Gujarat Launches Shared GPU Computing Platform to Accelerate AI Projects

Gujarat is making high-performance AI computing more accessible to government departments, educational institutions and selected startups with the launch of a statewide GPU-as-a-Service (GPUaaS) framework.

The initiative creates a shared pool of GPU computing infrastructure that eligible organizations can use for approved AI and data-driven projects, rather than requiring every department or institution to build expensive computing infrastructure of its own.

The framework is effective immediately and forms part of the infrastructure pillar of Gujarat's broader AI Action Plan.

What Has Gujarat Launched?

Under the new framework, Gujarat Informatics Limited (GIL) will manage centralized GPU infrastructure and handle provisioning, operations, monitoring and administration.

GPU resources will be allocated based on factors including:

• Technical feasibility

• Available computing capacity

• Approved AI use cases

• Priority of requirements

• Applicable government policies

The idea is relatively straightforward: create a common computing layer that different government organizations and approved AI projects can access when they need significant processing power.

For AI development, that can be important.

Training, fine-tuning and running modern AI models can require expensive GPU infrastructure that smaller organizations and startups may struggle to access independently.

Government and Educational Institutions Get Fully Subsidized Access

One of the biggest parts of the framework is its subsidy model.

Eligible Gujarat government departments, government institutions and government educational institutions will receive 100% subsidized access to the GPU infrastructure.

Startups selected through the AI Innovation Challenge at Gujarat's AI Centre of Excellence (AI CoE) will also receive fully subsidized GPU access for production rollout of their selected AI use cases.

Other startups may be allowed to use the infrastructure depending on resource availability and the conditions of the framework.

However, those companies will have to pay the actual GPU service cost along with the applicable GIL consultancy fee.

The framework provides for GIL to receive a 7% consultancy fee on the actual GPU service cost procured from cloud service providers empanelled under IndiaAI.

Why Shared GPUs Matter for AI Development

GPUs have become one of the most important pieces of infrastructure behind modern AI.

Building AI models, training computer vision systems, fine-tuning language models and running large-scale inference workloads can require significant computing power.

That creates a barrier.

An organization may have a useful AI idea, data and engineering talent but still struggle to move from a proof of concept to production because of infrastructure costs.

Shared GPU infrastructure changes some of that equation.

Instead of every government department purchasing and maintaining its own hardware, compute capacity can be allocated centrally based on actual project requirements.

For startups, access to shared infrastructure could also reduce some of the upfront cost of testing and deploying compute-intensive AI applications.

Gujarat Is Building More Than Just GPU Infrastructure

The GPUaaS framework fits into Gujarat's broader AI strategy.

The state's AI Action Plan is structured around six areas:

Data, Infrastructure, Capacity Building, Foundational R&D and Use Case Development, Deeptech Startup Facilitation, and Safe & Trusted AI.

Gujarat has also established an AI Centre of Excellence at GIFT City, Gandhinagar, which was launched in January 2025.

The centre is intended to support AI adoption across government and the wider innovation ecosystem, including developing proofs of concept, training and fine-tuning AI models, supporting startups and deploying successful AI projects into production.

The state's AI Innovation Challenge is another part of that ecosystem, connecting government problem statements with startups capable of developing practical AI solutions.

Security and Responsible AI Are Part of the Framework

Access to compute does not mean unrestricted use.

Organizations receiving GPU resources will be required to use them for approved purposes and comply with applicable policies covering:

• Information security

• Data privacy

• Cybersecurity

• Responsible AI

GIL is expected to issue detailed guidelines covering eligibility, applications, allocation, pricing, operational requirements and monitoring.

That governance layer will become increasingly important as government agencies begin moving AI projects beyond experimentation and into production systems.

Gujarat's Move Reflects a Bigger AI Infrastructure Shift in India

Gujarat's initiative also fits into a larger national effort to expand access to AI compute.

Under the IndiaAI Mission, India's shared AI compute capacity had grown to more than 45,000 GPUs as of June 2026.

By August 2026, 237 projects had accessed subsidized AI computing resources, representing approximately 9.3 million GPU hours.

The broader goal is to reduce the infrastructure barrier for Indian startups, researchers and organizations developing AI models and applications.

Gujarat is now creating a state-level mechanism around a similar idea: AI compute as shared infrastructure rather than infrastructure every organization needs to build independently.

Why This Matters

There is a lot of discussion about building AI models, launching AI startups and adopting AI inside government.

But all of those ambitions eventually depend on infrastructure.

Making GPU capacity easier to access could help more organizations move from an AI idea to an actual proof of concept, and from a proof of concept to a production deployment.

For Gujarat, the interesting part will now be what gets built on top of this infrastructure.

If the GPU pool results in useful government applications, stronger university research and more opportunities for local AI startups, it could become an important piece of the state's emerging AI ecosystem.

The infrastructure is now being put in place.

The next question is what Gujarat's developers, startups and institutions build with it.