AI infrastructure startup Nava has raised $22 million in a funding round led by Greenoaks Capital, according to Tech-Economic Times. The financing included participation from RTP Global and Unicorn India Ventures. The company will use the capital to expand its GPU compute and AI data centre capabilities and hire talent. Nava is expanding beyond its earlier software-led GPU cloud offering toward a vertically integrated model, with infrastructure offerings aimed at enterprises building AI models and applications.
Funding Round Details
The $22 million round reflects investor interest in AI infrastructure providers. The stated use of funds is specific: expand GPU compute and AI data centre capabilities and hire talent. In practical terms, this points to two linked areas of execution: scaling the underlying hardware and data centre operations that support accelerated workloads, and building the technical teams that can operate and optimize those environments.
The investment is capacity-driven, addressing a core constraint in AI infrastructure: availability of accelerated compute resources. AI model development and deployment cycles can be limited by GPU capacity availability. If Nava’s data centre expansion aligns with its compute expansion, it could reduce friction for customers who need GPU capacity for training and application workloads in the regions and configurations Nava supports.
Shift to Vertically Integrated Infrastructure
Nava is expanding beyond its earlier software-led GPU cloud offering to a vertically integrated model. This signals a change in how the company intends to control the stack around GPU compute. A software-led model typically emphasizes orchestration, provisioning, and management layers while relying on external hardware supply. A vertically integrated approach suggests the company is moving closer to owning or directly managing more of the underlying infrastructure needed to deliver GPU compute services.
The shift is connected to Nava’s planned expansion of AI data centre capabilities and GPU compute. This combination suggests the company is aligning its business model with the operational requirements of running accelerated workloads: data centre capacity, hardware availability, and the platform layers that expose that capacity to customers.
Service Offerings and Target Market
Nava targets enterprises building AI models and applications. The company offers infrastructure through two models: GPU-as-a-service and bare-metal compute.
GPU-as-a-service is a managed model where customers access GPU resources through a service interface rather than directly provisioning hardware themselves. Bare-metal compute allows customers to run workloads on physical servers without virtualization abstraction layers. The combination of both service types suggests Nava aims to serve multiple deployment preferences—ranging from teams that prefer managed access to teams that require direct control over compute environments.
These service types can influence engineering decisions. GPU-as-a-service can simplify scaling and operational management, while bare-metal compute can be important for workloads requiring specific performance characteristics or environment control. The availability of both options indicates Nava is positioning itself to address different customer needs.
Market Implications
In AI infrastructure, capacity and delivery models determine which workloads can be served reliably. Nava’s plan to use new funding to expand GPU compute and AI data centre capabilities while hiring suggests it is investing in the operational foundation required to serve enterprise AI demand. The company’s vertically integrated direction could potentially translate into faster provisioning, more consistent availability, or improved alignment between customer needs and underlying hardware.
The funding round’s leadership and participation—Greenoaks Capital, RTP Global, and Unicorn India Ventures—signals continued market interest in platforms that deliver accelerated compute. Nava’s move toward vertically integrated infrastructure could indicate a broader industry pattern: providers may seek more control over the hardware and data centre layers that support AI workloads. This strategy could strengthen a provider’s ability to support enterprise pipelines for AI model and application development.
Source: Tech-Economic Times