Economic Survey warns GPU shortages could delay India’s data centre and AI expansion plans

At a time when the demand for compute is rising sharply with the rapid adoption of artificial intelligence across sectors, the Economic Survey flags the cost of compute as a growing bottleneck for AI expansion.

The Survey points to surging demand for key inputs, driven by higher GPU requirements, alongside constrained supply conditions caused by shortages of high-bandwidth memory chips and storage.

According to the Economic Survey, this combination is pushing up costs and is bound to have a ripple effect on the expansion of compute capacity in India, potentially turning financial viability into a binding constraint.

The Survey notes that higher demand from overseas buyers can choke the available supply of GPUs for domestic players. It adds that even if the sovereign, domestic investors or financial institutions are willing to finance data centre expansion, projects may need to be deferred until GPU supplies are secured.

Stress tests cited in the Economic Survey reveal that across scenarios, developments in global GPU supply chains have a direct bearing on AI infrastructure expansion.

The Survey underlines that capacity growth is influenced by the structure and concentration of global hardware supply chains, making a strong case for policy measures that enhance supply-side resilience in access to advanced compute.

The Government of India is boosting AI infrastructure through the IndiaAI Mission by securing high-performance GPUs. Companies like Jio Platforms, NxtGen, E2E Networks have already offered GPUs, while Nvidia has committed to a major role, providing chips such as H100, H200 and Blackwell models.

These GPUs are being made available via the IndiaAI Compute Portal at subsidised rates for startups, researchers and academic institutions. The government is also including GPUs from AMD and Intel to ensure a diverse, reliable supply.

A late-mover advantage in AI strategy

Against this backdrop, the Economic Survey argues that India’s late entry into the AI race could work to its advantage. The Survey observes that early adopters scaled AI when capital was cheap and regulation limited, locking themselves into energy-intensive architectures, rising financial commitments and uncertain revenue models.

As these investments expanded, the Economic Survey notes, risks also increased, with advanced economies now debating government backstops to contain potential financial fallout.

The Survey argues that India does not need to follow this path. Its comparative advantage, according to the Economic Survey, lies not in building frontier-scale models but in application-led innovation, the productive use of domestic data, strong human capital and the coordinating capacity of public institutions.

A bottom-up approach built on open and interoperable systems, sector-specific models and shared physical and digital infrastructure, the Survey says, offers a more credible route to value creation.

This allows India, the Economic Survey concludes, to design AI systems that are more resource-efficient and aligned with public objectives, while avoiding costly and hard-to-reverse dependencies.