As India accelerates toward a formal regulatory framework for Artificial Intelligence, the Ministry of Electronics and Information Technology (MeitY) has taken a decisive step to secure the country’s digital future. Under the aegis of the ambitious IndiaAI Mission, the government has officially initiated the process of empaneling AI Service Providers to supply a massive infrastructure backbone: 10,000 AI Compute Units (AICU).
This move, executed through the Centre for Development of Advanced Computing (C-DAC), represents a strategic pivot from mere policy-making to active infrastructure provisioning. By centralizing the procurement of GPUs, TPUs, and specialized accelerators, the government aims to democratize access to high-performance computing (HPC) for the country’s research institutions, startups, and public sector departments.
The Core Mandate: Scaling India’s AI Compute
The Request for Empanelment (RFE) issued by C-DAC is not merely a procurement document; it is a blueprint for India’s sovereign AI capability. C-DAC has been tasked with building a shared, secure, and scalable ecosystem to support the next generation of AI-driven applications.
Defining the AICU
The RFE categorizes AI Computing Units (AICUs) as GPUs, TPUs, and other high-performance hardware accelerators essential for training and deploying large-scale AI models. The target is an initial capacity of 10,000 units, with a clear provision for expansion. As noted in the tender document, the government reserves the right to scale this capacity significantly beyond the initial 10,000 threshold based on the evolving demands of academia, industry, and governmental bodies.
Cluster Architecture
To ensure resilience and efficiency, the infrastructure is to be organized into modular clusters of 4,000, 1,000, or 512 units. The RFE mandates that each cluster must consist of at least 512 AICUs housed within a single data center located in India. This geographical distribution requirement is designed to ensure that the infrastructure is not only robust but also localized, mitigating latency issues and adhering to national data sovereignty norms.
Chronology: A Rapid Push for AI Supremacy
The recent RFE is the latest in a series of fast-tracked initiatives by MeitY to position India as a global AI hub.
- August 2026: MeitY issues a landmark tender inviting AI/ML professionals to form a "Unified AI Team." This team is tasked with embedding AI capabilities into essential government services, including the UMANG and DigiLocker platforms, signaling a transition toward AI-integrated public service delivery.
- October 7, 2026: C-DAC releases the RFE for AI Service Providers, effectively inviting private players to help build the underlying hardware infrastructure required to support the government’s AI goals.
- October 8, 2026: Union IT Minister Ashwini Vaishnaw addresses the World Development Report 2026, confirming the government’s demand estimate of 10,000 GPUs. He clarifies that the initial phase of procurement would target 5,000 units, emphasizing an "empowerment-based" approach rather than traditional, rigid procurement models.
- Late October 2026 (Forthcoming): MeitY is expected to release a comprehensive consultation paper on AI regulations, which will define the ethical and legal boundaries for the use of the compute power being procured today.
The "Empowerment" Philosophy: A Shift in Governance
One of the most significant aspects of the current strategy is the government’s departure from traditional tendering processes. Minister Ashwini Vaishnaw has explicitly stated that the government will adopt an "empowerment approach."
In practice, this means moving away from the "lowest-bidder" (L1) model that often stifles innovation in technology procurement. By creating a panel of service providers, MeitY intends to ensure that the government has access to the latest, most efficient hardware without being locked into inflexible contracts. This allows for a more agile system where the government can tap into the best available computing power as the market evolves.
Data Sovereignty and Security: The Non-Negotiables
The RFE places a heavy emphasis on security and data sovereignty. Any private entity participating in this empanelment must adhere to stringent protocols:
- Logical Separation: The infrastructure must provide a secure, multi-tenant environment. While multiple users (startups, researchers, departments) will share the platform, the system must ensure that data remains strictly siloed. Cross-sharing of data between users is explicitly prohibited.
- Domestic Hosting: All compute clusters must be located within India. This is a strategic move to ensure that the compute backbone—and the data processed through it—remains within Indian legal jurisdiction, aligning with the broader national objective of data sovereignty.
- Cross-City Distribution: The infrastructure is required to be spread across multiple cities to avoid single points of failure. While the RFE does not name specific cities, the requirement implies a distributed cloud architecture that can provide consistent service regardless of regional disruptions.
Implications for Stakeholders
For Startups and Academia
For a long time, the barrier to entry for AI research in India has been the prohibitive cost of high-end GPUs. By providing subsidized or centralized access to 10,000 AICUs, the IndiaAI Mission effectively lowers the cost of innovation. Startups can now experiment with training large language models (LLMs) and specialized AI agents without the capital expenditure of building private data centers.
For Ministries and Public Services
The integration of AI into platforms like DigiLocker and UMANG will likely transform citizen engagement. Imagine a version of UMANG that uses a localized, government-trained AI agent to help citizens navigate complex bureaucratic processes, file taxes, or claim benefits in their native language with near-zero latency.
For Private AI Service Providers
The empanelment process provides a massive opportunity for data center operators and cloud service providers in India. By participating in this program, firms are not just vendors; they are partners in the nation’s technological infrastructure. This provides a stable, long-term revenue stream and positions these companies at the center of the country’s AI revolution.
Challenges Ahead: The Road to Implementation
While the vision is ambitious, the execution will face hurdles.
- Hardware Supply Chain: The global demand for high-end GPUs (like those produced by NVIDIA or AMD) is immense. Securing 10,000 high-performance units in a competitive global market will require the government and its empaneled partners to navigate complex supply chain dynamics.
- Energy Consumption: AI compute is notoriously energy-intensive. Scaling to 10,000+ units will require a sustainable energy strategy, likely involving green data centers and high-efficiency cooling systems to meet India’s commitment to sustainability.
- Talent Gap: While the government is hiring for a "Unified AI Team," the actual maintenance of this infrastructure requires specialized hardware engineers and cloud architects. Ensuring that the human capital keeps pace with the hardware acquisition is vital.
Conclusion: A New Era for Digital India
The issuance of the RFE by C-DAC is a landmark moment in India’s digital evolution. It signals that the government is no longer just a regulator of the digital space but a proactive architect of it. By focusing on the foundational layer of AI—the compute power—MeitY is building the "digital highway" upon which the next decade of Indian innovation will travel.
As the government prepares to release its consultation paper on AI regulations, the establishment of this compute infrastructure provides the necessary physical foundation to ensure that the regulation is meaningful. In the coming months, as the first 5,000 units are rolled out, the focus will shift from procurement to deployment, and eventually, to the tangible impact on the millions of Indians who rely on government systems every day.
The message from MeitY is clear: India’s AI future will be home-grown, secure, and—above all—empowered by a world-class computing infrastructure.
For those interested in the technical specifications and the full scope of the RFE, the official document can be accessed via C-DAC’s portal/MeitY documentation.
