New Delhi: Enterprises and startups looking to run AI workloads inside India will have a new home for their compute from early 2027, after data centre operator STTGDC said on Wednesday that Neysa, an India-based AI cloud company, has chosen its Noida campus for a large GPU deployment.
The facility, STT Noida 2, will support up to 9,000 GPUs and run Neysa's Velocis AI cloud platform. The first phase is expected to go live in the first quarter of calendar year 2027 and is built for high-density AI training and inference, the two companies said.
The pitch is straightforward: production AI workloads that need predictable latency and data residency can run close to where the users and data sit, rather than being shipped abroad.
Bimal Khandelwal, CEO (India) of STTGDC, said India's AI ambitions would depend on the digital infrastructure needed to support increasingly complex, compute-intensive workloads at scale, as much as on access to compute. Bringing Neysa's specialised AI capability together with STTGDC's infrastructure would help enterprises 'move from experimentation to real-world applications', he said.
Sharad Sanghi, CEO of Neysa, said the company was building for the long term, for Indian businesses and global compute customers. He said customers want capacity they can plan around, reliable operations and room to scale as their workloads grow.
The announcement lands in a week of heavy activity in India's AI infrastructure space. Earlier, Submer India and BSNL Tower Corporation Limited signed a memorandum of understanding at the India Mobile Congress 2026 to explore sovereign, AI-ready data centres, including GPU-as-a-Service and facilities of up to 20 MW.
For Indian enterprises, the Noida deployment is a test of whether homegrown AI capacity can keep pace with demand.