The IFSCA has officially categorized GPUs and data center infrastructure as financial products, enabling new leasing models in GIFT City. This change allows companies to lease rather than buy expensive AI hardware, potentially reducing upfront capital spending. The move aims to attract global investment to boost India's AI infrastructure capacity.
The International Financial Services Centres Authority (IFSCA) has officially categorized Graphics Processing Units (GPUs) and related data center infrastructure as "financial products." This regulatory change, which became effective on October 29, 2026, allows finance companies operating within the GIFT City International Financial Services Centre (IFSC) to engage in operating and hybrid leasing models for high-end AI hardware.
By treating these components as financial assets, the regulator is establishing a clear framework for companies to own, lease-in, or lease-out AI infrastructure. This mechanism mirrors how specialized leasing for aircraft and ships already operates within the GIFT City framework. For organizations building AI data centers, this shift provides a strategic alternative to heavy upfront capital spending. Instead of paying the full cost to purchase expensive AI servers, processors, and network equipment, operators can now enter into leasing arrangements. This helps preserve cash flow and reduces the immediate pressure on balance sheets.
The strategic objective is to lower the barrier to entry for establishing AI-ready compute power in India. As of recent data, India holds a small fraction of global AI-optimized compute capacity. By creating this flexible leasing structure, the government aims to integrate India into the global AI supply chain and attract global firms that require tax-efficient, stable environments to manage their high-value hardware cycles.
While the policy supports capital-intensive AI projects, investors should note the specific operational risks inherent in this model. The most significant challenge is the rapid pace of technological obsolescence. GPU generations evolve quickly, and high-performance chips can lose value or demand as newer, faster alternatives are introduced by global manufacturers. Lessors, or the companies that own and lease out these assets, face the risk that the equipment may become less valuable or harder to re-lease before the end of its useful life.
Furthermore, the sector remains heavily dependent on imports. This exposes the leasing model to supply chain constraints and currency fluctuations, which can impact the cost-efficiency of long-term contracts. The current framework focuses on physical equipment leasing rather than 'GPU-as-a-Service' models, meaning leasing firms bear the responsibility of managing, maintaining, and eventually remarketing the hardware. The next phase of this development will depend on how quickly finance companies develop the technical expertise to manage these specialized assets and whether data center operators find these leasing terms attractive enough to scale their capacity.