For an average user, competition issues in AI are not abstract. They directly affect choice and innovation. For instance, if a dominant company integrates its own AI assistant into its search engine by default, users are more likely to rely on that tool instead of exploring alternatives. Over time, this may reduce competition and limit innovation which goes against a core objective of CCI which is to promote innovation. The concern is not just who builds AI, but who controls how it reaches users.
In 2025, the Competition Commission of India published its market study on Artificial Intelligence and Competition[1], and what it revealed was not reassuring. The study does not merely map the AI landscape but exposes the structural limits of the very framework charged with regulating it. The study identifies that the AI ecosystem in India is structured around a small number of hyperscalers who simultaneously supply the infrastructure on which competitors depend and compete against those same competitors in downstream markets. Control over data, compute, and foundation models is concentrated at the top of the stack. Startups seeking to build AI products must enter contractual arrangements with the very incumbents against whom they nominally compete, arrangements that may channel data back upstream, deepening the concentration they were meant to disrupt. To put it plainly, imagine a toll road, where the toll booth operator is also your biggest competitor in the market you are trying to reach. This is the structural condition that CCI has identified in India's AI economy for which there is no ready answer.
The Competition (Amendment) Act 2023[2] introduced the Deal Value Threshold ('DVT') of INR 2,000 crore precisely because transactions like Facebook's acquisition of WhatsApp and Microsoft's acquisition of LinkedIn had escaped the Indian merger scrutiny entirely. The DVT was designed to bring data-rich transactions into view. Yet a threshold that determines notifiability does not address the underlying structural condition, that is data concentration compounds faster than legislative responses arrive, and the competitive harm it creates may be unreachable by the time the law turns to look at it[3].
The window for meaningful legislative response is closing faster than the current pace of reform acknowledges. India's competition framework increasingly recognises the risks of data concentration but remains doctrinally and procedurally ill-equipped to address the infrastructure-layer power it produces.
- From Recognition To Reluctance: India And Its Encounter With Data Concentration
- The Doctrinal Gap In AI Markets
The CCI's AI market study is candid about the merger control problem. It acknowledges that established firms may acquire emerging competitors to forestall future competition, and traditional thresholds may not capture where value resides in data pipelines and model architectures. The Deal Value Threshold (DVT) attempts to fix this by bringing such transactions in notification requirement. However, it only answers when a transaction must be reported, not how the deeper problem of data concentration should be addressed.
Algorithmic systems operating on concentrated data pools can coordinate market outcomes without an agreement in the traditional legal sense[4] evading the requirements of Section 3 of the Competition Act entirely. The NCLAT in Samir Agarwal v. Competition Commission of India, confirmed that dynamic algorithms do not constitute an 'agreement' or concerted practice under Section 3 as they respond to market conditions rather than explicit collusion[5]. The law, therefore, requires proof of agreement while the market has moved to coordination by architecture.
Consider how this plays out in practice. When Zomato and Swiggy adjust delivery charges in real time such outcome may be produced by algorithms responding to the same market signals. Both companies act independently, not by any agreement, but the result looks like coordination. The law will however require proof of an agreement something algorithms, by design, don't produce.
The ICRIER policy brief on digital competition reinforces this diagnosis from the legislative side[6]. Data-sharing obligations proposed under the Digital Competition Bill faced sustained backlash on the ground that they would adversely affect MSMEs dependent on targeted advertising through digital intermediaries. The Digital Competition Bill remains a draft. Meanwhile, CCI's own study shows that the AI stack is not waiting.
Beyond the substantive doctrinal gap, the procedural architecture compounds the problem. India's merger notification regime gives merging parties considerable discretion to define the scope of disclosure. Pipeline assets, adjacent data markets, and vertical relationships may never surface in a Phase I filing. By contrast, the Hart-Scott-Rodino framework in the United States empowers the FTC and DOJ to issue binding second requests compelling the submission of additional documents and internal communications wherever competitive risks are suspected[7]. India's framework incentivizes parties to structure filings that avoid triggering Phase II scrutiny entirely. In data-driven markets, where the competitive harm is often in what is not disclosed rather than what is, that asymmetry is not a procedural detail. It is a structural vulnerability.
- BookMyShow: The framework Identifying Harm It Cannot Reach
In India most movie tickets are booked on BookMyShow. Very few people might have noticed that BookMyShow is essentially controlling 70-80% of India's online ticketing market. When a new platform tries to enter, theatres often already are committed to exclusive deals with BookMyShow. In 2022, CCI noticed this too.
In 2022, the CCI looked at BookMyShow[8]. Data exclusivity, it found, would entrench network effects and generate monopoly rents. The commission saw the structural harm clearly and directed investigation. Four years later, the final order arrived and the 2022 finding was nowhere in it[9]. Instead, the Commission asked whether BookMyShow had treated similarly situated cinemas unequally under Section 4(2)(a)(i). Single-screen cinemas and multiplexes are different, it found. No discrimination was established.
The original question in the initial order, that is whether exclusive data ownership by a dominant platform entrenches competitive harm, was never answered. Instead, it was swapped for a question the commission could comfortably resolve and resolved the same in the platform's favour.
This is not about the commission getting the law wrong. On the narrow question it asked, the reasoning holds. The problem is the question itself. Data exclusivity is not a discrimination story but rather a compounding market power story. What the 2022 order identified, is exactly what the AI market study warns is happening at scale across India's digital economy. The commission saw it in 2022. The toolkit did not reach it in 2026. In AI markets, that gap will be considerably more costly.
- A Pattern, Not An Accident
The connection between the AI market study and the BookMyShow sequence reflects a consistent pattern in how India's competition framework encounters data-driven market power. It identifies the concern, flags it for attention, and then finds the available instruments do not reach it.The DVT was adopted without the institutional conditions that made equivalent instruments effective in Germany and Austria, whereas the Standing Committee on Finance was informed[10], Germany had identified only a handful of DVT-triggered notifications and Austria was yet to find a single anti-competitive combination that caught its equivalent provision. India adopted the form without the substance[11]. The Digital Competition Bill calibrates its framework to avoid burdening emerging Indian enterprises with the capacity to grow into global players. This framing has consistently been applied to preserve incumbent flexibility and delay enforcement.
The result is a framework that is structurally more legible to concentrated interests than to diffuse ones. Not because the CCI lacks capability, but because the architecture of consultation, the ideology of development, and the rhythm of legislative signalling each independently produce outcomes that favor incumbents. BookMyShow and the AI market study are not isolated data points. They are the pattern made visible.
- The Window Is Closing: What India Must Do Before The Market Structure Hardens
The BookMyShow order is not a failure of judgment. The commission resolved the question before it correctly on the evidence available. But correctness on a narrow question and adequacy as a response to structural harm are different things and the gap between them is precisely where India's data economy is consolidating.
What the CCI's AI market study implicitly acknowledges, and what the BookMyShow litigation makes explicit, is that India's competition framework is encountering a structural condition it was not designed to address. That is data as infrastructure, controlled by incumbents, generating advantages that accumulate faster than legislative responses arrive. In the 1990s, regulatory inaction during the MRTP transition allowed intra-group consolidation to lock in before the Competition Act commenced[12]. In the digital decade, the framework's calibration to an earlier economic era allowed platform concentration to consolidate before the DVT reached transactions of consequence. In AI markets, the same dynamic is operating at the infrastructure layer, and the legislative response is arriving after the market structure it nominally targets has begun to solidify.
India does not need another market study to identify the problem. The CCI has already done that, twice with the AI study and in BookMyShow's 2022 order. What it needs is a data exclusivity theory of harm under Section 4(2)(b) that the Commission can deploy without requiring class comparisons between counterparties. It needs AI-specific merger guidelines clarifying how the DVT applies to model acquisitions and data pipeline transactions.
The BookMyShow order leaves one question unanswered. This is if a dominant platform's data practices entrench network effects and generate monopoly rents, as the commission itself observed in 2022, which provision of the Competition Act reaches that harm?
Until the Commission answers that question, the framework will keep seeing the harm clearly—and remain unable to touch it.
CCI, Market Study on Artificial Intelligence and Competition (CCI 2025). ↑
Competition (Amendment) Act 2023 ↑
Ayush Raj, 'Merger Control and Competition (Amendment) Act, 2023: Analysing the Amendment and Advancing Post-Amendment Considerations for the Commission for an Immaculate Combination Regime' (2024) NLIU Law Review ↑
Sakshi Gupta, 'Artificial Intelligence and Competition Law in India: A Legal Response to Algorithmic Market Collusions' (2025) 1 European Economic Letters. ↑
LiveLaw, Supreme Court Upholds CCI Order Dismissing Complaints Alleging Anti-Competitive Practices by Uber & Ola (2020) ↑
Payal Malik, Competition Issues in Digital Markets (2024) ICRIER Policy Brief ↑
FTC, FTC Finalizes Changes to Premerger Notification Form (10 October 2024) ↑
CCI, Vijay Gopal (Proprietor, Vanila Entertainments) v. Big Tree Entertainment Pvt Ltd (BookMyShow), Case No. 46 of 2021 (Order dated 16 June 2022) ↑
LiveLawBiz, CCI Clears BookMyShow Of Abuse Of Dominance Allegations Despite Dominance In Online Movie Ticket Booking Market (12 March 2026) ↑
Vidhi Centre for Legal Policy, Regulating India's Digital Markets: A Framework for Ex-Ante Regulation (2024) ↑
Tarushi Tewari, Evaluating the Deal Value Threshold under the Competition (Amendment) Act, 2023: A Comparative Analysis and Recommendations for India's Merger Control Regime (2025) IJIRL ↑
Nemika Jha, Political Economy of Takeover Regulation in India: How Good is India's Mandatory Bid Rule? (2024) NLS Business Law Review ↑
Views are personal.