Ian Bremmer’s technopolar world describes a condition in which power has migrated from states to the technology corporations that control digital infrastructure. A conventional response to this situation would be for a state to reassert its authority over technology corporations. For instance, the US has been putting export controls on technology corporations intending to trade with China, a ‘small yard, high fence’ approach since 2022. Over time, this US approach is becoming increasingly exclusionary towards countries beyond China. For instance, in June 2026, the US government issued an export control directive to Anthropic to suspend access to its AI models to “any foreign national, whether inside or outside the United States, including foreign national Anthropic employees”. This approach has helped catalyse a broader turn towards ‘digital sovereignty’, where states increasingly seek to reduce dependence on privately controlled technological infrastructures that may be subject to the interests or regulatory authority of another state.

This dynamic is pushed by technology companies equally, because of their ‘sovereign AI marketing strategy’. Chipmakers and large technology firms have deliberately promoted the language of technological self-reliance to encourage states to invest in homegrown AI infrastructure, while ensuring those investments remain dependent on proprietary chips, cloud services, and vendor ecosystems. 

India’s AI governance discourse has not been immune to this capture. The rhetorical emphasis on sovereignty and self-reliance has proceeded alongside deepening infrastructural dependence on technology corporations. The Indian government has been promoting ‘AI sovereignty’ by providing grants, compute support, and priority in public procurement to indigenous solutions. It has also been promoting the use of the ‘Digital Public Infrastructure’ (“DPI”) approach in achieving the goal of democratising access to AI infrastructure to a wide range of actors. However, there is little independent empirical evidence of whether the current Indian experience with the DPI approach actually democratises access to infrastructure for smaller entities.

This blog argues that what presents itself as digital sovereignty in India is better understood as a political arrangement in which the state provides legal legitimacy, infrastructure subsidy, and a captive user base, while private actors retain operational control and capture the resulting value. This is similar to ‘Sovereignty-as-a-service’, a phenomenon described by Rafael Grohmann and Alexandre Costa Barbosa, which highlights a form of discursive capture where Big Tech companies co-opt political terms to sell back ‘control’ and ‘compliance’ as modular, branded products. 

The Indian arrangement of sovereignty-as-a-service operates through two mutually reinforcing mechanisms. The first is the substitution of rule-based regulation with best-practice governance that include voluntary frameworks, ethical guidelines, and self-certification regimes that perform accountability. The second is the systematic encroachment of private actors into sovereign functions under the Digital Public Infrastructure framework, where public-private partnerships shield these arrangements from competitive scrutiny and democratic oversight.

These mechanisms allow private actors to escape regulation and produce a specific distribution of risk and reward: the state absorbs the risk, while private actors capture the reward. 

Governing by Guidelines

India’s approach to AI governance has been pro-innovation and voluntary, instead of a rule-based system. The government has been favouring industry-led voluntary guidelines, ethical principles and governance best practices to be adopted at the discretion of industry, following an explicit logic of ‘Innovation over Restraint’. It is, however, important to understand what kind of existing power structures in the industry these voluntary governance frameworks promote. 

The global AI industry is an oligopoly structured around a ‘bigger-is-better’ paradigm in which access to compute and proprietary data functions as both a barrier to entry and a proxy for legitimacy. Voluntary frameworks are promoted by states to allow smaller players a ‘level-playing field’ to promote innovation in the AI industry. However, in practice, these frameworks do not apply equally to all actors. They will apply most easily to those organisations that have resources to perform compliance: to hire AI ethicists, publish responsible AI frameworks, and participate in standard-setting bodies. For smaller firms, civil society, and affected communities, voluntary frameworks provide little to no meaningful lever of contestation. As a result, bigger technology companies can potentially use their resources to gain momentum in innovation and function as de-facto private regulators by setting priorities for a sector, determining access, and allocating resources through their decisions. 

In India, many public sector AI solutions are being developed and deployed by the government in collaboration with private technology companies. Without binding standards, there is no clear benchmark against which the design or performance of these systems can be assessed, making the attribution of accountability difficult. In addition, existing legal mechanisms can potentially shield these public sector AI solutions from regulatory scrutiny. For example, Section 54 of the Competition Act exempts enterprises acting in the discharge of public interest, security of state, or sovereign function from competition scrutiny. Although no such exemption appears to have been granted for public sector AI projects, the provision shows that these initiatives could, in principle, be exempt from competition law oversight. Since competition law can scrutinise anti-competitive arrangements in public–private collaborations, including exclusive partnerships and preferential access to public resources, such exemptions could remove an important layer of accountability.

A voluntary guidelines approach and potential reduced scrutiny for entities that collaborate with the government can allow certain private companies to gain an advantage, creating barriers for smaller entities seeking to enter and develop in the sector.

The Indian state’s reliance on voluntary guidelines, limited scrutiny of state-adjacent enterprises, and the absence of binding algorithmic standards represents a broader regulatory choice. Although framed as necessary for technological markets to mature, this approach can entrench existing market arrangements by treating dominant structures as the default conditions of governance rather than as arrangements open to democratic challenge.

Digital Public(?) Infrastructure

While soft governance including voluntary governance frameworks allow the state to avoid directly regulating technology companies, DPI can allow the state to actively support and promote certain actors within the sector. The Indian DPI model separates public purpose from private operation: the state provides legal recognition, infrastructure, and scale, while private actors develop and implement services. As Smriti Parsheera argues, this has contributed to the emergence of ‘alt big tech’ in India, i.e., systems that may reproduce the power of traditional Big Tech through data aggregation, control over digital networks, and influence over technical standards, while deriving their legitimacy from state endorsement rather than purely market dominance. Unlike traditional Big Tech whose power stems from economic scale, data advantages, and early market entry, alt big tech derives influence from its infrastructural role, state-backed legitimacy, and ability to shape participation and technical rules. This creates new concentrations of power with potential implications for competition, innovation, and public interest.

The state-backing of alt big tech extends to financial subsidy as well. For instance, the IndiaAI Mission was launched by the Indian government in March 2024 with its stated purpose as democratising access to AI compute infrastructure and subsidising access to high-end GPU resources available to Indian startups and researchers to contest the concentration of AI capability in larger, international firms. This approach, however, reproduces the problem it claims to solve. Public funds are allocated to selected private recipients who in return, make selected models open source. However, the Mission provides critical infrastructural de-risking: the state absorbs the capital cost of the most expensive phase of AI development, and private actors capture the resulting market position. When those actors subsequently raise private capital, they reinvest in the same compute infrastructure the public already subsidised, deepening the concentration the IndiaAI Mission was designed to break.

The DPI framing wraps government support to private entities in the language of sovereignty, inclusion, and public good. Digi Yatra illustrates this shift: a biometric identity system developed through a government-led initiative and implemented through a public-private institutional structure is framed primarily as a seamless, secure, and citizen-centric innovation. It has also been recognised by the government for its innovative use of AI in improving citizen services. As the platform is envisioned as a broader identity and verification layer beyond aviation, it is positioned not merely as a service but as a foundational component of India’s digital infrastructure. 

A similar framing appears in emerging AI infrastructure initiatives, where state support for access to computing resources and technological capacity is described as enabling “inclusive innovation” and building national AI ecosystems. In both these cases, the language of public good and technological progress shifts attention away from the underlying questions about institutional control, allocation of public resources, private participation, and accountability. What is celebrated as a neutral expansion of digital capacity is therefore also a set of choices about how technological power is organised and governed. These choices are often made in the name of public benefit, while citizens are positioned primarily as recipients of technological systems rather than participants in determining how these systems are designed, governed, and deployed.

The DPI architecture which is routinely celebrated for its digital inclusivity has produced documented patterns of exclusion: marginalised communities are routinely denied welfare entitlements at the enrolment stage due to biometric failures or absence of documentation. These are the predictable consequences of deploying technically complex, privately managed systems on populations with little capacity to contest failures and the most to lose from them. Further, when responsibility is distributed across multiple actors within the digital stack, the affected individuals, many of whom have just been introduced to the digital space, find it hard to identify and hold any single actor accountable. 

Reposing the Question

The two mechanisms this piece has traced, including (a) the retreat from binding regulation, and (b) the encroachment of private actors into sovereign function through the DPI approach, are not policy failures awaiting correction. They are deliberate policy choices that are often presented as inevitable instead of being contested. Understanding their continued acceptance requires examining the structures that enable and normalise them.

Richard Barbrook and Andy Cameron’s introduction of the concept of the ‘Californian Ideology’ describes how the fusion of countercultural bohemianism and authoritarian capitalism produces a worldview in which faith in technological progress and individual empowerment displaces political contestation with technological solutionism. In this worldview, markets and technology are not domains to be governed through democratic deliberation but autonomous evolutionary forces whose progress should be facilitated, making democratic oversight appear unnecessary or even obstructive. In this worldview, the role of the state, therefore, becomes not to regulate but to enable by clearing the path for innovation and subsequently stepping aside. What Barbrook and Cameron identified as the emergence of a ‘virtual class’ of techno-aristocrats who are insulated from accountability through the ideology of meritocratic disruption also finds resonance in an Indian governance culture that often treats market-led AI development as inherently progressive and regulatory intervention as inherently retrogressive.

This technological solutionism is also enabled by what Nathan Schneider has termed innovation amnesia; i.e., the institutional tendency to forget past social arrangements and patterns of market failure to treat each new technological moment as unprecedented, requiring fresh deference to industry. In India, this is structurally embedded in the ‘Innovation over Restraint’ logic across governance documents. It explains why voluntary guidelines are presented not as a choice among regulatory options but as the obviously correct response to technological novelty. It also explains why DPI, despite raising familiar concerns about private control over public infrastructure, is framed not as a governance choice with distributional consequences but as a neutral public good.

The technopolar debate, as framed by Bremmer and echoed in much of the international governance literature, asks how states can reassert sovereignty over technology. India’s experience reveals a slightly newer version of this debate. The problem is not that the Indian state has been displaced by technology corporations. It is that the state is constitutive of the arrangement being critiqued by providing the legal authority, the infrastructure subsidy, and the ideological cover that allows private actors to operate in spaces that are sovereign in name but unaccountable in practice.

Governance frameworks that focus on accountability presuppose a state legible as a regulator, one whose interests are distinct from the private actors it governs, and which can therefore be held to account for regulatory failure. Where the state and private actors construct a mutual legitimation arrangement, an examination of which actors do the governance structures enable and serve becomes important.

Harleen Kaur is a Research Fellow with the Centre for Digitisation, AI and Society at Ashoka University. Her research examines the societal impact of artificial intelligence and emerging technologies, with a particular focus on healthcare, ethics, and governance.