In a first, India notifies cloud, data centre and AI ethics standards

A government notification dated 25 February and reviewed by Mint shows the standards establish common definitions for cloud systems, introduce performance metrics for data centres, and embed ethical considerations within AI system design, formalizing the technical and governance backbone of India’s AI expansion.

The standards have been notified under Sub-rule (1) of Rule 15 of the Bureau of Indian Standards (BIS) Rules, 2018. They are Indian Standards derived from international ISO/IEC frameworks.

For now, they are not mandatory. To make compliance compulsory, the government would need to issue a Quality Control Order (QCO). Government officials say notifying the standards is a first step that could pave the way for future QCOs, should the government decide to mandate adherence.

The move comes as Indian conglomerates accelerate investments in AI-ready data centres and digital infrastructure, signalling an effort to formalize the technical and governance backbone of India’s AI expansion.

"Since AI is developing at a rate that no other technology has, regulations must keep up without impeding innovation. A practical step in that direction is the government's decision to review AI regulations and bring them into compliance with international frameworks,” said Rajesh Chhabra, general manager, APAC, Large Markets, Acronis, a provider of AI-powered cybersecurity for businesses and individuals.

However, given the speed at which cyber threats are increasing, integrating security into AI governance from the start will be essential to establishing long-term digital trust, he said.

Queries sent on Sunday to Bureau of Indian Standards, and to the ministries of commerce and MeitY, remained unanswered till press time.

It has also laid down a methodology for calculating the Cooling Efficiency Ratio (CER), which measures how efficiently data centres remove heat relative to the electrical energy consumed. Processing large volumes of data generates significant heat, making cooling central to operations. As AI workloads intensify, energy efficiency and thermal management have become key operational concerns.

A standardized metric allows operators and clients to compare cooling performance within a recognised framework, the official said.

Globally, data centre efficiency is assessed using metrics such as Power Usage Effectiveness (PUE), Water Usage Effectiveness (WUE) and Carbon Usage Effectiveness (CUE). Officials said globally used metrics will continue to be followed in India.

The timing is significant. According to a Deloitte–NITI Aayog report, India’s data centre capacity is projected to rise from about 1.5 GW in 2025 to 8–10 GW by 2030. AI-led expansion alone could require an additional 40–45 terawatt hours (TWh) of power in 2030, compared with 10–15 TWh in 2024. Data centres’ share of total electricity consumption could rise from around 0.8% at present to roughly 2.5–3% by 2030.

These standards outline how technology should be designed and deployed in systems that support independent living. They describe key building blocks, including devices, sensors, software platforms and users, and define how components connect and exchange information. They also provide guidance on fairness, transparency, privacy, safety and accountability when AI systems are used to monitor or assist individuals.

Officials said the standards address ethical considerations such as protecting personal privacy, ensuring secure data handling, promoting transparency in AI decision-making, avoiding bias or unfair treatment, and maintaining human oversight in critical decisions.

While the cloud and data centre standards do not directly regulate AI applications, they define the operating environment in which AI systems function. Benchmarks on infrastructure efficiency and system architecture can influence procurement norms, certification requirements and investor confidence, the official said.

The notified standards define core cloud concepts, establish a methodology for measuring cooling efficiency in data centres, provide guidance on asset management, prescribe testing methods for optical fibre and coaxial cables, and lay down technical specifications for USB 4 interfaces used for data and power transmission.

Industry executives said regulatory clarity could help reduce uncertainty for startups and enterprises.

“The priority should be simple — enable experimentation while protecting users. Regulations must support builders, not overwhelm them, and encourage responsible innovation without slowing momentum,” said Ankit Aggarwal, founder, Unstop, an AI-powered platform for students and freshers to upskill.

Artificial intelligence has emerged as a core government priority, reflected in FY26 Budget allocations including ₹2,000 crore for the IndiaAI Mission and ₹500 crore for a new AI Centre of Excellence.