Data Vault, positioning itself as Pakistan’s first sovereign, AI-ready data center and GPU-as-a-Service provider, said the partnership is designed to deliver a hyperscaler-like experience comparable to Azure, AWS, and Google Cloud: self-service access for developers and data science teams, faster provisioning, and managed platform operations—without exporting sensitive data outside the country.
Beyond infrastructure, the companies are pitching the partnership as a catalyst for AI innovation in Pakistan, reducing barriers that slow down experimentation and production deployment. By packaging GPU compute and Kubernetes into a governed, self-service platform, Data Vault says organizations can move from pilots to production faster, and enable more local teams—startups, universities, enterprises, and service providers—to build and deploy AI services inside Pakistan.
NVIDIA has highlighted the broader industry shift toward self-service GPU clouds and “private cloud” experiences for AI teams, arguing that instant provisioning and a platform layer are essential to improve developer productivity and optimize GPU utilization. Data Vault says its sovereign AI cloud is being built on NVIDIA-accelerated computing, with Rafay’s platform providing the orchestration layer that helps deliver cloud-like consumption and governance.
Mehwish Salman Ali, Founder and CEO of Data Vault Pakistan, stated that the partnership aims to make sovereign AI consumption as easy to consume as any hyperscale cloud, while ensuring that sensitive workloads remain securely within the country.
The initiative is designed to deliver a hyperscale-like experience comparable to global services such as Azure, AWS, and Google Cloud, offering self-service access for developers and data science teams, rapid provisioning, and managed platform operations–all without data ever leaving Pakistan, she added.
Key expected outcomes include faster time-to-market for AI use cases in banking, telecommunications, healthcare, manufacturing, and government sectors. The platform also seeks to lower friction for startups and researchers in accessing GPU capacity and modern MLOps and Kubernetes environments.
Haseeb Budhani, CEO of Rafay Systems, emphasized the balance of speed and control, noting, “Companies want AI outcomes quickly without compromising security or control.”
The joint offering will include GPU-as-a-Service and AI-as-a-Service capabilities for training, fine-tuning, and inference, along with managed Kubernetes and platform services covering lifecycle management, security policy enforcement, and observability. It will also provide self-service consumption with built-in guardrails, allowing rapid access for development teams while maintaining centralized governance and cost control. Furthermore, the partnership will enable infrastructure providers to monetize AI resources through multi-tenant service delivery models.
