Deepandru Singh is a fourth-year and Bhaumik Pratap Singh is a second-year students at Maharashtra National Law University, Nagpur
A New Era of AI-Telecom Partnerships
Imagine your gadgets not just featuring calls and data but also a great AI assistant to answer questions, compose emails and summarize reports. For 360 million Bharti Airtel customers, that is now a reality thanks to the company’s partnership with Perplexity AI, providing a year of Perplexity Pro free of charge.
The alliance is clearly beneficial from a business perspective. Perplexity gains immediate access to one of the world’s largest telecommunications user bases, while Airtel benefits from a distinctive service by which it can carve out its niche in a highly competitive market. Reports suggest that similar AI partnerships including potential integration of Reliance Jio with OpenAI’s ChatGPT and Meta’s LLaMA, prove that AI integration in telecom services will become business as usual in near future.
But behind the marketing veneer is a series of legal and regulatory issues. The Competition Commission of India (CCI), with the Competition Act, 2002, including 2023 amendments granting it more powers, now has a wider mandate to inquire into such tie-ups. With “material influence” now officially established as a test of control and a new deal value threshold aimed at detecting big-ticket digital deals, the CCI is poised to probe tie-ups that may redefine competition in the AI and telecom sectors.
Why the 2023 Amendments Matter
Two of the amendments to the Competition Act are particularly relevant to AI collaborations. The first is the direct inclusion of “material influence” under Explanation (a) to Section 5 as a control, as it covers the ability to influence another enterprise’s management or strategic decision in any manner. The second is the inclusion of deal value threshold under Section 5(d), requiring notification to the CCI for any transaction exceeding ₹2,000 crore with a target having “substantial business operations in India”, even if the normal turnover or asset thresholds are not exceeded. Although the deal-value threshold is formally sector-agnostic, its utility is clearest in digital and new-age markets, where competitive influence often exceeds what the balance sheet reflects. The Delhivery-Ecom Express deal approval illustrates this: despite Ecom Express’ weakened financial position, the Commission reviewed the transaction because its digital logistics infrastructure carried strategic weight. The case underscores why high-value collaborations in technology-driven sectors merit scrutiny even when traditional turnover or asset tests are not met.
These provisions imply that even in the absence of majority ownership, a structure can be deemed to be a notifiable combination if one has strategic control over the other’s decisions. This can be the governance rights, contractual terms, or privileged access to confidential business secrets.
How CCI Interprets ‘Material Influence’
European Commission has rightly pointed out that even limited rights over training data can give meaningful leverage where digital innovations shape competitive outcomes. However, CCI’s concept of material influence has matured through a series of precedential orders. In UltraTech–Jaiprakash , the Commission found that even a minority shareholding combined with presence on the board and industry status could confer material influence. In the Jet–Etihad itself, 24% shareholding along with rights of governance such as veto over strategic decisions and a seat on the board were enough to warrant scrutiny. In SCM Soilfert–Deepak Fertilizers , again 24.46% holding was sufficient to trigger the threshold because it was coupled with veto rights in relation to strategic decisions.
The January 2025 Goldman Sachs–Biocon Biologics deal sent an even stronger message. In this matter, gun-jumping arose when companies acted as one before the merger deal was approved for instance, by sharing confidential business information or coordinating their commercial decisions, which competition law does not permit. Thus, Goldman Sachs was fined by the CCI for “gun-jumping” when it bought only 3.81 per cent of the firm but bought rights overboard information, observer rights, and veto rights much higher than a minority shareholder would otherwise enjoy. These guidelines make it clear that even a minority holding with qualitative rights such as access to proprietary training data and AI computing hardware, rights affecting model development and deployment parameters through cloud infrastructure may be on the CCI’s radar in telecom-AI partnerships.
The Deal Value Threshold in the Digital Context
The value threshold was introduced to close enforcement loopholes that are created due to “killer acquisitions,” whereby larger incumbents take over startups before they can become competitively threatening. As per the 2025 Merger FAQs of the CCI, in digital markets, “substantial business operations” can be formed if a party has more than 10 per cent of global users or turnover in India.
This has repercussions for AI alliances. For instance, if Airtel were to acquire a stake in Perplexity for more than ₹2,000 crore, notification would be necessary even if Perplexity’s present Indian turnover is marginal, as long as its Indian base exceeds the 10 per cent mark. The idea is to catch deals early in the expansion cycle, when they are not large enough to change market structure in irreversible ways.
Sections 3 and 4: Behavioural Scrutiny of Partnerships
Even in the absence of equity, AI partnerships may be examined under Section 3 on anti-competitive agreements and Section 4 on abuse of dominance. Agreements of exclusive supply or distribution are subjected to a “rule of reason” test under Section 3(4) based on Section 19(3) considerations, i.e., foreclosure of competition, consumer well-being, and technological advancement. Unless Airtel-Perplexity collaboration disallow either of them to work with other telecom or AI partners , the CCI would examine whether any exclusivity between them shuts out rival collaborations or substantially forecloses competition. In Meru Cabs v. Ola/Uber, the Commission declined to intervene against exclusivity agreements where consumers and service providers could use multiple platforms, a system known as multi-homing. Similarly, in AI search engines, substitutes such as ChatGPT, Gemini (Google Bard), and Bing AI are still broadly available, which currently reduces foreclosure risk.
Bundling is also under the spotlight. Airtel’s inclusion of Perplexity Pro free with its data offers is a bundling, but non-coercive, because customers are not required to sign up for the service. If a dominant telephony player were to make use of one AI service as a default one with favourable speeds or zero-rating, this would be objectionable under Section 4(2)(c) for denial of market access or Section 4(2)(d) for abuse of dominance in one market to enter another.
Data Sharing and Algorithmic Risks
The 2025 FAQs explained the definition of “affiliate” to include those who have access to one another’s commercially sensitive information. Access to user data in AI alliances can create competitive advantages that are difficult to replicate by others. For instance, if a telecom operator allows its AI partner access to anonymised but granular user search queries, network usage patterns, or language preferences, the AI system can refine its models faster and offer more accurate or localised outputs. Competing AI providers without access to comparable datasets may find it significantly harder to match this quality, even if they possess similar technical capabilities. The CCI’s Market Study on the Telecom Sector in India also highlights how control over consumer data and traffic patterns can strengthen market power and raise entry barriers, particularly in digital ecosystems where data-driven feedback loops reinforce scale advantages.
Another emerging concern is algorithmic collusion, a concept that may be understood as a situation where pricing or competitive outcomes become aligned not because firms explicitly agree with each other, but because their algorithms learn to respond in similar ways to market signals.. The Chairperson of CCI has warned that AI systems can, without human collusion, converge to outcomes such as prices, reducing competition in a way that is hard to detect. While Airtel and Perplexity are in different markets and are not competitors in the classical sense, the risk emerges if the same AI model is licensed to competing operators. In such scenarios, regulators may need to examine whether reliance on common technological platforms inadvertently softens competition or facilitates coordinated market behaviour, even in the absence of any explicit agreement.
Lessons from Global Regulators
India is part of a broader global trend of heightened scrutiny of AI-related deals. In the United States, Federal Trade Commission Chair Lina Khan has prioritised enforcement of generative AI and emphasized the importance of acting early to preserve competition. In the United Kingdom, the Competition and Markets Authority applies a material influence test similar to that used by the CCI. This is evident from cases such as Amazon–Deliveroo case, where a 16 per cent stake was investigated due to concerns around network effects, and Meta–Giphy (2021) case, where divestment was ultimately ordered. In the European Union, while merger control is based on the test of “decisive influence,” the European Commission has encouraged Member States to refer below-threshold acquisitions in digital markets. The Illumina–Grail case in the health-tech sector further shows that transactions may be scrutinised even without an obvious dominance issue when they involve strategically important technologies.
These examples underscore that competition authorities across the globe will enforce interoperability, non-discrimination, or even transaction reversal if a merger injures competitive access to essential technologies or information.
Potential Remedies and Future Scenarios
If the Airtel–Perplexity merger was to emerge as a competition problem, the CCI might think in terms of solutions less drastic than blocking the merger. One such option is behavioural commitments, which are ongoing obligations placed on the parties to regulate how they conduct their business post-transaction, rather than altering the structure of the deal itself. In practice, this could include requiring Airtel to give competing AI products equal promotional attention. Information firewalls might be implemented to prevent the transfer of sensitive business data. Non-discrimination provisions would ensure the equal treatment of competing AI programs on the telephony infrastructure.
In the future, material influence is obtained through reliance on technology rather than typical shareholding or control rights. If the AI tool of a company is a vital infrastructure for multiple industries, command over access to the tool could be as good as command over the market. The regulators would need to be sensitive to this change in the sources of competitive power, particularly where control over AI systems, data, or digital infrastructure creates sustained dependency across markets and enables firms to influence competitive outcomes without relying on traditional ownership or control rights, a concern even highlighted by OECD.
Striking the Right Balance
Now, the Airtel–Perplexity tie-up seems to offer clear benefits for consumers.. By bundling an AI assistant that can answer queries, summarise information, and assist with everyday tasks, the partnership enhances the overall telecom experience while encouraging wider adoption of AI-driven services. But the larger context of the CCI means that even such healthy tie-ups are not isolated from regulatory scrutiny. The regulators’ challenge is to encourage innovation without such tie-ups becoming exclusionary gatekeepers, for instance by limiting rival access to essential data, infrastructure, or distribution channels.
Prior practice, e.g., in Meru, indicates the CCI likes to keep a watch on potentially risky structures rather than intervene at any premature stage, intervening only when there is tangible proof of harm. Competition law in the age of AI will have to evolve to deal with situations where influence is exercised by controlling algorithms, data, and vital technological infrastructure by enabling ex-ante algorithmic audits to detect collusive or exclusionary effects (suggested by CCI itself), and by imposing targeted data-access and data-sharing obligations where exclusive control over datasets operates as an essential facility or entrenches market foreclosure.
By keeping a close eye, the CCI can ensure the AI revolution is as much about consumer choice and open competition as it is about tech innovation, and that a few, select alliances don’t have disproportionate power over the future of digital markets. Ultimately, as AI becomes more deeply embedded in telecom services, partnerships like Airtel–Perplexity illustrate how innovation and competition concerns will increasingly intersect, making careful regulatory assessment of such collaborations central to the future of digital markets.