Nava is building AI infrastructure for enterprises that need GPU compute, orchestration software, and data-center capacity without stitching the whole stack together themselves. The Nava neocloud platform has now raised $22 million, or about ₹204 crore, in a Series A round led by Greenoaks, with RTP Global and Unicorn India Ventures coming back in. AI adoption is running ahead of local compute supply in India and across Asia. Founded in 2025 by Abhinav Sinha, Vamshidhar Reddy, and Abhijeet Singh, the company is trying to sell a fuller answer to that bottleneck than a plain GPU rental business.
What is the Nava neocloud platform and how does it work?
The Nava neocloud platform is a full-stack AI compute system that combines infrastructure, data services, and developer tooling in one product. On its public product architecture, Nava splits that stack into three blocks: AI Factory, Data Platform, and Core Compute. In practice, an enterprise can rent compute and deploy models from one environment. It can also manage data and run inference there instead of bolting together multiple vendors.
The AI Factory layer is built for model serving and training. Nava offers inference as a service through model endpoints that auto-scale and rebalance in real time, along with platform services for deployment and monitoring. Under that sits GPU Fabric and AI-optimized data-center infrastructure. That's the heavy plumbing that determines whether an AI workload runs fast or turns into an expensive mess.
Then there’s the data side. Nava lists parallel file systems, a lakehouse setup, model registry tools, and managed vector databases as part of its platform. AI teams usually don’t just need raw GPUs. They also need storage for large datasets and version control for models. Retrieval systems for production inference matter too. Nava’s pitch is that those pieces should be built together, not added later as afterthoughts.
Before the rebrand, the company described the product as an intent-driven AI-native private cloud where developers set performance, cost, or compliance goals and the software handles deployment and orchestration across hybrid, on-prem, edge, and sovereign environments. The newer Nava site shows that same logic, expanded into a more capital-heavy model that includes data centers and bare infrastructure. It also bakes in zero-trust controls like role-based access, GPU isolation, signed model artifacts, and encrypted vector stores. Enterprise buyers will ask about that before they sign anything.
Who founded Nava and what did they build before this?
The company started as Kluisz.ai
Nava began in 2025 as Kluisz.ai and has now rebranded as it shifts from a software-led GPU cloud product to a broader neocloud strategy. At the same time, the startup has set up Singapore as its regional headquarters so it can sit closer to APAC customers and talent. That’s not cosmetic. The company doesn’t want to stay boxed into India alone.
Why this founding team has real market fit
Abhinav Sinha, Nava’s CEO and co-founder, previously served as Global COO and CPO at OYO and earlier worked at BCG. Vamshidhar Reddy, also a co-founder, is a former McKinsey partner with AMD in his background. Abhijeet Singh, the third co-founder, previously led cloud infrastructure at Jio and earlier worked at AT&T. Between them, Nava has operations experience and direct cloud infrastructure exposure. It also has consulting and systems design depth.
Nava’s leadership page adds a bit more color. Sinha studied at Harvard and IIT Kharagpur. Reddy is listed with Stanford and IIT Kharagpur. Singh is also shown as an IIT Kharagpur alumnus. That doesn’t guarantee execution. But it reinforces the investor story here: this is a founder group built to sell hard technical infrastructure to serious buyers, not just spin up another AI wrapper.
Early traction, fundraising, and the crowded field
The product launched earlier in 2026, and Nava has begun onboarding paying customers, though it still hasn’t disclosed revenue. The company is also in advanced talks around rolling out GPU-based AI infrastructure offerings, and it plans to add senior talent across AI data-center design, GPU engineering, and go-to-market functions. In Singapore, it aims to hire 10 to 15 people by the end of the year.
On the financing side, Greenoaks led this new $22 million Series A, with RTP Global and Unicorn India Ventures participating. Nava’s total funding now stands at $31.6 million, after a $9.6 million seed round in July 2025 that RTP Global led and that also included Unicorn India Ventures, Blume Founders Fund, and Climber Capital. That seed round stood out at the time because it was one of the larger AI startup seed deals in India.
How does Nava compare with Neysa, Yotta, and E2E?
Neysa already sells a full-stack AI cloud system with GPU-as-a-service, orchestration, cost controls, and security. E2E Networks pitches itself as India’s GPU cloud for AI and machine learning. Yotta is pushing Shakti Cloud as a sovereign AI factory that covers training, fine-tuning, and deployment on Indian infrastructure. So Nava isn’t walking into an empty market. Not even close.
Nava is betting on tighter vertical integration and regional reach. Hyperscalers like AWS, Microsoft Azure, and Google Cloud are the obvious default alternatives, but they can be expensive, generic, and awkward for customers that want local deployment, lower latency, or tighter control over compliance and infrastructure choices. Nava’s pitch is that it can sit between those global clouds and old-school enterprise data-center projects. It offers more purpose-built AI features than one, and less complexity than the other. That’s a smart thesis. It’s also brutally hard to execute because infrastructure companies don’t get much room for error.
Why this Nava funding round matters now
A $22 million Series A matters here because Nava isn’t just polishing software anymore. It’s moving into the expensive part of the stack AI-optimized data centers, GPU infrastructure, orchestration, inference layers, and developer tooling. That kind of roadmap needs more than a seed-round budget. Greenoaks leading the round suggests investors think Nava could become more than a niche cloud product.
For customers, the round should mean faster product depth and more local capacity. Nava wants to build its AI compute platform for Asia, not just India, and that matters for companies training models or running inference close to end users. If the company delivers reliable GPU-as-a-service and bare-metal compute with decent developer ergonomics, it could win teams that are tired of juggling fragmented infra contracts.
For the category, this round is another sign that investors are warming up to infrastructure again but only when the story goes deeper than “we rent GPUs.” The core bet on Nava is that software alone won’t be enough, and hardware alone won’t differentiate either. The edge comes from owning both layers in markets where AI demand is rising faster than supply.
Why India needs more AI compute and neocloud capacity
India’s AI data-center push is getting big, fast. The backdrop to Nava’s raise is a broader buildout that the source article pegs at more than $200 billion in planned investment, while India’s AI market is projected to reach $126 billion by 2030 and contribute up to $1.7 trillion to GDP by 2035. That’s the opportunity side of the story. The problem is the infrastructure side still looks thin.
Sinha put the gap pretty bluntly: India has roughly 1 megawatt of compute capacity per million residents, versus more than 100 megawatts in the US. That’s a staggering difference. It helps explain why Indian AI startups and enterprises keep talking about sovereign infrastructure, not just cloud credits. Existing facilities were largely built for conventional cloud workloads. AI training and inference demand denser power, faster networking, and more specialized hardware.
This isn’t just an India story, either. JLL now expects global data-center capacity to climb from 103 GW to 200 GW by 2030, with AI workloads rising from about 25% of total capacity in 2025 to 50% by 2030. Asia-Pacific is part of that surge. So Nava’s timing makes sense.
The real test for the Nava neocloud platform
Nava has a credible founding team, returning investors, a fresh lead in Greenoaks, and a product vision that matches where enterprise AI is heading. That’s the bullish case.
But the Nava neocloud platform now has to prove it can do the ugly part too — secure capacity, ship reliably, price competitively, and stand out in a market where better-funded rivals are already moving. The next thing to watch isn’t another brand refresh. It’s customer adoption, infrastructure rollout, and whether Nava can turn a strong narrative into real compute on the ground.
Read how Astranova Mobility Raises ₹60 Cr for EV Fleet Leasing to expand its electric vehicle leasing and fleet management platform.
FAQ
What funding did Nava raise?
Nava raised $22 million in a Series A round announced in April 2026, led by Greenoaks with participation from RTP Global and Unicorn India Ventures. That brought its total disclosed funding to $31.6 million after a $9.6 million seed round in July 2025.
What does the Nava neocloud platform actually do?
It gives enterprises an integrated AI infrastructure stack instead of just raw GPU rentals. Nava combines inference services and GPU compute in one platform. It also includes Kubernetes and virtual machines, storage, vector databases, model registry tools, and enterprise security controls.
Who are the founders of Nava?
Nava was founded in 2025 by Abhinav Sinha, Vamshidhar Reddy, and Abhijeet Singh. Sinha came from OYO and BCG, Reddy from McKinsey and AMD, and Singh from Jio’s cloud infrastructure team and AT&T.
Is Nava an AI cloud startup or a data center company?
It’s trying to be both. Nava started with a software-led cloud approach under the Kluisz.ai name, and it’s now expanding into a full-stack neocloud model that includes AI-optimized data centers, GPU-as-a-service, bare-metal compute, orchestration, and inferencing software.
