At least six global neocloud companies backed by Nvidia are in discussions with Indian data centre operators to lease compute capacity for artificial intelligence inference workloads, marking a fresh signal that India is moving from being primarily a large AI consumption market to a possible deployment base for AI infrastructure.
The companies in discussions include CoreWeave, Together AI, Nebius, Lambda Labs, Crusoe and Nscale. The Indian and India-linked operators named in the discussions include Sify Technologies, Yotta Data Services, CapitaLand Investment, CtrlS Datacentres and Tata Communications. The talks are at varying stages and centre on leasing capacity rather than necessarily building owned facilities, a structure that would allow neocloud providers to expand faster in a market where demand for AI workloads is rising but power, land, GPU supply and execution timelines remain binding constraints.
The development sits inside a broader shift in global AI infrastructure strategy. Neocloud companies have grown rapidly by renting accelerated computing capacity to AI labs, enterprises and developers that need GPU access without directly owning large-scale data centre assets. India’s appeal is tied to the scale of its digital economy, the growth of enterprise AI pilots moving toward production, and regulatory as well as commercial interest in keeping more workloads closer to domestic users and data.
For Indian operators, the discussions could broaden demand beyond traditional enterprise hosting, cloud availability zones and hyperscaler commitments. AI inference workloads differ from training-heavy deployments: they require low-latency serving, predictable uptime, efficient cooling and scalable GPU-backed clusters that can be consumed by multiple customers. If leases materialise, Indian facilities would need to support dense power loads, specialized cooling and networking architectures suited to accelerated computing.
The talks also come as India’s data centre market is drawing capital from domestic conglomerates, global infrastructure investors and cloud providers. Operators such as Yotta, CtrlS, Sify and Tata Communications already sit within a competitive field shaped by AI, cloud migration, data localisation, digital public infrastructure and enterprise demand from banking, telecom, manufacturing, healthcare and public-sector use cases. CapitaLand’s presence reflects the continued entry of global real-asset platforms into India’s digital infrastructure market.
The immediate transaction structure remains to be finalized. The current development is best read as a capacity-sourcing move: global AI infrastructure providers are assessing whether Indian operators can support workloads at the scale, reliability and cost profile required for commercial inference. If executed, it would add another layer to India’s AI stack, linking domestic data centre capacity with global GPU-cloud platforms and enterprise AI demand.
