Bridging Innovation And Competition: The Cross-Licensing Dilemma In India’s Digital Economy
Introduction
In January 2024, the Federal Trade Commission (‘FTC’) ordered five big firms to reveal details about their partnerships and agreements with GenAI companies and significant cloud service providers. The main objective of this investigation was to uncover the financial magnitude of these transactions and understand their strategic drivers and real-world ramifications. Further, in January 2025, FTC Chair Lina Khan specifically remarked on the dangers posed by such agreements, stating “how partnerships by Big Tech firms can create lock-in, deprive startups of key AI inputs, and reveal sensitive information that can undermine fair competition.” Statements like these reflect a broader concern; while cross-licensing might be legal and even efficient, it can also become a vehicle for exclusionary practices, especially when employed by firms with outsized market power.
Historically, both competition law and intellectual property law have engaged in a tenuous tug of war, with each attempting to achieve a distinct but equally important policy objective. Competition law was formulated to optimise public interest by maintaining efficiency, consumer welfare, and market contestability. It aims to stop cooperation, abuse of power, and monopolistic behaviour. On the other hand, the foundation of intellectual property is based on the notion that artists and innovators should be given exclusivity to appropriately safeguard their investments. This uneasy coexistence makes their intersection one of the most contested areas of modern legal scholarship.
This intersection becomes quite pronounced, especially in cases of cross licensing agreements where the line between legitimate collaboration and anti-competitive collusion often becomes blurred. A cross-licensing agreement can be formally defined as a contractual arrangement in which two or more parties mutually grant licenses to their respective intellectual property rights, thereby allowing each party to access and utilise the other’s protected technologies, inventions, or works of authorship.
In this newfound era defined by big tech consolidation, cross-licensing is on the rise. While in principle, such agreements facilitate innovation by reducing transaction costs and avoiding litigation, they also raise serious antitrust concerns. Agreements that involve Artificial Intelligence (‘AI’) present a newfound problem because of the growing entanglement of these AI models with data ownership. Data ownership in this context does not merely mean possession of raw data, but extends to control over how datasets are collected, processed and deployed to train these models by big tech consolidations. This control transforms data into a decisive competitive asset.
This article, therefore, examines the issue through four key lenses. First, it highlights the inherent tensions between IP law and competition law, focusing on how exclusivity and market fairness collide in licensing agreements. Second, it defines the antitrust risks present in cross-licensing arrangements, including cross-jurisdictional analysis. Third, it recognises data as the new competitive frontier in big tech agreements and its modern-day implications and fourth, it evaluates a potential framework that can be built upon to address the evolving changes in this dynamic field.
Antitrust And Intellectual Property: A Delicate Balancing Act
The intersection of IP law and antitrust is a recurring phenomenon in cross-licensing where the line between competitive cooperation and anti-competitive collusion is blurred. In India, this tension can be seen in Section 3(5) of the Competition Act, 2002, which allows IP holders to impose reasonable conditions for the protection and exercise of their IPRs without coming under antitrust scrutiny, though the provision remains ambiguous, as it neither defines the term “reasonable” condition, nor does it clearly delineate the scope of the IP protection.
Consequently, the Competition Commission of India’s (‘CCI’) case-by-case assessment has often resulted in inconsistent outcomes. For example, in K Sera Sera Digital Cinema Private Limited v. Pen India Limited & Ors., IP restrictions to prevent piracy were upheld, while in FICCI – Multiplex Association of India Federation House v. United Producers/ Distributors Forum, refusal to supply and price-fixing were denied cover under Section 3(5). Moreover, IP exclusivity continues to create friction in marketplaces because of the resultant practices of tying, refusal to license, and exclusionary cross-licensing or patent pooling that raises entry barriers. This has prompted antitrust regulators around the world to deal with this dilemma and scrutinize the scope of the IP protection claimed. The relationship between the two thus remains a contentious one, with evolving jurisprudence aiming to find a middle ground in this conflict.
Cross-Licensing: Collaboration Or Collusion?
A cross-licensing arrangement refers to a contractual understanding through which two or more entities permit one another to use their respective patents, technologies, or creative works, thereby ensuring reciprocal access to each other’s protected innovations. These arrangements sometimes raise anti-competitive concerns due to heightened risk of collusion, cartelization, and price-fixing. However, authorities have not been able to rely on a formal rigid criterion because these agreements also have major pro-competitive effects, especially in technological transfers. Cross-licensing arrangements result in goods being produced at lower costs by potentially more firms. Thus, these transfers have proven useful in Information Technology (IT) Industries where a product is owned by a large number of parties, such as the semi-conductor or mobile phone industry. Further, sharing through licensing can also result in better service delivery by enabling firms to personalise offerings, reduce error rates, and improve responsiveness.
Nevertheless, cross licensing agreements having both pro-competitive and anti-competitive concerns have still come under the scrutiny of antitrust regulators. In 2021, Acting FTC Chairwoman Rebecca Kelly emphasized upon the need for strict enforcement of antitrust laws to prevent anti-competitive practices by dominant firms, especially in high-technology markets and those involving intellectual property. The European Commission has also recognized that cross-licensing agreements have the potential to allow dominant firms to leverage their power to gain further market advantage at the expense of smaller rivals or facilitate tacit collusion. It clarified in a report that cross-licensing may have exclusionary effects if there is a discrepancy in the cross-licensing terms offered to different firms, due to which some players may assume a greater competitive advantage. In India, Micromax v. Ericsson held that asymmetric royalty rates i.e. charging different royalty rates or different commercial terms to licensees in the same category amounts to anti-competitive conduct.
The European Union’s (‘EU’) Technology Transfer Block Exemption Regulation (‘TTBER’) is an instructive example of how licensing frameworks can build structured oversight. The Regulation explicitly defines technology transfer agreements as ‘agreements which provide for the licensing of technology for the production of goods and services’. It introduces a safe harbor threshold providing that the combined market share of the parties to the agreement should not exceed 20% in the relevant market. This provision exempts parties falling under this limit without any antitrust scrutiny under Article 101(1) and Article 101(3). For agreements exceeding the threshold, it provides for a case-by-case analysis of the terms of the agreement. The regulation finds relevance in the present context because while there is a difference between transfer of technology and data as an input, the distinction becomes increasingly blurred in digital markets where data-driven algorithms significantly influence competition because of access to large datasets.
Similarly, the U.S Antitrust Guidelines for the Licensing of Intellectual Property also contain a ‘safety zone’, where a licensing agreement is granted a presumption of legality if the combined market share of both entities is below 20%. Both of these jurisdictions therefore recognise the need to draw a line that clearly demarcates creative innovation from anti-competitive conduct.
In the Indian landscape, the CCI has closely examined the impact of cross-licensing agreements in a number of industries in India, including the technology, biotechnology, and pharmaceutical sectors. A 2019 report analysing such arrangements in the biotech industry stated that cross-licensing among major firms often results in the creation of entry barriers for smaller firms, which can lead to the disincentivization of firms to innovate. It clarifies that major agricultural firms engaged in cross-licensing of agricultural technologies (transgenic technologies, genome editing technologies, etc.) with start-ups has also led to a large number of ‘non-merger mergers’, i.e. there is no change in ownership, but they still have the potential to result in vertical integration and exclusive product ranges, thus becoming anti-competitive Thus, a careful and nuanced approach by antitrust regulators globally is required to balance all the variables.
Data As The New Competitive Frontier In Cross-Licensing
In the landmark Meta-CCI case in 2021, the CCI for the first time explicitly recognized data as a major non-price factor in antitrust inquiries. This was a watershed moment in Indian competition law, as it shifted the focus away from traditional price-centric assessments to the realities of the digital economy where user data is the most valuable asset. In the landmark case of Matrimony.com Limited v. Google LLC & Others, which involved Google’s misuse of power by imposing unfair terms on its partners in search intermediation, the CCI noted that “it would not be out of place to equate data in this century to what oil was to the last one”. In modern markets, data functions as the new currency. Control over large volumes of data can create strong network effects, increase switching costs for consumers, and ultimately entrench market dominance.
This case reflected how companies such as Meta build and maintain their market power by collecting and controlling vast quantities of user data across platforms. In the era of AI-driven markets, this dynamic becomes even more complex. AI tools rely on enormous datasets to function effectively, and Big Tech firms often channel user data into their AI systems, outsourcing its processing to proprietary algorithms. In this way, control over data is increasingly mediated by AI solutions, which are operated by dominant firms. While legal ownership of data continues to vest with the firm, AI enhances the efficiency and scale with which this control is exercised, allowing firms to extract greater strategic value from the same datasets. This makes anticompetitive conduct harder to detect and regulate, not because AI constitutes a separate market, but because exclusionary effects now arise through algorithmic design choices and internal data governance rather than through overt contractual restrictions. What is framed as product improvement or optimisation may, in fact, be the strategic use of data to entrench market power and gradually foreclose rivals. As discussed beforehand, there is no formal, rigid criterion for these markets. This leads to uncertainties, as regulators have acknowledged that it may be premature to conclude that access to data from one app or service, such as a messaging platform or a social media feed, always acts as an entry barrier. However, when diverse datasets are combined to train models, their competitive value multiplies, giving incumbents an advantage that is extremely daunting for new entrants to replicate.
This is precisely where cross-licensing in AI adds another layer of difficulty. Cross-licensing in traditional industries such as semiconductors or pharmaceuticals was already complex due to overlapping patent portfolios and risks of foreclosure. However, in case of AI, the stakes are higher. Licensing arrangements now cover dynamic and interdependent inputs: training datasets, compute capacity (GPUs and TPUs), proprietary algorithms, and cloud infrastructure. While these spare commutes are a fully-fledged business model, these agreements often involve bundled or preferential access to these inputs. This in turn, has anti-competitive concerns, as it blurs the lines between legitimate upstream supply and strategic control over these critical resources. Each of these is indispensable to building and scaling AI models, and cross-licensing agreements often extend far beyond a simple patent swap. They effectively decide who controls the pipelines of innovation.
For example, granting access to a proprietary dataset or model is not the same as licensing a chip design. It can determine the direction of a competitor’s innovation, especially if exclusivity clauses tie access to a dominant cloud provider or limit interoperability with rival platforms. A recent example is the EU’s probe into anti-competitive practices in the AI chip market. This investigation was brought about by the fact that Nvidia, the largest manufacturer of AI chips, holds about 80% of the market share in the GPU market, prompting a question as to whether intervention is required or not. This is relevant in the context of cross-licensing because Nvidia’s dominance is not limited to chip manufacturing alone but extends to its proprietary software stack and licensing ecosystem, such as CUDA (Compute Unified Device Architecture). It also extends to related IP, as Nvidia’s graphics processing units are found in all global power AI applications such as ChatGPT.
In this sense, AI cross-licensing has become more than just a mere legal mechanism for sharing IP. It is a tactical tool to influence the innovation pipeline by merging IP exclusivity and data control. For firms with established market positions, such arrangements may also function as a mechanism to preserve technological parity with close competitors, preventing disruptive shifts in market leadership. A useful illustration of these incentives can be seen in markets characterised by overlapping patent portfolios and high innovation costs, such as semiconductors or telecommunications. Consider a situation where two leading firms each hold a set of patents that are essential for developing next-generation technology. A cross-licensing agreement would present as a dual-win approach for both parties involved, as it will secure uninterrupted access to each other’s technologies, which would allow them to continue innovating their products without the risk of injunctions. It is far more difficult for regulators to balance genuine collaboration and exclusionary control in this dual control setup. Thus, there is a need for a new framework to solve this conundrum.
Navigating The Road Ahead: Regulatory Pathways For India’s Digital Market
The challenge for India is that no such structured guidance currently exists under the Competition Act. Section 3(5) provides only a broad exemption for ‘reasonable conditions’ tied to IP protection, but the absence of technical standards means cross-licensing agreements in digital markets are evaluated on an ad hoc basis. This is problematic when the stakes involve AI models and cloud infrastructure that underpin the next wave of innovation. Unlike conventional markets where harm can be easily quantified, AI and data-driven industries evolve so rapidly that anti-competitive effects may become entrenched before regulators intervene.
CCI Chief Ravneet Kaur stated in a recent conference that a forward-looking approach needs to be enabled in the intersection of AI with antitrust laws. This intersection becomes important as unlike traditional markets, digital markets, by virtue of their multisided nature and the presence of strong network effects, often experience a ‘winner takes all’ phenomenon. This is where a platform garners enough market power in a relatively shorter duration so as to create a lock-in effect for its consumers, and successfully drives out competitors and creates entry barriers in the market. While such winner-takes-all dynamics would prima facie present strong incentive for firms to maintain exclusivity, dominant firms often engage in strategic licensing of AI and new technologies with niche firms to further their data advantage, and thus entrench their position in the market.
Naturally, a straightforward option would be to directly draw from an established framework such as the safe harbor mechanism employed in the EU. However, as seen in the failed example of the introduction of the ex-ante regime in the form of the Digital Competition Bill which was largely modelled after the EU’s Digital Markets Act, this would not be the most optimal approach. The primary reason for this can be attributed to the fact that India’s marketplace significantly differs from that of the EU. Internet penetration and the familiarity of consumers with digital platforms is still growing and evolving in India, resulting in varying preferences and developments. This is because internet accessibility still remains a myth for many parts of the country. Additionally, the CCI’s position is that an IP licensing agreement would be unlikely to invite scrutiny for having an Appreciable Adverse Effect on Competition (AAEC) unless there is substantial market power held by the licensor. Thus, a block exemption mirroring the TTBER could undermine the case-to-case effects-based approach prevailing currently. Thus, a one-size-fits-all blanket regulation would not be well-suited for India due to the fragmented nature of the tech and AI market.
Therefore, the authors suggest the obvious first step towards regulation of licensing agreements in India should be for the CCI to provide some clarity on the matter in the form of its Frequently Asked Questions (FAQs) or a set of guidelines which effectively define a ‘reasonable condition’ as stipulated under the Act. Further, it must delineate the scope of such conditions and better assist the stakeholders before pushing for a major overhaul with the help of the legislature.
A useful tool that can be used in such cases are here is the “ability, incentive, and effects” framework, which is commonly used for the assessment of vertical mergers. First, authorities must assess whether a partnership impacts a firm’s ability to impact market outcomes based on the amount of market power it holds. For instance, a licensing deal tied to telecommunication technologies can tilt the balance of power in favour of one partner by controlling essential inputs. Second, regulators must examine how the partnership alters the incentives of the parties, whether it nudges them towards cooperative rather than competitive behaviour. Lastly, it is necessary to assess the overall impact on customers and competition, with attention to whether the agreement creates exclusivity around critical resources. The framework will allow the authorities to operate at an earlier, more diagnostic stage and to flag competitive risks arising from licensing arrangements even before a firm clearly satisfies the dominance threshold under Section 4.
To further elaborate on the telecommunications example and to show its implementation in the ‘ability, incentive, and effects framework,’ let us consider a licensing agreement related to mobile communication, such as 4G or 5G infrastructure. If a dominant firm with substantial market power licenses such technology on preferential terms to a selective network operator, it may acquire the ability to influence downstream market outcomes by controlling access to critical input. Competitors who are denied similar access may be compelled to operate under technological or cost disadvantages, thereby weakening their competitive position. Such arrangements also alter incentives. Rather than competing independently on price, quality, or innovation, the parties to the agreement may have incentives to align their commercial strategies, delay rival entry, or soften competition in adjacent markets. From an effects-based perspective, the long-term implications for consumers and market structure become critical, particularly where these agreements result in exclusivity over network infrastructure or interoperability standards.
The jurisprudence of AI and competition law is evolving at an unprecedented pace, with the challenge of Big Tech cross-licensing agreements poised to become one of the most pressing dilemmas for antitrust authorities worldwide. This concern is not merely theoretical, with several cross-licensing agreements worldwide being criticised for entrenching market power and restricting competition. The most prominent example is the cross-licensing of Standard Essential Patents (SEPs) by the telecom giant Ericsson with various tech companies like Lenovo, Intex, etc. These IP licensing arrangements have invited antitrust scrutiny from regulators around the world for alleged excessive royalty conditions and exclusionary conduct. Another recent example is the FTC probe into the Microsoft-OpenAI deal, focusing on the software licensing practices between the two, which has led to concerns about market dominance and restrictive data sharing agreements.
Jurisdictions such as the EU and the US have already taken proactive steps, setting out structured frameworks to address these concerns. India, by contrast, still lacks a dedicated mechanism. Failure to do so will inevitably convert today’s regulatory gap into tomorrow’s regulatory crisis.
Conclusion
Cross-licensing agreements hold a distinctly complex place in the overlap of intellectual property and competition law, especially in India’s rapidly changing digital market. While such arrangements promote innovation and reduce unnecessary litigation, but they also raise several antitrust concerns. These concerns magnify substantially in AI-based and data-centric markets, wherein access to datasets and algorithmic control not only holds a significant value in the market, but also influences the very direction of the market.
When combined with the presence of strong network effects, feedback loops, and the dynamic nature of digital markets, such agreements pose a threat to fair market practices by raising entry barriers and discouraging innovation.
In India, the current antitrust framework centered around the Competition Act, 2002 does not provide enough clarity and direction to address such concerns. A balanced effects-based approach which draws inspiration from global practices like those of the EU but still takes into account India’s distinct market conditions is the need of the hour. CCI should aim to develop more robust standards for assessing such arrangements, and the broader intersection of competition law and IPR. Such an approach is imperative for ensuring that innovation acts supplementary to competition, rather than as its enemy.