India's Datacentre Boom Is a Sovereign Infrastructure Moment.

In February 2026, India hosted its most significant AI policy moment to date. The Global IndiaAI Summit convened policymakers, technologists, researchers, and industry leaders to discuss India's trajectory as a leading AI nation. The conversation was ambitious, appropriately so. India has the scale of digital infrastructure, the engineering talent, the sovereign data assets through programmes like Aadhaar and UPI, and now the fiscal commitment — through Budget 2026–27's landmark tax holiday until 2047 for foreign companies providing cloud services from Indian data centres — to become a genuine global node in AI infrastructure.

The numbers behind this moment are extraordinary. India's installed data centre capacity has tripled since 2019. Combined hyperscaler and domestic DC investment commitments are projected in the range of US $200 billion. Maharashtra, where the CII Datacenter Blueprint Summit convened last month, is the country's largest and fastest-growing data centre market — with Navi Mumbai, the BKC corridor, and emerging Pune clusters anchoring a physical infrastructure story that is, by any measure, world-class.

And yet. The harder question — the one that the investment story, the capacity charts, and the fiscal incentives do not answer — is this:

This paper poses that question directly — drawing on India's own governance architecture, on what Europe has learned from a decade of building sovereign digital infrastructure, and on the IndiaAIStack IS-1 to IS-9 Sovereignty Capability Framework.

IndiaAIStack brings complementary practitioner lenses to this question. The initiative works on the governance architecture and EU policy dimensions — what the frameworks, assurance mechanisms, and control plane sovereignty questions look like in practice — and also on the supply chain, procurement, and physical infrastructure dimensions — what sovereign delivery actually requires at the level of hardware sourcing, site selection, and procurement continuity. Together, the IS Framework reflects both perspectives.

India's recent AI Summit and Budget 2026–27 have generated a productive infrastructure conversation. The tax holiday to 2047 is a bold fiscal signal. Hyperscaler investment commitments are real. Domestic DC operators are building at meaningful scale. IndiaAI Mission is creating the compute access architecture. Beneath the investment story, key governance questions should receive the analytical attention they require. They are capability architecture questions — and they will determine whether the infrastructure India is building in 2026 constitutes genuine sovereign capability or expensive dependency.

Question 1: Who controls the control plane?

India's Budget 2047 tax holiday incentivises foreign cloud companies to build data centres in India. This is a legitimate strategy for anchoring physical infrastructure within Indian jurisdiction. Indian DC majors — CtrlS, Yotta, STT GDC, AdaniConneX — will own the land, the power contracts, the buildings. The control plane — the management interfaces, hypervisor access, network routing decisions, configuration authority, and audit logging — is the key.

Question 2: Data residency is not data sovereignty. Is India conflating the two?

India's DPDP Act 2023 and sectoral regulators (RBI, IRDAI, DoT) mandate data localisation across critical sectors. This is necessary. But localisation — requiring data to be physically stored within Indian borders — is not sufficient for sovereignty. India's legal sovereignty framework must address foreign legal instruments such as the US Cloud Act. India's DPDP Act gives the government strategic control over cross-border transfer whitelists. But without infrastructure audit rights that verify the legal chain of custody from physical storage through logical access, localisation laws remain partially unenforceable.

Question 3: India's AI compute is growing. But who governs access to it?

India will, in the next five years, have more AI compute available within its borders than at any point in its history. The governance question that this capacity growth does not answer is: who gets access to it, under what terms, with what assurance of continuity? Sovereignty requires a governance model for compute access, not just a compute procurement budget.

Question 4: India's supply chain for AI infrastructure — how sovereign is it?

Every data centre being built in India depends on a hardware supply chain that is, for critical components, concentrated in geographies that India does not control. GPU accelerators, high-bandwidth memory, advanced cooling systems, structured cabling, and precision power infrastructure are sourced from supply chains that run through the United States, Taiwan, Japan, and South Korea.

India's PLI programs and semiconductor strategy are building long-term design and packaging capability. But the near-term supply chain for the infrastructure being built under Budget 2026–27 incentives is not sovereign. An export control decision, a logistics disruption, or a geopolitical event affecting Taiwan Strait shipping lanes has direct consequences for India's ability to maintain and expand its AI infrastructure. The physical irreversibility of data centres as assets makes supply chain planning a sovereign obligation, not merely an operational procurement matter.

Question 5: India's AI infrastructure investment is large. But is it sufficiently governed?

US $200 billion in projected AI and cloud infrastructure investment is a national asset. But the governance architecture surrounding that investment — procurement assurance, accountability frameworks, public interest obligations, sovereignty compliance verification — is not commensurate with the scale of investment.

MeitY's empanelment framework approves cloud providers through a process-based evaluation. It does not assess sovereignty capability systematically. There is no equivalent of the EU's Cloud Sovereignty Effective Assurance Level (SEAL) — which gives government procurers a structured tool for comparing how 'sovereign' a cloud solution actually is across defined parameters. Without equivalent tools, India's public sector AI procurement is making billion-dollar infrastructure decisions with an incomplete governance framework.

Europe's digital sovereignty journey is instructive not because India should replicate it — the two contexts are fundamentally different — but because Europe has spent a decade and tens of billions of euros learning which governance instruments actually work, which are aspirational rhetoric, and which create genuine sovereign capability. India can compress that learning curve.

The EU's digital sovereignty ecosystem — built around GAIA-X, EuroHPC, the Digital Europe Programme, SIMPL, and EuroStack — rests on a core strategic philosophy that differs from India's emerging approach in one fundamental way:

Neither model is superior. But Europe's model is more advanced in the governance architecture layer — precisely the layer where India's gap is greatest.

The IndiaAIStack Indian Sovereign (IS) Framework™ is a practitioner-built diagnostic tool for assessing India's sovereign AI infrastructure capability across nine parameters. It is not a policy proposal. It is an analytical instrument designed to make the governance gap conversation precise — to move it from abstraction to assessment.

Each parameter names a distinct dimension of sovereignty that India's infrastructure conversation tends either to conflate with others or to omit entirely. Together, they provide a 9-parameter checklist against which any AI infrastructure program — a data centre, a cloud procurement, an AI compute allocation — can be evaluated for its Sovereignty Assurance level.

The framework is designed to be applied at the program level — to a specific data centre procurement, a specific cloud service contract, a specific AI compute allocation. Its primary value is as a governance diagnostic: when India's decision-makers evaluate a proposed infrastructure investment, the IS parameters provide the questions that commercial procurement criteria will not surface.