Neysa
Neysa provides an India-based sovereign AI acceleration cloud platform (Velocis) offering GPU-as-a-Service, inference, MLOps, and LLM security to enterprises, AI-native startups, research institutions, and government clients across BFSI, insurance, retail, and AI-native segments.
- Company typePrivate
- Founded2023
- HeadquartersMumbai, India
- Headcount101–250
- GTM typeB2B
- OfferingSoftware
What Neysa does
Neysa Networks Private Limited is an India-based AI Acceleration Cloud provider founded in 2023 and headquartered in Mumbai. The company offers the Velocis platform, which bundles GPU-as-a-Service, AI inference, MLOps tooling, and the proprietary Aegis LLM Shield for model security into an integrated sovereign-AI cloud stack. Neysa currently operates 1,200 GPUs with a stated expansion plan to over 20,000 units, and serves approximately 75 customers across BFSI, insurance, eCommerce/retail, manufacturing, education/research, AI-native startups, and government.
The technical foundation rests on inference optimization via vLLM and SGLang with custom CUDA kernels, framework-agnostic AI tooling, and applied research collaborations — most notably with the Indian Institute of Science (IISc), which has produced the O3SLM small language model. The platform is certified to IRDAI, SOC2, and ISO standards and operates in compliance with India's DPDP Act, making it eligible for regulated and data-residency-sensitive workloads. Revenue is generated through quote-based, multi-year enterprise contracts with logos including HDFC Bank, PhonePe, Justdial, BharatGen, and AI-native customers such as Smallest.ai, Navana.ai, and ITQ (the latter reportedly processing ~141B tokens/month of inference volume).
The business model is capex-intensive and scale-driven: prior capital of approximately $44M was deployed into GPU infrastructure, and the February 2026 $1.2B Blackstone-led round is intended to fund the 20,000+ GPU build-out, expansion of owned data centers, and an upcoming AI Marketplace. The combination of sovereign-AI positioning, regulatory certifications, and a vertically templated go-to-market (e.g., the DSW Insurance AI Cloud partnership) targets the structural demand for in-country AI compute from enterprises that cannot or will not route workloads to global hyperscalers.
Neysa firmographics
Firmographics- Name
- Neysa
- Legal name
- Neysa Networks Private Limited
- Website
- https://neysa.ai
- Company type
- Private
- Founded year
- 2023
- Operating status
- Operating
- Headcount range
- 101–250 employees
- Short description
- Neysa provides an India-based sovereign AI acceleration cloud platform (Velocis) offering GPU-as-a-Service, inference, MLOps, and LLM security to enterprises, AI-native startups, research institutions, and government clients across BFSI, insurance, retail, and AI-native segments.
- Ownership category
- akta.pro rank
Where Neysa is headquartered
LocationHeadquarters
- HQ city
- Mumbai
- HQ country
- India
- HQ region
- Asia
Offices3 records
Markets served
Neysa business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Infrastructure, Technology or R&D, Personnel, Marketing or Sales, Operations
Revenue model
- GPU-as-a-Service (GPUaaS): On-demand hourly access to L4, L40S, H100, H100 NVL, H200 SXM, and AMD MI300X GPUs as VMs, containers, or bare metal; plus 1–36 month committed/reserved plans with transparent per-hour billing and no egress or control-plane fees.
- Inference-as-a-Service: Dedicated single-tenant inference endpoints running vLLM or SGLang on reserved monthly GPUs; hybrid pricing against a committed spend pool with scale-to-zero capability.
- AI Platform-as-a-Service (AI PaaS): Managed development environment with Jupyter, PyTorch, TensorFlow, Hugging Face, MLflow, and Kubeflow pre-integrated; sold as a recurring platform subscription.
- Dedicated Private Cluster Agreements: Long-term enterprise contracts providing fully isolated GPU clusters (with air-gapped options for regulated sectors); includes SLA commitments, dedicated support, and professional services for deployment and optimization.
- Professional Services: Optional deployment and optimization services bundled into enterprise contracts — model fine-tuning, kernel engineering, benchmarking, RLHF, and compliance support.
- Startup Launchpad Program: $5,000 in AI Cloud Credits plus technical support, PoC guidance, early feature access, co-marketing, and a dedicated startup success manager for eligible AI-native and digital startups.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Usage-based | Pay-as-you-go | On-demand GPU instances and Inference endpoints |
| Subscription | Monthly | Reserved instance contracts |
| Hybrid | Multi-year contract | Long-term committed plans (1–36 months) |
| Other | Multi-year contract | Dedicated private cluster agreements |
| Freemium | Pay-as-you-go | Startup Launchpad credits |
| Other | Annual | Education & research pricing |
Go-to-market motion6 records
Distribution channels5 records
Marketing channels8 records
Neysa product offering
Product offeringCore offering
Neysa sells the Velocis AI Acceleration Cloud System, a full-stack AI cloud platform combining GPU-as-a-Service on NVIDIA and AMD accelerators, AI Platform-as-a-Service, Inference-as-a-Service, MLOps orchestration, unified monitoring, an AI Catalog, an AI Marketplace, and the Aegis LLM Shield security layer. The platform is sold to enterprises, regulated industries, AI startups, and research institutions that need sovereign, in-country AI infrastructure with predictable economics.
Product overview
Neysa offers a unified AI Acceleration Cloud platform called Velocis, which serves as the core offering and is built on a platform-plus-modules architecture. Velocis integrates GPU-as-a-Service (GPUaaS), AI Platform-as-a-Service (AI PaaS), Inference-as-a-Service, Orchestration & MLOps, Unified Monitoring & Management, Catalog, and AI Infrastructure Security into a single cloud system purpose-built for AI workloads. On top of this core, Neysa offers add-on modules including Aegis LLM Shield (a security layer for LLM endpoints) and the AI Marketplace (for deploying vetted partner apps and agents). The platform also extends through customer programs such as the Neysa Startup Launchpad (AI Velocity Program) for AI-native startups, and is complemented by industry-tailored solutions (e.g., Insurance AI Cloud with DSW, BFSI cloud) and bespoke partnerships like the real-time inference platform co-built with Pipeshift. Together, these named products form a vertically-integrated sovereign AI infrastructure layer for Indian enterprises, startups, and government entities.
Differentiator
Problem solved
Functional benefit
Brands
- Neysa Velocis: Flagship AI Acceleration Cloud System combining GPU infrastructure, AI Platform-as-a-Service, Inference-as-a-Service, orchestration, observability, optimisation, and AI/ML security within a unified enterprise AI environment.
- Aegis LLM Shield
Products and services
- Neysa Velocis AI Acceleration Cloud Full-stack AI Acceleration Cloud System that unifies the entire AI workflow from training to deployment, combining AI infrastructure, inference, orchestration, observability, optimisation, and AI/ML security within a single cloud environment purpose-built for AI workloads and sold to Indian enterprises, regulated industries, AI startups, and research institutions.
- GPU-as-a-Service (GPUaaS) Instant on-demand access to NVIDIA L4, L40S, H100, H100 NVL, H100 SXM, H200 SXM, and AMD MI300X GPUs available as VMs, bare metal, or containers with 3200 Gbps interconnects, NVMe-backed storage, and on-demand, reserved, or 1–36 month commit pricing with up to 40–60% lower unit economics than hyperscalers.
- AI Platform-as-a-Service (AI PaaS) Managed AI development environment providing one-click dev environments (Jupyter, PyTorch, TensorFlow, Hugging Face), MLflow and Kubeflow integrations, CI/CD for models, and Git integration for moving from training to production with zero setup effort.
- Inference-as-a-Service Real-time inference endpoints with custom CUDA kernels and engine selection between vLLM and SGLang, providing single-tenant endpoints with dedicated SLAs, hybrid pricing against committed spend pool, and OpenAI-compatible endpoints for open-source and custom models.
- Aegis LLM Shield Security add-on integrated into the Velocis Developer Hub that enforces security policy at every LLM endpoint, blocking prompt injection and jailbreaks, redacting PII, and providing one-click compliance for DPDPA, GDPR, and CCPA with built-in safeguards for HIPAA and PCI-DSS.
- Neysa Startup Launchpad (AI Velocity Program) Exclusive program for eligible AI-native and digital startups offering $5,000 in AI Cloud Credits, technical support and PoC guidance, early feature access, co-marketing, ecosystem amplification, and a dedicated Startup Success Manager.
Quantifiable outcome
- 40% reduction in total cost of ownership vs general-purpose cloud GPUs (ITQ case study)
- +8 more outcomes
Companies that use Neysa
Customer profileNamed customers27 records
Segments7 records
Ideal customer profiles7 records
Neysa technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration38 records
Feature11 records
Neysa partnerships and signals
Strategic signalPartnerships
Six partnerships are on record, tiered flagship and minor.
- PipeshiftflagshipStrategic partnership announced 27 May 2026 to launch production-grade real-time inference infrastructure fully deployed within India. Pipeshift's managed inference platform (MAGIC containers) runs on Neysa's AI-native GPU infrastructure, enabling single-tenant inference endpoints for open-source models (Gemma, Qwen, GPT-OSS, Llama, DeepSeek, Mistral) via OpenAI-compatible APIs. The platform is tuned at the kernel and inference-engine level for production traffic and is targeted at voice AI, enterprise copilots, customer support, and regulated AI workloads. Early deployments include Nurix AI (3x TTFT reduction) and Arrowhead AI (multilingual inference with sub-day deployment).
- Data Science Wizards (DSW)minorCo-development partnership to launch an advanced Insurance AI Cloud platform for the Indian insurance sector, combining DSW's insurance domain AI capabilities with Neysa's Velocis infrastructure.
- Indian Institute of Science (IISc) BangaloreminorResearch partnership in which IISc's Visual Computing Lab ran end-to-end training of the O3SLM sketch vision models (7B and 13B) on Neysa Velocis — including 33 million sketch-image synthetic data generation, two-stage instruction tuning, and four-benchmark evaluation. The models were accepted at AAAI 2026 and outperformed GPT-4o and Gemini 1.5 Pro on four sketch benchmarks. IISc is publicly listed as a Neysa customer logo on neysa.ai.
- DC AdvisoryminorDC Advisory advised Neysa.AI on the $1 billion+ capital raise led by Blackstone and co-investors, supporting Neysa's deployment of over 20,000 GPUs in India.
- TT&A / Trilegal / Gibson Dunn & CrutcherminorMultiple law firms including TT&A, Trilegal, and Gibson Dunn & Crutcher advised on the Blackstone-led $1.2 billion Neysa investment transaction in February 2026.
- Anup PurohitminorVeteran technology executive (Wipro, YES Bank, RBL Bank, Barclays, JPMorgan) appointed as Strategic Advisor on 11 June 2026; guides expansion of Velocis and provides strategic counsel to enterprises integrating AI.
Scale indicators12 records
Recent moves7 records
Expansion highlights7 records
Neysa competitors and assessment
Company assessmentDirect peers
- CoreWeave: US-based GPU cloud specialist offering NVIDIA H100/H200 bare-metal and virtualised compute with Kubernetes-native orchestration. Closest functional analogue to Neysa Velocis in scale, technology stack (GPUaaS + Inference + MLOps), and focus on AI workloads.
- Lambda Labs: GPU cloud and deep-learning workstation provider running NVIDIA H100/H200 fleets with on-demand and reserved pricing. Comparable to Neysa on GPU-as-a-Service economics and AI-native customer base, though Lambda is primarily US-focused.
- Yotta Data Services: Indian data-centre operator offering NVIDIA H100 GPU cloud, sovereign AI infrastructure, and partnerships with NVIDIA and the IndiaAI Mission. Most direct India-headquartered competitor to Neysa on sovereign AI infrastructure and BFSI/government workloads.
- E2E Networks: India-listed GPU cloud provider offering NVIDIA H100/H200 instances, AI/ML platform services, and the country's first IRDAI-aligned AI cloud. Competes head-to-head with Neysa on price-conscious Indian AI workloads and sovereign deployment.
- Together AI: AI cloud platform providing dedicated GPU clusters, optimised inference for open-source LLMs (vLLM/SGLang), and fine-tuning infrastructure. Closely comparable to Neysa's Inference-as-a-Service and Pipeshift-integrated real-time inference layer.
- CtrlS Datacenters: India's largest hyperscale data-centre operator now adding GPU-cloud and sovereign-AI offerings. Competes with Neysa on private-cluster deals with BFSI, government, and large enterprise customers seeking in-country AI infrastructure.
Broad incumbents
- Reliance Jio Cloud: Indian hyperscaler offering Jio AI Cloud, sovereign compute, and bundled AI services targeting BFSI and consumer apps. A broader incumbent than Neysa — overlapping on AI infrastructure and data-residency but not specialised in GPUaaS/inference.
- Amazon Web Services (AWS): Global hyperscaler with EC2 GPU instances, SageMaker, Bedrock and growing India-region presence. Competes with Neysa on GPU capacity, MLOps and managed inference, but as part of a much broader cloud portfolio rather than AI-specialised focus.
Others
- Microsoft Azure: Global hyperscaler offering Azure ML, Azure OpenAI Service and Azure ND-series GPU VMs with India data-residency. Direct competitor to Neysa on enterprise AI workloads, particularly cited by customers (GreyLabs, Navana.ai) as the legacy alternative Neysa is displacing.
Emerging players
- RunPod: Developer-focused GPU cloud offering on-demand H100/A100 instances and serverless inference for AI startups. Comparable to Neysa's Startup Launchpad and self-serve motion, with overlapping AI-native startup customer base but no sovereign/India specialisation.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks6 records
Key highlights1 record
Customer concentration
Neysa social profiles
Digital presenceNeysa financial estimates
Financial estimateRevenue estimate
Valuation estimate
Neysa leadership team
Management profileNumber of profiles
Profiles5 records
Neysa funding detail
Funding detailFunding overview
Funding rounds3 records
Investors9 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Neysa M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Neysa
What does Neysa do?
Neysa sells the Velocis AI Acceleration Cloud System, a full-stack AI cloud platform combining GPU-as-a-Service on NVIDIA and AMD accelerators, AI Platform-as-a-Service, Inference-as-a-Service, MLOps orchestration, unified monitoring, an AI Catalog, an AI Marketplace, and the Aegis LLM Shield security layer. The platform is sold to enterprises, regulated industries, AI startups, and research institutions that need sovereign, in-country AI infrastructure with predictable economics.
Is Neysa a public or private company?
Neysa is a private company. It is classified as private equity controlled and is currently operating.
When was Neysa founded?
Neysa was founded in 2023. It employs 101 to 250 people.
Where is Neysa based?
Neysa is headquartered in Mumbai, India, in the Asia region.
How does Neysa make money?
Six revenue lines are on record. GPU-as-a-Service (GPUaaS) is the primary driver. The others are inference-as-a-Service, AI Platform-as-a-Service (AI PaaS), dedicated Private Cluster Agreements, professional Services and startup Launchpad Program.
Who are Neysa's main competitors?
Direct peers on record are CoreWeave, Lambda Labs, Yotta Data Services, E2E Networks, Together AI and CtrlS Datacenters. Broad incumbents are Reliance Jio Cloud and Amazon Web Services (AWS). Microsoft Azure is listed as an others. RunPod is listed as an emerging player.
Does Neysa have an API?
Yes. Neysa Velocis offers OpenAI-compatible inference endpoints with custom CUDA kernels and engine selection (vLLM, SGLang). The platform is API-first and extensible, exposing core functions via secure APIs for integration with CI/CD pipelines, monitoring stacks, IDEs, and existing ML workflows. Supports HuggingFace checkpoints, custom fine-tuned weights, custom containers, and Helm charts. Deployment includes single-tenant endpoints with dedicated SLAs and vLLM tensor parallelism across NVIDIA GPU clusters.