Much of India’s digitalisation rests heavily on the network architecture that supports it. Rapid digital transformation, and AI-heavy workloads contingent on 5G network deployment, have reignited debates on efficiency, management and net neutrality.Traditional mobile networks faced an irreducible provisioning dilemma as the network was designed to serve heterogeneous traffic like consumer broadband, IoT sensors, enterprise applications and emergency communications, which had to be provisioned for peak demand of the most demanding use case. Network slicing converts this static overprovisioning problem into a dynamic allocation problem. The same physical resources can be reconfigured, in real time, to serve whichever workload is actually present.5G functionality allows telcos to divide one physical network into multiple, independent virtual networks by using different ‘slices’ of the same spectrum band. Each slice is isolated, self-contained and secure, and can be independently configured to meet specific performance needs of different services or user groups.Operators would be able to charge differently based on the service characteristics of the slice such as latency requirements, throughput, level of security, degree of mobility and geographical reach. Users of slice would be able to reach new business segments benefiting from enhanced service customisation, greater scalability, and improved network efficiency and allocation.Network slicing also allows operators to charge premium pricing and increase per-user revenue. While it ensures that critical applications like autonomous vehicles and robotic surgeries can get uninterrupted service guarantees, projections indicate that it can increase average revenue per user (Arpu) by 30-40% from enterprises.In any case, for such services, the ability to carve guaranteed low-latency slices is technically necessary. These are not cases where ‘good enough’ shared bandwidth will do. The International Telecommunication Union (ITU) and most 5G standardisation bodies treat slicing as a core 5G feature.But this creates a tension with net neutrality rules that by design mandate content- and application-agnostic non-discriminatory tariffs. This sits alongside the economic sustainability rationale for operators to recoup return on investments for their 5G infrastructure.Also, India’s operators ostensibly face a spectrum crunch, with insufficient contiguous spectrum, especially in the mid-band, to build genuinely high-capacity 5G networks at scale.In a spectrum-abundant environment, an operator would simply expand capacity rather than monetise scarcity through priority tiers. In a capacity-constrained environment, economics push toward selling priority on a deliberately or unavoidably congested network.When a telco sells ‘priority’ to consumers on a congested network, there is an implicit corollary: everyone not on the priority tier experiences worse service than they otherwise would. The network hasn’t become larger; the slice has just redistributed who gets the good part.Network slicing, while technically required and economically desirable, could impact basic internet users unlike other spectrum-sufficient jurisdictions where network neutrality rules and operational efficiencies from slicing can exist parallelly.Moreover, whether the network is truly spectrum-constrained or capex-constrained — insufficient cell site density given subscriber growth — is analytically distinct. If congestion is avoidable through investment that operators are not making, then permitting priority tiers on the basis of that congestion is not solving a technical problem; it is allowing revenue from the problem to substitute for the incentive to solve it.Rent-seeking, and not innovation, then, is the motivation. It isn’t clear where the evidence stacks up.India’s market structure adds another dimension to the issue. With effectively two dominant operators, Reliance Jio and Bharti Airtel, the coordination risk that operators in concentrated markets have incentives to collude on quality tiers is more plausible than in more competitive markets.Network slicing for enterprise, industrial and genuinely specialised services has a defensible technical and economic rationale that doesn’t threaten the open internet. Evidence for harm is weak in these contexts. But slicing deployed as a consumer retail priority tier on a congested or spectrum-constrained network rests on a much less certain foundation.The mechanism by which it harms non-priority users is real. Incentive for operators to engineer, or tolerate, congestion is structurally present. India’s market concentration amplifies both concerns.Telecom Regulatory Authority of India (Trai) has not resolved this ambiguity proactively. A rigorous regulatory approach would require operators to demonstrate, before any priority tier is permitted, that:They have deployed all held spectrum in congested areas at technically feasible density.Their capex trajectory per subscriber has been maintained relative to subscriber growth.Congestion events being cited occur despite — not in absence of — that deployment.None of that evidence has been placed on record.(Disclaimer: The opinions expressed in this column are that of the writer. The facts and opinions expressed here do not reflect the views of www.economictimes.com.)