. NIBV, and subsequently Netflix US, were directly responsible for providing access to the Netflix Service to end users. They were also solely responsible for establishing, managing, and maintaining all infrastructure and acquiring content necessary for the streaming service. Netflix India’s role did not extend to these areas. Indeed, the group entities (NIBV and Netflix US) have paid the applicable equalisation levy on the distribution fees received from Netflix India.
12. Under the agreements, Netflix India was required to remit a distribution fee to NIBV/Netflix US. The distribution fee was computed as the total subscription revenue collected from Indian customers, net of local costs incurred, plus a fixed return on Indian sales. Thus, Netflix India was guaranteed a fixed profit margin after its operating costs a structure consistent with its characterisation as a limited-risk distributor. Its profit margin for the year under consideration was 1.36% on sales, which was fully cost-insulated.
13. Netflix India’s Functions, Assets, and Risk (FAR) profile substantiates this characterisation:
(a) Functions:
Netflix India performed limited functions including: entering into Terms of Use with customers; issuing invoices and collecting subscriptions; hiring approximately 60 employees (39 in sales and marketing, one in relation to transfer of OCAs, and 20 in administrative functions); marketing and promoting subscriptions as per global strategies; liaising with telecom operators and ISPs for business development; arranging, though not providing, customer service; and procuring and transferring Open Connect Appliances (OCAs) specialised cache devices used to reduce network congestion for ISPs. It also undertook local legal and compliance work such as securing licences and approvals.
(b) Assets:
Netflix India neither owned nor developed any intangible assets. All intellectual property rights, content, service architecture, and trademarks were owned and controlled by Netflix US and NIBV. Tangible assets comprised routine office equipment, computers, fixtures, and leasehold improvements, totaling Rs.75.48 crores as of 31 March 2021. Of this, less than 10% pertained to OCAs and network gear. In contrast, as on 31 December 2020, Netflix US’s total assets amounted to USD 3,928 crore (approximately Rs.3,92,80,359 thousand), of which content assets alone were USD 2,53,83,950 thousand, clearly indicating that the value-creating assets resided entirely outside India.
(c) Risks:
Netflix India was risk-insulated. Its entire cost base was reimbursed by the Associated Enterprises, along with a fixed mark-up on sales. It bore only limited operational risks such as minor market or regulatory risks incidental to its distribution activity, while all critical entrepreneurial, service liability, and investment risks were assumed by NIBV/Netflix US. Thus, Netflix India had no exposure to losses and functioned as a cost-plus, limited-risk distributor.
13A. On a comparative scale, the overall asset base of Netflix India was negligible vis-a-vis that of Netflix US barely 1 crore USD compared to over 3,928 crore USD globally, making the Indian entity almost 4,000 times smaller. The revenue from India constituted less than 1% of Netflix’s global turnover. Likewise, its human capital was insignificant, with only 64 employees compared to approximately 9,400 globally (including 5,258 in the US), representing merely 0.68% of the global workforce and 1.22% of the US staff mostly engaged in marketing and administration, not content or technology functions.
14. In sum, Netflix India performed the typical and routine functions of a distributor without ownership of any intangible or participation in development, enhancement, maintenance, protection, or exploitation (“DEMPE”) of intellectual property. Its profit and cost structure reflected complete financial insulation.
15. For benchmarking its international transaction of payment of distribution fee to its Associated Enterprises, Netflix India applied the Transactional Net Margin Method (TNMM) as the Most Appropriate Method (MAM) under section 92C of the Income Tax Act, using Operating Profit/Operating Revenue (OP/OR) as the Profit Level Indicator (PLI).
16. The assessee undertook an exhaustive multi-step search process to identify appropriate comparables:
| • | Step I: Search for entities performing activities similar to Netflix India (distribution of audio/video streaming services) -rejected due to non-availability of suitable comparables. |
| • | Step II: Search for media distributors – rejected since none passed the quantitative and qualitative filters. |
| • | Step III: Google-based search for similar entities – rejected for same reasons. |
| • | Step IV: Selection of entities engaged in distribution of software and related products – accepted as they were functionally analogous to Netflix India’s distribution role. |
17. Based on 17 such software-distribution comparables, the operating margin ranged between 1.88% to 2.23%, with post working-capital adjustment margins between 0.77% and 1.47%. Against this, Netflix India’s margin stood at 1.36%, squarely within the arm’s-length range, thereby establishing that the payment of distribution fee was at arm’s length.
18. The assessee contended that its study fully complied with the statutory framework under Chapter X and that its FAR profile, cost-plus structure, and benchmarking methodology had been correctly and scientifically undertaken. The dispute, however, arose when the Transfer Pricing Officer (TPO) chose to disregard these findings and recharacterised Netflix India’s profile an aspect discussed subsequently.
TPO’s Analysis and Finding
19. The Transfer Pricing Officer (TPO), while examining the international transactions between Netflix India and its Associated Enterprises (NIBV and later Netflix US), departed from the assessee’s declared functional characterization as a “limited-risk distributor of access to Netflix Service.” The TPO rejected this characterization and embarked upon a recharacterization exercise, asserting that Netflix India bore significant entrepreneurial, regulatory, and operational risks, and was in effect not a mere distributor but the principal service provider of the Netflix content and platform in India.
20. In arriving at this conclusion, the TPO relied heavily upon specific clauses of the Distribution Agreement, particularly clauses 4.1(b), (d), (e), (g), (h), (i), (l), and (m) (TPO order, pp. 64-65), contending that these provisions revealed Netflix India’s independent and risk-bearing role. According to the TPO:
| • | Netflix India was providing services to Indian subscribers on its own accountability (clause 4.1(b)); |
| • | It was responsible for making available the Netflix Service to Indian customers by leasing or licensing the services and products from the ultimate or designated owner (clause 4.1(d)); |
| • | It had the obligation to promote and market the Netflix Service in India (clause 4.1(h)); |
| • | It could, at its own discretion, issue purchase gift subscriptions, offer discounts and decide pricing, thereby exercising control over the pricing mechanism (clause 4.1(e)); |
| • | It entered into agreements with Indian subscribers on its own account, without the authority to bind or obligate the AEs (clause 4.1(g)); |
| • | It provided customer support directly to Indian users (clause 4.1(i)); |
| • | It bore legal risk in relation to subscriber agreements; and |
| • | It was required to procure licenses, permissions, and infrastructure necessary for distributing the Netflix Service in India (clause 4.1(m)). |
21. On this basis, the TPO concluded that the assessee was not a limited-risk distributor but an independent entrepreneur performing multiple functions, employing valuable assets, and assuming substantial risks. The TPO observed that Netflix India, through its agreements and operations, controlled the marketing strategy, pricing decisions, customer relations, and local infrastructure (including Open Connect Appliances or OCAs), which were indispensable for the streaming of Netflix content in India.
22. Proceeding further, the TPO held that Netflix India was not merely distributing access to the Netflix platform but was providing the Netflix content and the platform itself to Indian subscribers. He reasoned that Indian customers paid their subscription fees exclusively to Netflix India for the right to view Netflix content; none of the subscription revenue was paid directly to the Associated Enterprises. Thus, in substance, Netflix India was delivering a complete content package to customers for a stipulated period.
23. To substantiate this view, the TPO cited the assessee’s substantial expenditure on streaming, communication, and networking costs, which, according to him, could not have been incurred merely for a low-risk distribution role. He further alleged that the assessee’s self-classification as a distributor was a deliberate structuring device to shield its AEs from royalty-related obligations under Indian tax law.
24. The TPO also concluded that there was a transfer of intellectual property rights in the content to Netflix India. He stated that the provision of the service by Netflix India to Indian customers was identical to the provision of the service by Netflix US to US customers with the only difference being that Netflix US owned the content. As per him, the AEs had granted a licence to Netflix India to make the content available to Indian subscribers. Copies of the content were stored on OCAs owned by Netflix India, which were essential for streaming. Consequently, the TPO asserted that Netflix India had acquired both content and the technological platform on licence from its AEs and was required to pay royalty or licence fees for the same.
25. The TPO therefore rejected the TNMM adopted in the Transfer Pricing Study Report (TPSR). He observed that the assessee’s comparables largely software and hardware distributors were inappropriate because Netflix India was not trading in goods but providing complex, integrated services in the media and entertainment industry. According to him, software distribution entities operated in a B2B environment on credit terms and earned low margins, whereas Netflix India operated in a B2C model directly serving millions of subscribers, involving content licensing, technology integration, and marketing. Hence, the TPO held that the traditional TNMM was unfit for benchmarking such an intricate model.
26. Invoking the flexibility of Rule 10AB of the Income-tax Rules, 1962, the TPO adopted the so-called “Other Method” as the Most Appropriate Method (MAM). He contended that this approach was better suited for complex transactions where direct comparables were unavailable. Citing the OECD Guidelines, he noted that the OTT business rested upon three critical pillars content, technology platform, and marketing & sales. Since Netflix India allegedly obtained content and technology on licence from its AEs, the TPO opined that the arm’s-length price (ALP) should be determined on the basis of royalties payable for these components.
27. For this purpose, the TPO sourced six uncontrolled royalty agreements from the RoyaltyStat database: three concerning licences for content distribution rights and three for technology platform rights. On analysing these agreements, the TPO computed the royalty rate for content rights at 48.75 % of revenue, and for technology platform rights at 8.37 % of revenue, aggregating to 57.12 % of Netflix India’s total revenue. Accordingly, he proposed a massive transfer-pricing adjustment of Rs. 4,44,93,42,724.
28. In further support of this computation, the TPO prepared a notional margin attribution table, allocating percentages to various functions allegedly performed by Netflix India such as content storage, CDN & ISP contracts, invoicing, customer support, marketing, copyright protection, regulatory approvals, and technology maintenance cumulatively attributing 43 % of total margins to the Indian entity. The TPO’s analysis thus portrayed Netflix India as a full-scale, entrepreneurial operator of the Netflix Service in India, performing all value-creating activities locally.
DRP’s Direction
28. The Dispute Resolution Panel (DRP), while dealing with the assessee’s objections, substantially endorsed the TPO’s findings. In its extensive directions (pp. 56-73; 99-107; 150166), the DRP recorded that Netflix India undertook a plethora of functions including (i) securing orders on behalf of its foreign AEs; (ii) entering into user agreements; (iii) promoting the Netflix Service and issuing gift subscriptions; (iv) maintaining digital content stock for distribution to endusers; (v) deciding pricing and discounts; (vi) billing, fund collection and transfer; (vii) providing infrastructure support, including OCAs and ISP arrangements; (viii) offering customer support; (ix) feedback and reporting; (x) distribution of the Netflix Service; (xi) obtaining licences and permissions; (xii) monitoring legal compliance; (xiii) securing regulatory approvals; (xiv) maintaining infrastructure and resources; and (xv) negotiating with Internet Service Providers.
29. The DRP placed particular emphasis on the OCAs and the ISP arrangements, observing that these formed the “backbone” of Netflix’s streaming service in India. According to the Panel, the ownership of OCAs by Netflix India rendered it an “extremely significant contributor” to the group’s OTT/VoD business in India. The DRP opined that by owning such infrastructure, Netflix India had accepted investment risk and was therefore far more than a routine distributor.
30. In conclusion, the DRP affirmed the TPO’s characterisation of Netflix India as a key operational hub within the global Netflix group a significant technological and asset service provider as well as an entrepreneurial distributor. It ruled that the assessee had under-reported its functions, assets, and risks, and that its Transfer Pricing Study Report was therefore inaccurate. Consequently, the Panel upheld the TPO’s adoption of the “Other Method” and confirmed the adjustment proposed on the royalty-rate basis.
31. Proceeding further, the DRP examined the TPO’s reasoning that the traditional transfer-pricing methods failed to capture the complexities of the streaming-media business. It observed that the Transactional Net Margin Method (TNMM) adopted by the assessee in its Transfer Pricing Study Report (TPSR) was “unscientific, misdirected, and incompatible with the business model actually carried out in India.” The Panel reasoned that the Assessee’s benchmarking on the basis of distributors of computer software and related products was fundamentally misplaced, since Netflix India was neither a trader of tangible goods nor engaged in the sale of software licences. Instead, it was a participant in a hybrid model of content distribution coupled with technology provision a model more akin to media and entertainment service providers.
32. The DRP emphasised that, according to the TPO, the comparables used by the assessee were unsuited to the Indian entity’s operations. Out of the 17 comparables, 14 were hardware traders, 2 were equipment traders, and only one was a software-solutions company. None of these entities, it found, passed the filters which Netflix India itself had set up. The DRP further noted that several of the comparables, such as Best IT World (India) Pvt. Ltd. and Computronix Infotech Pvt. Ltd., were persistent loss-makers and lacked reliable three-year financial data. Eight comparables did not even possess current-year data, and none satisfied the turnover filter.
33. The DRP further highlighted that Netflix India’s turnover exceeded Rs. 1,500 crores during the relevant year, yet its own study applied a turnover filter of only Rs.1 crore, which, in the Panel’s view, was “incongruous and manipulative.” The Panel also criticised the TPSR for employing the industry classification “Electronics, Software Products, and Database” to entities that bore no resemblance to an OTT or digitalmedia service provider. According to the Panel, these distortions resulted in an artificially depressed median margin below 1%, effectively pre-determining an arm’s-length comfort zone rather than objectively establishing one.
34. The DRP went on to note that the assessee’s adoption of the Operating Profit to Sales ratio (OP/OR) as its Profit-Level Indicator (PLI) was “mechanical and erroneous,” because Netflix India, according to the Panel, undertook complex functions integrally tied to the Netflix Service. In such a scenario, a PLI based purely on operating margins could not, in its view, capture the true value of the intertwined activities performed by Netflix India. The DRP observed that the assessee’s role was functionally inseparable from that of the non-resident entities, and therefore, the assessee’s standalone profitability could not be benchmarked in isolation.
35. The DRP then considered the asset-intensity and marketing-intensity adjustments filed by the assessee during the proceedings. It rejected both sets of workings, holding that the asset adjustments had “no factual or economic basis,” since the comparables did not employ any infrastructure even remotely comparable to that used by Netflix India. It also held that the marketing-intensity adjustments were defective because the advertisement, marketing, and promotion (AMP) expenditure of Netflix India had not been properly considered, while the comparables being hardware and software distributors incurred little to no AMP expenditure. Consequently, it held that the TNMM failed in both its quantitative and qualitative dimensions.
36. Having thus discredited the assessee’s benchmarking analysis, the DRP upheld the “Other Method” adopted by the TPO as the Most Appropriate Method under Rule 10AB, observing that this method provided greater flexibility for determining prices in complex, multi-component transactions where no external comparables were available. The Panel recorded that the OECD Guidelines, too, recognised the “Other Method” as an accepted approach in such situations, particularly in digital-platform economies where intangibles and technological integration dominate the value chain.
37. The DRP further reiterated that the OTT streaming business, by its very nature, was built upon three essential pillars content creation and distribution, technology platform infrastructure, and marketing & sales interface and that Netflix India had participated in all three. It held that the assessee had acquired the right to use both the content library and the technological platform under a distribution licence, and therefore, the royalty model adopted by the TPO reflected a more realistic approximation of arm’s-length pricing than the TNMM.
38. In reinforcing its conclusion, the DRP stated that Netflix India’s role was not comparable to a traditional distributor but rather akin to a de-facto supplier of content in the Indian market. It asserted that Netflix India’s operations exposed it to the “full spectrum of contractual, regulatory, and singleseller risks” typical of a primary operator, not of a limited-risk distributor. It observed that Netflix India was actively involved in three critical functions (i) content procurement and provision, (ii) technological platform management and streaming infrastructure, and (iii) customer-facing activities such as billing, marketing, and customer service each of which entailed independent risk exposure and required commensurate returns.
39. The DRP also emphasised that Netflix India had been “granted a licence under the Netflix Service Distribution Agreement (NSDA) to distribute Netflix Service, which effectively constitutes a distribution of media and entertainment content and not merely the provision of access.” The Panel noted that both Netflix India and NIBV appeared to possess similar rights to provide Netflix Service within their respective territories, which, in its view, belied the assessee’s claim of being a mere facilitator. It therefore characterised Netflix India as a full-scale entrepreneurial entity possessing substantial assets, contractual obligations, marketing resources, customer-service infrastructure, and hosting capabilities, thereby rejecting the notion of a routine distributor.
40. Ultimately, the DRP upheld the TPO’s determination that 57.12% of Netflix India’s total revenue represented the arm’s-length price for the royalty or licence fee payable to its Associated Enterprises for content and technology. In doing so, the Panel also endorsed the TPO’s detailed allocation table that attributed functional margins aggregating to 43% for the Indian entity, which, according to it, corroborated the overall adjustment figure.
41. Without prejudice to the above, the DRP proceeded to suggest an alternative ad-hoc benchmarking purportedly to “corroborate” the reasonableness of the ALP determined by the TPO. In this alternative model, the Panel sought to attribute approximate margins to various functional clusters such as content distribution, infrastructure, marketing, legal and regulatory support, customer management, and technological operations, without relying upon any external comparable data. However, the ad-hoc approach was not based upon any of the prescribed methods under Rule 10B or Rule 10AB, nor did it contain any quantitative analysis, filters, or empirical data.
42. In essence, both the TPO and DRP’s approach rested upon the foundational premise that Netflix India was not a distributor of access but a provider of content and technology, operating as an entrepreneurial entity bearing substantial investment and operational risks. The arm’s-length computation was therefore undertaken not on a transactional profit basis but by imputing a hypothetical royalty percentage derived from non-comparable third-party licensing agreements.
Issues Involved
43. The present controversy arises out of a fundamental divergence between the assessee’s declared characterization of its role and the Revenue’s re-characterization of the same. At the heart of the matter lies the question whether Netflix Entertainment Services India LLP functions merely as a limited-risk distributor of access to the Netflix Service, or whether, as alleged by the Transfer Pricing Officer and affirmed by the Dispute Resolution Panel, it operates as an entrepreneurial content-and-technology service provider, bearing significant risks and entitled or liable to commensurate remuneration under the arm’s-length principle.
44. The first and primary issue therefore concerns the correct functional characterization of the Indian entity. The assessee asserts that its operations are confined to the solicitation and promotion of subscriptions to the global Netflix Service, the execution of Terms of Use with Indian subscribers, and the collection of subscription revenues on behalf of its Associated Enterprises (AEs). All other valuecreating functions content creation, curation, technological development, infrastructure ownership, and strategic decision-making rest entirely with the AEs abroad. Conversely, the Revenue contends that Netflix India’s agreements, infrastructure ownership, and conduct indicate that it performs a bouquet of vital functions ranging from content dissemination and customer management to pricing, promotion, and network facilitation thus constituting a full-fledged operator in the Indian market.
45. Flowing from this is the second issue, namely, whether Netflix India acquires any rights, title, or licence in the intellectual property or technology constituting the Netflix Service. The TPO has alleged that the Indian entity has obtained on licence both the content and the technological platform, and hence the payments made to its AEs partake of the character of royalty within the meaning of section 9(1)(vi) of the Act and Article 12 of the relevant tax treaties. The assessee, on the other hand, asserts that it acquires no such rights; that the content and technology remain exclusively owned and controlled by the AEs; and that it merely facilitates access to subscribers without any right to copy, reproduce, distribute, or modify any element of the service. The resolution of this issue determines whether the underlying transaction is one of distribution of access or one of exploitation of intellectual property.
46. The third issue pertains to the treatment and functional significance of the Open Connect Appliances (OCAs). The Revenue has viewed the OCAs servers installed with Internet Service Providers to enhance streaming efficiency as substantive technological assets owned by Netflix India, conferring upon it a pivotal infrastructural role in the group’s global operations and exposing it to investment risk. The assessee disputes this, asserting that OCAs are merely cache devices akin to logistical tools for temporary content storage, devoid of processing capability or customer data, and that they serve only to assist ISPs in bandwidth management. Whether these appliances are to be viewed as core technological assets or routine logistical aids bears directly on the FAR profile of the Indian entity.
47. The fourth issue concerns the validity of the benchmarking methodology employed. The assessee benchmarked its international transaction of payment of distribution fees using the Transactional Net Margin Method (TNMM), identifying functionally comparable software and product distributors. The TPO, however, rejected this method as inapplicable, invoking instead the “Other Method” under Rule 10AB based on notional royalty rates sourced from third-party agreements in the RoyaltyStat database. The DRP endorsed this approach and even proposed an ad-hoc corroborative allocation of margins across various functional clusters. The question thus arises: whether the TPO and DRP were justified in discarding the TNMM universally regarded as the most pragmatic for routine distribution functions in favour of an untested and empirically unsupported royalty based construct.
48. The fifth issue relates to the selection and rejection of comparables. The assessee’s comparables consisted of seventeen entities engaged in software or related product distribution, yielding an average margin within the arm’s-length range. The TPO and DRP dismissed these on multiple grounds functional dissimilarity, inadequate turnover filter, data unreliability, and alleged cherry-picking. Whether such rejection was based on objective criteria or on conjecture and whether, in the absence of industry-specific comparables, the use of software distributors was a legitimate proxy, are matters that require judicial determination in light of precedents such as Turner International India (P.) Ltd. and Star Den Media Services Pvt. Ltd.
49. The sixth issue arises from the re-characterization of the transaction itself. The TPO has effectively recast the distribution arrangement into a composite licence transaction involving both content and technology. The propriety of such re-characterization especially when the contractual documents unequivocally describe the assessee as a distributor of access must be tested against settled jurisprudence that forbids tax authorities from disregarding genuine contracts unless they are proven to be sham or colourable. This raises the larger question whether the Revenue can, by mere inference, rewrite the legal relationship between the parties and substitute its own economic characterization.
50. The seventh issue pertains to the quantum and reasonableness of the adjustment. The TPO’s computation of 57.12 per cent of total revenue as arm’s-length royalty, culminating in an adjustment exceeding Rs. 444 crores, is founded upon third-party royalty agreements bearing no functional comparability to Netflix India’s operations. The correctness of extrapolating royalty rates for content-library licences or technology platforms none of which the assessee owns or exploits to a distribution-of-access model, lies at the heart of this controversy.
51. The eighth issue concerns the applicability of the Supreme Court’s ruling in Engineering Analysis Centre of Excellence (P.) Ltd. v. CIT (SC), which clarified that payments by distributors for non-exclusive, non-transferable licences do not constitute royalty. The assessee contends that its factual situation is even narrower than that considered by the Apex Court, since it neither receives nor transfers any copyright or licence but merely facilitates access. The Revenue, however, seeks to distinguish that decision on the ground that the streaming service model is technologically and commercially distinct. The question whether the ratio of Engineering Analysis governs the present facts is thus directly in issue.
52. The ninth issue concerns the conceptual soundness and evidentiary foundation of the DRP’s ad-hoc alternative benchmarking. By attributing arbitrary percentages to various functional heads without any supporting empirical data or adherence to Rule 10B, the Panel effectively introduced a non-statutory mechanism of profit allocation. The legal sustainability of such an approach, which bypasses both statutory methods and economic comparability, constitutes an important question of principle.
53. Finally, the tenth and overarching issue is whether, on the totality of the circumstances, the Revenue was justified in disregarding the assessee’s contemporaneous documentation, prepared in accordance with Chapter X, and substituting it with an uncorroborated hypothetical model divorced from market realities. This encompasses the broader question of administrative overreach in transfer-pricing re- characterization and the limits of the TPO’s discretion under section 92C read with Rules 10B and 10AB.
Arguments on behalf of the Assessee
54. At the very outset, learned Sr.counsel for the assessee Shri Porus Kaka painstakingly drew our attention to the paperbook, TPSR and submissions made before the TPO and DRP and elaborated upon the assessee’s business model, the role of Netflix US and Netflix International B.V. (NIBV), the functions performed, the assets employed, and the risks assumed by each entity. He contended that the entire factual and legal framework was misconstrued by both the TPO and the DRP, who, by selective reading of contractual clauses and unfounded presumptions, had re-characterised a simple distribution arrangement into a complex licence transaction of content and technology.
55. He emphasised that Netflix India is merely a distributor of access to the Netflix Service, functioning as a limited-risk entity with routine marketing and administrative responsibilities. It neither owns nor controls any intellectual property, nor does it acquire or transfer any copyright in the content, technology, or trademarks forming part of the Netflix Service. The “Netflix Service” as defined in the Distribution Agreement is nothing more than access to a global video-on-demand platform for personal streaming, and the assessee’s limited role is confined to distributing that access in India.
56. Learned Sr. Counsel pointed out that the Terms of Use entered with Indian subscribers unequivocally clarify that users are granted only a limited, non-exclusive, non-transferable licence to view content, and that all ownership and intellectual property rights remain vested exclusively with Netflix US or NIBV. Netflix India neither licenses nor sublicenses any content; it merely facilitates the subscription interface and billing process. Accordingly, there can be no inference of transfer of any copyright or technology rights to the assessee or to the subscribers.
57. Reliance was placed upon the authoritative judgment of the Hon’ble Supreme Court in Engineering Analysis Centre of Excellence (P.) Ltd. (supra), wherein it was held that payments made by distributors for the resale of software without the transfer of any copyright do not constitute “royalty.” It was submitted that Netflix India’s role is even narrower, since it neither receives nor resells any software but merely facilitates access to a service owned, hosted, and controlled by its Associated Enterprises. The entire technology stack operates from servers located outside India, primarily on Amazon Web Services (AWS) infrastructure, over which Netflix India has no ownership or operational control.
58. The assessee thus maintained that the TPO’s finding of an implied licence for content or technology was factually unsustainable and legally untenable. The TPO’s conclusion that Netflix India “obtained both content and platform on licence” was contradicted by the very preamble and appointment clauses of the Distribution Agreement (PB pp. 102-112), which describe Netflix India as a non-exclusive distributor of access to the Netflix Service and not as a licensee of any intellectual property.
59. On the aspect of Open Connect Appliances (OCAs), Shri Porus Kaka submitted that the TPO and DRP had gravely erred in treating these as critical technological assets or evidence of entrepreneurial investment. The OCAs, ld.counsel explained, are merely cache servers essentially storage devices placed at ISP nodes to locally store frequently streamed content and reduce network congestion during peak hours. They perform no data processing, contain no customer data, and neither modify nor reproduce any content. All processing, algorithmic recommendation, encryption, and playback functions occur through software hosted and operated by Netflix US outside India. The OCAs therefore serve purely as logistical enablers akin to a distributor’s warehouse; ownership of these devices does not translate into ownership or control over the underlying content or technology.
60. He further submitted that the TPO’s allegation of pricing autonomy and marketing discretion was misconceived. Although the Indian entity could offer minor discounts or gift subscriptions to customers as part of localized marketing campaigns, all such initiatives were executed within the strict parameters prescribed by the parent entity’s global marketing policies and under advance budgetary approvals. The subscription fee itself was determined centrally by the Netflix group’s global pricing algorithm based on uniform tier structures; the Indian entity had no power to unilaterally change or fix the price.
61. It was argued that the re-characterization of the assessee’s functions and risks was perverse, as it ignored the commercial and contractual realities. The group’s organizational model clearly segregated responsibility for content creation, technological innovation, and platform management with Netflix US, while the distribution and customer-facing functions were decentralised to country-level entities such as Netflix India on a limited-risk basis. Every rupee of cost incurred by Netflix India whether marketing, administrative, or infrastructural was reimbursed by the AEs, together with a fixed return on sales (ROS) of 1.36 %, thereby fully insulating it from business risk.
62. Learned counsel next addressed the benchmarking methodology, defending the use of the Transactional Net Margin Method (TNMM) as the Most Appropriate Method (MAM). It was pointed out that no direct comparables exist for the distribution of digital streaming services; therefore, distributors of software and related products whose business model likewise involves reselling access rights to intangible products without ownership were selected as functionally analogous. The assessee undertook a detailed multistep search across AceTP and Capitaline databases and identified seventeen comparables whose average margins fell between 1.88 % and 2.23 % (post-working-capital adjusted range: 0.77 %-1.47 %). Against this, the assessee’s margin of 1.36 % squarely fell within the arm’s-length band.
63. It was vehemently contended that the TPO and DRP erred in rejecting the TNMM on unfounded generalities. Both authorities disregarded settled judicial precedents wherein software distributors have been accepted as valid comparables for benchmarking distribution activities in the broadcasting and entertainment sector. Reference was made to a catena of decisions including Turner International India (P.) Ltd. v. Asstt. CIT (Delhi – Trib.), Star Den Media Services (P.) Ltd. v. ACIT (Mumbai – Trib.), Sony Pictures Networks India (P.) Ltd. v. Asstt. CIT (Mumbai – Trib.) and MSM Discovery (P.) Ltd. v. Assistant Commissioner of Income-tax (Mumbai – Trib.). The principle affirmed therein that when no direct comparables exist, distributors of software products may serve as functional analogues for media-content distributors was binding, yet ignored.
64. Shri Porus Kaka also criticised the TPO’s adoption of the “Other Method” under Rule 10AB, founded upon royalty agreements extracted from the RoyaltyStat database. He submitted that none of the six agreements relied upon by the TPO were comparable either in nature or substance: they involved transfers of actual content libraries, film rights, or technology platform licences transactions wholly absent in Netflix India’s factual matrix. Several agreements were unsigned, outdated, incomplete, or non-contemporaneous, lacking crucial details such as territorial scope, exclusivity, or duration. Moreover, the TPO offered no explanation as to why royalty agreements for unrelated sectors could provide a reliable benchmark for a distribution-of-access model. He thus argued that the TPO’s computation of 57.12 % of revenue as arm’s-length royalty was entirely arbitrary. There was no evidence of any royalty payable by Netflix India; the entire model was based on a cost-plus distribution return. The so-called “functional margin attribution table” devised by the TPO was a notional construct without legal or empirical basis, assigning percentages to random activities and thereby fabricating profitability where none existed.
65. As to the DRP’s ad-hoc corroborative approach, it was urged that the Panel’s exercise of attributing margins to assorted functional clusters was contrary to Rule 10B, Rule 10AB, and every canon of transfer-pricing analysis. No comparable data, quantitative filters, or economic reasoning supported the allocation. The DRP, it was contended, had effectively substituted a statutory method with a fictional arithmetic of convenience.
66. Finally, learned counsel submitted that the entire edifice of the Revenue’s case rested upon an impermissible re-writing of the parties’ contracts. The agreements were genuine, contemporaneous, and consistently acted upon; they could not be disregarded merely because the tax authority preferred a different economic interpretation. In the absence of any allegation of sham or collusion, the Revenue had no jurisdiction to re-characterise the transaction beyond its contractual contours.
67. On these premises, it was earnestly contended that the TPO’s and DRP’s conclusions were unsustainable both in law and on facts, that the assessee’s TNMM analysis stood uncontroverted, and that no adjustment under section 92CA was warranted.
D.R’s Arguments
68. Per contra, the learned Departmental Representative (DR) stoutly defended the findings and conclusions recorded by the Transfer Pricing Officer and subsequently endorsed by the Dispute Resolution Panel. It was submitted that the assessee’s self-characterisation as a “limited-risk distributor” was a strategic understatement, artfully crafted to minimize its Indian tax exposure and to shield its Associated Enterprises from the incidence of royalty taxation. The DR argued that, when one examines the true substance of the arrangement disregarding the formal nomenclature of the agreements it becomes evident that Netflix India is the operative face, arm, and engine of the global Netflix enterprise in India, performing the core revenue-generating and customer-facing functions essential to the group’s streaming business.
69. The DR emphasised that the clauses of the Distribution Agreement, particularly clauses 4.1(b), (d), (e), (g), (h), (i), and (m), clearly manifest that the assessee operates independently and bears entrepreneurial risks. The agreement mandates Netflix India to provide the service to Indian subscribers on its own accountability, to decide pricing, to offer discounts and promotional packages, to enter into agreements with subscribers in its own name, to provide customer support, and to procure the requisite infrastructure and licences for distribution in India. The cumulative effect of these clauses, the DR contended, is that Netflix India acts not as a mere distributor but as the primary supplier of the Netflix Service in the Indian market.
70. It was further contended that the flow of consideration itself substantiates this characterization. The Indian customers, it was pointed out, pay subscription fees directly to Netflix India. No portion of these fees is paid to the Associated Enterprises (Netflix US or NIBV) by the end users, nor does Netflix India collect any separate or earmarked payment on behalf of the AEs for content or technology. Thus, according to the DR, Netflix India effectively commercialises the content and platform in India, earning revenue entirely from local subscribers, while making internal payments to its AEs that are more appropriately characterised as royalty for the use of content and technology rather than distribution fees.
71. The learned DR submitted that Netflix India’s conduct in the market also belies its claim of limited functionality. The assessee is responsible for advertising, local market promotion, and brand development through massive online and offline campaigns tailored to Indian consumers. It undertakes public relations, social media engagement, and collaborations with telecom and consumer electronics companies for bundled subscriptions, all of which, according to the DR, represent entrepreneurial marketing initiatives rather than routine promotional activity. It was argued that such extensive marketing and customer acquisition efforts demonstrate that Netflix India functions as the economic entrepreneur of the Netflix brand in India.
72. Attention was drawn to the ownership and deployment of the Open Connect Appliances (OCAs). The ld.DR asserted that these are not minor logistical tools but core technological assets forming the backbone of Netflix’s streaming architecture. The OCAs store and deliver content to Indian subscribers, forming an integral part of the content delivery network (CDN). Without these OCAs, Netflix’s global content library could not be streamed efficiently within India. The DR argued that, by owning and maintaining these devices, Netflix India assumes not only investment risk but also technological and operational risks risks typically borne by an entrepreneur rather than a limited-risk distributor.
73. The ld.DR further contended that Netflix India is responsible for maintaining relationships and negotiating contracts with Internet Service Providers (ISPs), ensuring seamless content delivery across India’s bandwidth-constrained network. Such negotiations, the DR submitted, require technical expertise, capital investment, and regulatory compliance, which are indicative of infrastructure ownership and operational autonomy inconsistent with the role of a mere distributor.
74. In defending the TPO’s use of the “Other Method” under Rule 10AB, the ld.DR submitted that the streaming-media industry represents a new-age digital business model wherein traditional comparables such as software distributors or B2B resellers cannot capture the economic substance of the underlying transactions. The TNMM, according him, is ill-suited to evaluate a hybrid model involving both content licensing and technology exploitation. The TPO’s selection of the “Other Method,” which allows benchmarking through unrelated royalty agreements, was therefore both permissible and pragmatic.
75. It was emphasised that the OECD Transfer Pricing Guidelines recognise that in certain complex or unique transactions, where no external comparables exist, flexibility under the “Other Method” is warranted. The DR contended that Netflix India’s case typifies such complexity: the assessee has exclusive territorial responsibility for content distribution, maintains critical infrastructure, and controls customerfacing operations. Consequently, adopting royalty agreements representing payments for content and platform licences as comparables for determining arm’s-length pricing was, in the Department’s view, entirely justified.
76. The learned DR also justified the royalty-based ALP computation, observing that the TPO’s identification of six external agreements from the RoyaltyStat database three for content rights and three for technology platforms was a reasonable proxy for the value of intangibles used by Netflix India. It was submitted that the weighted-average royalty rate of 57.12 % of revenue reflected a fair market consideration for the combined use of content and technology, given that these two components are the primary value drivers of the streaming business. The ld.DR asserted that the massive transfer-pricing adjustment of Rs. 444.93 crores merely reflects the economic value extracted by Netflix India from its access to these intangibles, which it commercially deploys in the Indian market.
77. The Id.DR rebutted the assessee’s reliance on the Engineering Analysis judgment, contending that the said decision concerned shrink-wrapped software licences for installation on user devices a fundamentally different paradigm. The Netflix model, it was argued, involves real-time streaming and dynamic content consumption, dependent on continuous access to a global content library and technological infrastructure. Such continuous use, it was urged, constitutes use of copyright and technology, not mere sale of copyrighted articles. Hence, the ratio of Engineering Analysis was inapplicable.
78. With respect to comparables, the ld.DR supported the rejection of the seventeen companies adopted in the assessee’s TPSR, arguing that all were functionally incomparable. The comparables dealt in physical goods or software products, operated under B2B models, and earned wafer-thin margins based on trading economics. Netflix India, on the other hand, functions under a B2C subscription model delivering digital content directly to millions of consumers, thereby operating in a completely different market dynamic. The use of software traders as comparables, according to the DR, was an artificial construct designed to depress the arm’s-length margin.
79. He also defended the DRP’s alternative ad-hoc margin attribution, submitting that it was not arbitrary but merely an illustrative corroboration of the functional significance of Netflix India’s role. The percentage allocations, according to the DR, were drawn from a reasoned assessment of the assessee’s multifaceted responsibilities including content storage, distribution, customer management, and technological upkeep. The attribution merely demonstrated that the Indian entity’s share in the overall profit pool was substantial and not consistent with a low-risk distributor’s profile.
80. Lastly, the learned DR contended that the assessee’s claim of risk insulation was contradicted by its actual conduct. The Indian entity, it was noted, faces regulatory exposure under Indian laws relating to content certification, consumer protection, and taxation; it bears contractual obligations towards customers under the Terms of Use; and it manages local compliance, legal disputes, and data-security issues. Such exposures, it was argued, go far beyond the limited operational risk that a cost-plus distributor would ordinarily bear.
81. Summing up, the learned DR submitted that the TPO’s and DRP’s findings represented a faithful reflection of economic reality, that Netflix India’s FAR profile corresponded to that of a full-fledged entrepreneurial service provider, and that the adjustment proposed was both lawful and justified. It was therefore urged that the addition made under section 92CA be sustained in toto.
Decision
82. We have heard both the parties at length, perused the entire facts and material referred to before us and the observation and the findings given in the TPO’s order as well as DRP’s direction. Here in this case, the entire edifice of the transfer-pricing adjustment rests upon the recharacterisation of the assessee, Netflix Entertainment Services India LLP (“Netflix India”), from a limited-risk distributor to an entrepreneurial provider of content and technology in India. We, therefore, start by delineating the actual contractual framework, the functions, assets and risks (FAR) borne by each entity, and then testing, with granular precision, the validity of the TPO’s and DRP’s contrary findings.
83. The preamble of the Distribution Agreement unambiguously appoints Netflix India as a non-exclusive distributor of access to the Netflix Service in India. The definition clause defines “Netflix Service” as a global video-on-demand streaming service accessible via the internet for personal and non-commercial use. Clause 4.1 and its subclauses delineate operational obligations such as promotion, collection of subscription revenue, local invoicing, and customer support, but significantly, reserve all intellectual property rights (IPRs) including content, technology, software, and trademarks exclusively to Netflix International B.V. (“NIBV”) or Netflix US. The agreement nowhere confers upon the Indian entity any licence to use, reproduce, alter, or sublicense content or technology.
84. The Terms of Use entered with subscribers further reinforce this structure. Customers obtain only a limited, non-exclusive, non-transferable right to access and view content through the Netflix Service; no ownership or copyright in content is ever transferred to them. Netflix India acts merely as the distributor and invoicing entity, facilitating user access on behalf of its AEs.
85. On these very documents, the TPO nevertheless concluded that the assessee was providing “Netflix Service as a whole, including content,” and thus must be regarded as the primary provider of both content and platform in India. We find that such an inference is internally inconsistent because, the very paragraph quoted by the TPO begins by recognising that Netflix India “does not get access to content” yet ends by concluding that it does. Such self-contradiction, as the assessee rightly argued, betrays a perverse appreciation of record and an outcome-driven approach.
86. The TPO relied upon certain clauses particularly 4.1(b), (d), (e), (g), (h), (l) and (m) to assert that Netflix India bears greater risk, decides pricing, enters contracts on its own account, and licenses or procures the service for distribution. A contextual reading refutes this.
| • | Clause 4.1(b) merely allows the Indian entity to provide services to Indian subscribers on its own accountability in respect of billing and customer-service obligations; it does not allocate ownership or economic risk of the service. |
| • | Clause 4.1(d) obliges Netflix India to make the service available, meaning to facilitate access, not to supply or license the content. |
| • | Clause 4.1(e) empowers the assessee to issue gift subscriptions or discounts, yet expressly “within guidelines approved by the AEs”; this denotes tactical flexibility, not pricing strategy. |
| • | Clause 4.1(g) stipulates entry into user agreements “as per its own terms and conditions,” but the preamble clarifies that such Terms of Use are standard global templates, not independently authored. |
| • | Clause 4.1(l) on customer support and Clause 4.1(m) on regulatory approvals relate to routine distributor obligations compliance, billing, grievance redressal and not to any creation or exploitation of IP. |
87. The DRP, building upon these clauses, amplified the mischaracterisation by observing that Netflix India “undertakes all functions except content provision” and “owns critical technological assets (OCAs)” which form the backbone of streaming services. It enumerated fifteen “high-value functions” ranging from content storage and digital stock maintenance to ISP negotiations and regulatory approvals. This sweeping attribution, however, is contradicted by the record and exceeds the contractual remit.
88. The allegation that the assessee maintains digital content stock is demonstrably false. The evidence on record unchallenged before us shows that Open Connect Appliances (OCAs) are cache devices placed at ISP nodes to store temporary copies of data for bandwidth optimisation. They:
| • | contain no customer data, |
| • | perform no algorithmic processing, and |
| • | execute no playback or recommendation logic. |
All such functions are operated by Netflix US via software owned and hosted on AWS servers outside India.
89. The OCAs therefore act as mirror caches, analogous to logistical warehousing in physical distribution. Their local presence facilitates delivery efficiency, not value creation. The TPO’s and DRP’s characterisation of these caches as “critical technological assets implying entrepreneurial risk” proceeds on a misunderstanding: storage for bandwidth efficiency is not technological development or ownership. The logistics analogy advanced by the assessee is apt and remains unrebutted on facts.
90. It is further alleged by the DRP that ownership of OCAs and local ISP arrangements “made Netflix India the backbone of the group’s Indian streaming operations” and hence a riskbearing entrepreneur. This inference confuses operational indispensability with economic entrepreneurship. Every distribution network requires infrastructure at the destination market; such necessity does not, by itself, confer profitentitlement or risk ownership. The contractual and economic risks remain entirely insulated by the cost-plus remuneration structure.
91. On a factual reconstruction of the FAR, we accept the assessee’s summary that:
| • | Netflix India’s functions are limited to promotion, distribution of access, invoicing, local customer support, and regulatory compliance; |
| • | its tangible assets comprise office premises, IT equipment, and OCAs whose function is logistical; |
| • | its intangible assets are nil; |
| • | its risks are limited to routine operational and regulatory exposures, all fully indemnified by the AEs; and |
| • | it earns a Return on Sales (ROS) of 1.36 percent on a fully cost-insulated basis, consistent with a low-risk distributor profile. |
92. Further, the employee profile reinforces this characterisation: the Indian entity’s workforce of about sixty professionals performs marketing support, operations coordination, finance, and compliance. None are engaged in content acquisition, technology design, or platform development. The human-capital matrix thus negates the Revenue’s portrayal of a technology or content entrepreneur.
93. The TPO nevertheless discarded the assessee’s entire FAR and TNMM analysis, choosing instead to adopt the so-called “Other Method” under Rule 10AB. This was premised upon six RoyaltyStat agreements three concerning “distribution of content rights” and three concerning “use of technology platforms” from which the officer extracted notional royalty percentages of 48.75 percent and 8.37 percent, respectively, aggregating to 57.12 percent of revenue, leading to an adjustment of Rs. 4,44,93,42,724.
94. The assessee’s contention which stands unrebutted is that no search methodology, filters, or comparability analysis were ever disclosed. Several of the agreements are outdated, unsigned, or incomplete, and concern licences of films, music catalogues, or software codes, which are economically alien to the assessee’s mere distribution of access. Indeed, by treating such agreements as benchmarks, the TPO assumed the very fact in dispute that the assessee held a content/technology licence which neither exists in contract nor in conduct.
95. The DRP, instead of correcting this deviation, endorsed and expanded it. At page 108 of its directions, it replaced the royalty model with an attribution table assigning arbitrary percentages to multiple functions content storage 5%, CDN + ISP 2%, infrastructure 5%, customer agreements 2%, marketing 5%, technology 5%, etc. and ultimately concluded that 43 percent of total revenue should be attributed to Netflix India. This attribution, unsupported by any external comparables or Rule 10B methodology, is a non-sequitur masquerading as economic analysis.
96. Before the DRP, the assessee furnished working-capital, marketing-intensity, and asset-intensity adjusted margins for the software-product distributor comparables workings that the Panel itself had sought. These demonstrated that the assessee’s margin of 1.36 percent lies well within the interquartile arm’s-length range (-0.48 to +0.32 percent after adjustments). Yet the DRP, without assigning reasons, ignored these very workings and reverted to its ad-hoc allocation. Such disregard of the very data it solicited underscores the arbitrariness of the confirmation.
97. Having thus laid out the facts, we now proceed to the next segment to examine methodologically and jurisprudentially the validity of the TPO’s rejection of TNMM, the inapplicability of the “Other Method”, and the sustainability of the DRP’s ad-hoc attribution grid, while simultaneously evaluating the assessee’s extensive legal submissions supported by precedents and the methodological validity of the Assessing Officer’s and the TPO’s decision to discard the Transactional Net Margin Method (TNMM) and substitute it with the “Other Method” under Rule 10AB.
98. Our analysis proceeds from the jurisprudential premise that a benchmarking method must be grounded in functional comparability, not in speculative reconstruction of transactions that never existed. It is axiomatic that the choice of the Most Appropriate Method (“MAM”) must depend upon the nature of the transaction and the availability of reliable comparables. The assessee’s TNMM, with Operating Profit to Operating Revenue (OP/OR) as the Profit Level Indicator, is demonstrably suited to its role as a limited-risk distributor. The TPO, however, summarily discarded it on the ground that the assessee “is not a trader of goods” and that comparables comprising software distributors were functionally divergent because Netflix India operates in the media and entertainment industry. This reasoning betrays a fundamental misconception. As the coordinate benches have repeatedly emphasised, functional similarity not sectoral label is the touchstone of comparability.
99. The Assessee before us demonstrated, with empirical evidence, that it first conducted an exhaustive search for media-streaming distributors and finding none that met quantitative and qualitative filters, resorted to software and related product distributors, which mirror the same economic essence: distribution of intangible property under limited-risk conditions. The margins of seventeen such comparables, post working-capital and asset-intensity adjustments, ranged from 0.77 to 1.47 percent, within which the assessee’s 1.36 percent fell squarely.
100. The TPO’s rejection of these comparables proceeded on the mechanical assertion that “the assessee is not a trader of goods.” This reasoning ignores that, in transfer-pricing law, intangibles can be distributed without being “traded.” The Supreme Court in Engineering Analysis Centre of Excellence (P.) Ltd. (supra) has held that the distribution of access to software even where downloads occur is not a transfer of copyright. Netflix India’s function, being still narrower (distribution of access without any right of reproduction), stands on even firmer footing.
101. The DRP, while echoing the TPO, added that the comparables were “hardware traders” and “equipment dealers,” that many failed turnover and persistence filters, and that the assessee had cherry-picked to arrive at a median below 1 percent. Yet these observations were sweeping, not supported by any re-computation or alternative set. Crucially, the DRP ignored that it had itself directed the assessee to submit asset-intensity and marketing-intensity-adjusted margins; when those workings vindicated the assessee, the Panel dismissed them as “baseless” without analytical counter-workings.
102. We note that Rule 10B(2) and OECD Guidelines 2.59 accord primacy to functional comparability where product or market comparables are unavailable. A method that reasonably reflects the economic reality of the tested party cannot be rejected merely because the industry label differs. Consistently, this Tribunal in various cases viz., Turner International India (P.) Ltd. (supra), Star Den Media Services Pvt. Ltd. (supra), Sony Pictures Networks India Pvt. Ltd. (supra), and Pr. CIT v. Warnermedia India (P.) Ltd. (Delhi), affirmed that software distributors are valid analogues for media-content distributors in the absence of direct comparables.
103. Conversely, the “Other Method” adopted by the TPO is both factually and legally untenable. Rule 10AB contemplates a residual method applicable only when no recognised method can be reasonably applied. Here, TNMM was demonstrably workable; the TPO invoked the residual clause merely to justify an ex-post royalty mosaic. He selected six RoyaltyStat agreements three for “distribution of content rights” and three for “use of technology platforms” yielding blended royalty rates of 57.12 percent of revenue. This approach presupposes that Netflix India holds licences to content and platform technology, an assumption directly contradicted by the Distribution Agreement and Terms of Use
104. The Assessee’s rejoinder, fortified by documentary evidence, establishes that no licence transaction exists: there is no right to copy, adapt, sub-license, or modify any content or code; the OCAs remain group-owned caches; and all intellectual property is held by Netflix US/NIBV. The royalty comparables thus price an imaginary transaction. Furthermore, the RoyaltyStat agreements relied upon are non-contemporaneous, unsigned, and economically dissimilar some relate to film-library sales and others to franchise or app-deployment rights. None share the tested party’s limitedrisk distribution profile.
105. The DRP’s confirmation of this “Other Method” compounds the error. Instead of testing comparability, it devised an ad-hoc attribution table, assigning arbitrary percentages (5 percent to content storage, 2 percent to CDN contracts, 5 percent to marketing, etc.) and concluding that 43 percent of revenue be attributed to Netflix India. Such allocation has no mooring in Rule 10B/10AB; it lacks external comparables, cost-driver linkage, and risk-return rationale. It effectively manufactures a pricing mechanism out of thin air something the law does not permit.
106. At a conceptual level, we are of the opinion that Rule 10AB is not a licence for arbitrary attribution. The method must still rely on “comparable uncontrolled transactions” or reasonable quantitative adjustments. Neither the TPO nor the DRP has demonstrated even a single third-party agreement that mirrors Netflix India’s role as a distributor of access. Moreover, the OECD Transfer Pricing Guidelines 2022 2.149 caution that “Other Methods” should be invoked sparingly and only where they yield a higher degree of reliability than the established methods. Here, the TPO’s hybrid royalty construct reduces reliability by introducing incomparable property rights and by ignoring actual tested-party data.
107. The Id. Counsel rightly pointed out that the Revenue’s premise that Netflix India “obtains content and technology on licence” is internally inconsistent with the TPO’s own earlier finding that Netflix India “does not get access to content.” Such logical dissonance undermines the integrity of the adjustment.
108. We therefore hold that the TNMM remains the Most Appropriate Method, given the functional profile, availability of data, and jurisprudential acceptance of similar comparables. The RoyaltyStat-based Other Method and the DRP’s ad-hoc corroboration grid stand vitiated by nonapplication of mind and absence of comparability.
109. In the ensuing part, we shall address the specific arguments of both parties the Department’s claim that Netflix India is the “provider” of content and technology, and the Assessee’s exhaustive rebuttal demonstrating why this characterisation is factually misconceived and legally unsustainable.
110. The pivotal question that now engages our judicial scrutiny is, whether the Assessee Netflix Entertainment Services India LLP can be re-characterised as a contentprovider and technology entrepreneur for the Indian market, as asserted by the TPO and endorsed by the DRP, or whether its true role remains that of a limited-risk distributor merely facilitating access to a global service.
111. The TPO’s thesis rested on the assertion that Netflix India “is providing Netflix Service as a whole including content” and “obtains rights to content and technology through Netflix International B.V. („NIBV’) for onward streaming to Indian customers.” The officer relied on selected clauses of the Distribution Agreement particularly clauses 4.1(b), (d), (e), (g), (h), (l) and (m) to allege that the Assessee fixes subscription prices, issues gift subscriptions, contracts with users independently, and assumes legal and regulatory risks. This reasoning collapses upon inspection. The cited clauses, when read in pari materia with the Agreement’s preamble and Article 9 on ownership of intangible property, reveal that Netflix India’s discretion is purely operational, not entrepreneurial. The Assessee’s latitude to issue discounts or handle customer service cannot metamorphose into control over IP or content. Indeed, clause 9.1 explicitly reserves all intellectual-property rights including patents, copyrights, and trademarks to Netflix US/NIBV. The TPO’s inference therefore amounts to reading contractual autonomy into administrative convenience.
112. The DRP magnified this mischaracterisation by declaring that “all functions are carried out by Netflix India except content provision,” thereby imputing to the Assessee even “maintenance of digital content stock,” “content storage through OCAs,” and “technology functions forming the backbone of streaming”. The Panel further claimed that ownership of OCAs by the Indian entity “made it a significant contributor accepting investment risk.” Yet these findings disregard both the record and technical reality.
113. The Assessee has conclusively shown that an Open Connect Appliance (OCA) is a mere mirror cache, storing transitory packets to optimise network bandwidth. It performs no processing, programming, or data analytics; nor does it house subscriber data. Every algorithm contentrecommendation, compression, adaptive streaming is developed, owned, and operated by Netflix US on AWS servers outside India. We find this evidence as uncontroverted. To equate such caching devices with core technological assets is to mistake warehousing for authorship.
114. The TPO’s additional claim that Indian subscribers pay Netflix India for viewing “Netflix content” and not merely for access likewise fails the contractual test. The Terms of Use make clear that the end-user receives only a non-exclusive, non-transferable right to access and view; no part of the subscription constitutes consideration for transfer of copyright. Netflix India, having no copyright itself, could transfer none to others. This proposition is fortified by the Supreme Court’s ruling in Engineering Analysis Centre of Excellence (P.) Ltd. (supra), which held that a distributor of software or digital content who merely enables access does not acquire or convey copyright. The Assessee’s role is even narrower, as it does not even host or deliver the content; the streaming is effected by Netflix US’s global servers.
115. The DEMPE analysis (Development, Enhancement, Maintenance, Protection, Exploitation) further dismantles the Revenue’s case. The record shows:
| (a) | All Development and Enhancement of technology, algorithms, and user interface occur within Netflix US’s engineering teams; |
| (b) | Maintenance of the platform, bug fixes, and feature roll-outs are executed centrally; |
| (c) | Protection of IP, including registration of trademarks and enforcement of copyright, is undertaken by NIBV and Netflix US; |
| (d) | Exploitation through global licensing and monetisation remains wholly offshore. |
In India, no DEMPE function save routine regulatory facilitation is performed.
116. Asset data corroborate this conclusion: Netflix India’s total assets are approximately Rs. 75 crores (« USD 1 million), while Netflix US’s assets exceed USD 3,928 crores about 4,000 times larger. Content assets form the predominant share of the group’s balance sheet; Netflix India holds none. Even employee strength (64 in India versus 9,400 globally) and roles (predominantly marketing, administration, and compliance) demonstrate a purely supportive function, not IP creation.
117. Risk allocation, too, confirms the limited-risk profile. All critical entrepreneurial risks market, investment, service liability, technological obsolescence are borne by the AEs. Netflix India’s costs are reimbursed, and it earns a 1.36 percent Return on Sales, insulating it from losses. Such arrangements, akin to cost-plus contracts, negate any entrepreneurial exposure.
118. The DRP’s contrary finding that ownership of OCAs implies “investment risk” is untenable. The devices, being cost-reimbursed, entail no capital risk. Netflix India neither funds their acquisition nor controls their deployment strategy; they remain assets functionally akin to warehouses supplied for efficiency.
119. The TPO’s invocation of customer-research functions that the Assessee undertook studies of Indian viewer preferences to aid content curationis overstated. Marketfeedback activities form part of ordinary distribution and marketing support. They do not equate to contentdevelopment functions under DEMPE. No evidence exists of any budgetary control, decision-making, or intellectual contribution by the Indian entity toward production or selection of shows.
120. In sum, both authorities below have conflated facilitation with creation, logistics with technology, and compliance with entrepreneurship. Their reasoning is inconsistent with both the contracts and international transfer-pricing doctrine.
121. Consequently, we hold that:
| (a) | Netflix India performs routine distribution and marketing support functions under strict supervision of its AEs; |
| (b) | It owns no valuable intangible assets and undertakes no DEMPE functions; |
| (c) | Its risks are limited and cost-insulated; |
| (d) | Accordingly, its profitability benchmark under TNMM reflects an arm’s-length outcome. |
125. Having thus resolved the factual and functional disputes, we shall, now turn to the legal adjudication of the benchmarking methodology, the rejection of TNMM vis-a-vis Rule 10B, and the ultimate quantification of the arm’s-length margin. We have traversed the labyrinth of facts, contractual architecture, and technical operations, we now approach the legal heart of the dispute whether the TPO and DRP were justified in disregarding the Transactional Net Margin Method (TNMM) adopted by the assessee and replacing it with an artificial “Other Method” and ad-hoc attribution of profit percentages, and whether such an approach satisfies the discipline of Chapter X of the Act.
126. At the threshold, it is imperative to reiterate that Section 92C read with Rule 10B provides an exhaustive framework for determining the arm’s-length price (ALP). These rules confer no discretion upon the tax authorities to devise novel or hybrid methods divorced from recognised economic or accounting standards. The residuary “Other Method” under Rule 10AB may only be invoked where none of the prescribed methods can reasonably apply and where a demonstrably more reliable means is available. That safeguard is the legal bulwark against arbitrary attribution of income.
127. In the instant case, the TPO’s entire approach proceeds not from economic comparability but from functional mis-characterisation. Having wrongly presumed that Netflix India is a licensor or owner of content and technology, he lifted royalty rates from unrelated third-party licensing agreements concerning film catalogues and software platforms transactions wholly alien to the assessee’s actual profile and constructed from them a blended royalty of 57.12% of revenue. This percentage was then treated as the supposed ALP of distribution fees, yielding a transfer-pricing adjustment of ? 4,44,93,42,724.
128. Such methodology lacks any statutory anchor. The so-called “Other Method” used by the TPO does not demonstrate how those agreements represent comparable uncontrolled transactions as defined in Rule 10B(2). No filters, no functional similarity, no geographic or market comparability were established. Indeed, as the Assessee correctly pointed out, several agreements were non-contemporaneous, unsigned, incomplete, and related to outright sales or licensing of IP, whereas Netflix India has no IP to sell or license. The TPO’s method is thus a textbook instance of circular reasoning assuming the very fact in dispute to justify a notional benchmark.
129. The DRP, instead of correcting this procedural infirmity, proceeded to invent an allocation grid apportioning the group’s total revenue across 15 “functions” with arbitrary percentages (content storage 5%, technology 5%, marketing 5%, customer agreements 2%, copyright protection 2%, etc.), ultimately concluding that 43% of revenue ought to be retained by Netflix India. This ipse dixit exercise bears no resemblance to any recognised transfer-pricing method. It contains no external benchmarks, no economic rationale, and no linkage to risk or cost contribution. It is, at best, a spreadsheet fiction and at worst, a breach of statutory duty.
130. We cannot condone such departures from the law. Chapter X is not an invitation to economic imagination; it is a discipline founded upon objective comparability. As the Supreme Court observed in Vodafone India Services (P.) Ltd. (supra), transfer pricing adjustments cannot proceed on hypothetical or notional income. Similarly, in Maruti Suzuki India Ltd. v. CIT (Delhi), it was held that recharacterisation of a transaction is impermissible unless the arrangement is shown to be a sham or colourable device. Neither circumstance exists here. The agreements between Netflix India and its AEs are genuine, executed, and approved by regulatory authorities.
131. The Assessee’s TNMM analysis, on the other hand, complies meticulously with Rule 10B(1)(e). It benchmarks the Operating Profit to Operating Revenue (OP/OR) ratio against comparable distributors of software and related products a functional peer group supported by OECD guidance. The margins of comparables, after working-capital and assetintensity adjustments, ranged within an interquartile spread of -0.48% to +0.32%, within which the assessee’s 1.36% lies comfortably.
132. The DRP’s rejection of these comparables was perfunctory and self-contradictory. It had itself directed the assessee to furnish adjusted margins but, when those very workings vindicated the assessee, it refused to consider them. Such behaviour betrays non-application of mind and violates the statutory mandate of reasoned decision-making under Section 144C(8).
133. From a jurisprudential standpoint, the TPO’s and DRP’s methodologies also offend the arm’s-length principle codified in Article 9 of the OECD Model and accepted by Indian law. Transfer pricing seeks parity between controlled and uncontrolled transactions, not the creation of income through internal allocation. To attribute 43% of global subscription revenue to an entity that neither owns nor develops the underlying content or technology is to violate the symmetry between function, asset, and risk the triad that defines economic ownership.
134. In the realm of digital economy, tribunals worldwide WarnerMedia India Pvt. Ltd. (supra)(Delhi HC, 2023), Star Den Media Services Pvt. Ltd. (supra) (Mumbai ITAT, 2021), Turner International India Pvt. Ltd. (supra) (Delhi ITAT, 2019), have consistently held that entities engaged in marketing, promotion, and distribution of access to global OTT platforms are to be characterised as limited-risk distributors remunerated on a cost-plus or TNMM basis. None has endorsed a royalty-based attribution absent local IP ownership. Applying these precedents, we find that Netflix India’s FAR profile, asset composition, risk insulation, and contractual obligations unequivocally categorise it as a limited-risk distributor. Its selection of TNMM as the Most Appropriate Method is legally correct and economically justified. The “Other Method” and the DRP’s attribution model are unsustainable in law and fact.
135. Consequently, the entire transfer-pricing adjustment of Rs. 4,44,93,42,724 is hereby deleted. The ALP determined by the assessee under TNMM stands accepted. The recharacterisation of Netflix India as a full-fledged entrepreneur or content-provider is held to be contrary to record and law.
136. Before parting, we note with concern the increasing tendency of transfer-pricing officers to conflate technological presence with economic ownership. The mere existence of servers, caches, or support personnel in a jurisdiction cannot by itself confer value-creation status. Unless an Indian entity controls, develops, or exploits the underlying intangible assets, its remuneration cannot exceed a routine distributor’s return. This principle, though trite, appears to have been forgotten in the instant case.
137. The Assessee’s functions are confined to distribution of access, marketing support, invoicing, and regulatory compliance;
| • | It owns no intellectual property or critical intangible asset; |
| • | It bears no entrepreneurial risk, all costs being reimbursed by AEs; |
| • | The Transactional Net Margin Method remains the Most Appropriate Method; |
| • | The “Other Method” adopted by the TPO and the attribution grid by the DRP are invalid under law; |
| • | Accordingly, the impugned adjustment is set aside in entirety. |
138. Now in so far as enhancement by the DRP u/s.40(a)(i), it has been brought on record that based on order u/s.201/201(1A) of the act passed in assessee’s case in the same year appeal was filed by the ld. CIT(A) and the assessee has opted for settlement of this appeal under Vivad se Vishwas Scheme-2020, thus dispute relating to 201 has been resolved and the assessee has obtained certificate from the competent authority, the copy of certificarte has also been filed before us. Thus, as a consequence of the settlement impugned disallowance u/s.40(a)(i) does not survived and this position has been clarified by the CBDT in Circular No.12 of 2024 for (specifically in FAQ 2022), accordingly, ground No.14-16 have become academic and the same is dismissed.
139. In ground No.2-4, the assessee has assailed the final assessment being without jurisdiction however, it has been submitted that this issue has kept open and liberty may be granted to the assessee to raise in future proceedings in case any need arises, accordingly, these grounds are dismissed as academic in the aforesaid manner.
140. Coming to the ground No.17 wherein the assessee has contended that there is an error in computation of assessed income which has been stated that as per the intimation issued under section 143(1) of the Act, no adjustment was made to the returned income of INR 27,63,46,470. However, the assessment order incorrectly mentions the income computed under section 143(1)(a) of the Act as INR 29,63,00,220 to which the adjustments made in assessment order (of INR 4,44,95,50,224) are added, to take the assessed income to INR 474,58,50,440 (instead of INR 4,725,896,690). Thus, the assessment income is incorrectly higher by INR 19,953,750/-. From the perusal of the computation sheet attached to the assessment order, it is seen that it considers the assessed income after deductions under Chapter VIA correctly as In response to the 4,725,896,690/-. Thus, AO is directed to revise the assessment order to reflect the correct amount of income computed u/s.143(1A) as In response to the 27,63,46,470/-.
141. In so far as Ground No.18 relating to excess interest, the same is consequential and AO is directed to compute the interest in accordance with law.
142. Lastly, with regard to initiation of penalty proceedings, the same is premature and therefore, is dismissed.
143. Accordingly, the grounds of the assessee is allowed in the manner indicated above and ld. AO is directed to compute the income of the assessee in view of our findings and the directions given above.
144. In the result, appeal of the assessee is partly allowed.