Neysa

Neysa is India's first fully AI-focused cloud platform - a "neocloud" founded in 2023 and headquartered in Mumbai. Co-founded by Sharad Sanghi (who previously built Netmagic into India's largest data center company before its NTT acquisition) and Anindya Das, Neysa is designed from the ground up for AI and machine learning workloads, not retrofitted from a general-purpose cloud.

Its flagship product, Neysa Velocis, is a full-stack AI Acceleration Cloud System combining GPU-as-a-Service, AI Platform-as-a-Service, Inference-as-a-Service, Orchestration & MLOps, and the proprietary Aegis LLM Shield - a security layer purpose-built for enterprise AI governance. The platform supports fractional GPU billing, spot-aware scheduling, per-minute billing, and pre-configured AI stacks including PyTorch, TensorFlow, and Hugging Face.

In February 2026, Neysa closed a landmark $1.2 billion financing round led by Blackstone, with co-investors Teachers' Venture Growth, TVS Capital, 360 ONE Assets, and Nexus Venture Partners. The capital is earmarked for scaling GPU capacity to over 20,000 units across Indian data centers - positioning Neysa as India's largest dedicated AI cloud.

Neysa targets enterprises, regulated sectors (BFSI, healthcare, government), AI-native startups, and research institutions requiring a combination of high-performance compute, enterprise SLAs, Indian data sovereignty, and managed AI tooling - a higher-service tier compared to self-serve GPU clouds.

• Full-stack AI platform (Velocis) - not just raw GPU rental; includes MLOps, observability, model catalog, and inference endpoints
• Fractional GPU support - share a GPU for smaller workloads, pay only for what you use down to per-minute billing
• H200, H100, L40S, A100, L4, A30 all available - from lightweight inference to large-scale LLM training
• Aegis LLM Shield - built-in firewall for LLM governance, a unique feature not offered by other Indian providers
• 15-minute support response SLA - enterprise-grade hand-holding unmatched by hyperscalers or self-serve clouds
• Pre-configured AI stacks - PyTorch, TensorFlow, Hugging Face, vLLM, Stable Diffusion - ready in minutes, not hours
• Three deployment models: public cloud, private cloud, and hybrid - suitable for any compliance posture
• Full Indian data sovereignty - DPDP Act compliant, suitable for BFSI, healthcare, and government workloads
• $1.2B Blackstone-backed - serious capital for GPU capacity expansion; 20,000+ GPUs planned
• Inference-as-a-Service - deploy popular open-source models (Llama, DeepSeek, Mistral, Qwen) as managed endpoints
• Spot-aware scheduling - intelligently routes interruptible jobs to lower-cost capacity automatically
• No egress, no hidden I/O, no surprise API call charges - transparent pricing model

Neysa positions itself as an enterprise/managed AI cloud - pricing is customized for many configurations. The table below shows representative on-demand rates. Contact sales for reserved, private cluster, and volume pricing.

Deployment Models

Managed Kubernetes (VKE)

Neysa plans to deploy 20,000+ GPUs across Indian data centers as part of its $1.2B Blackstone-backed expansion. International expansion announced but timelines not yet confirmed.