MeitY extends deadline for Rs 10,000 crore GPU tender to November 12 amid industry concerns

He Ministry of Electronics and Information Technology has extended the deadline for submitting bids for the government’s Rs 10,000-crore graphics processing unit (GPU) tender from November 12 to October 16, according to a government official.

The tender aims to establish a public AI cloud infrastructure with a capacity of at least 10,000 GPUs through a public-private partnership under the India AI Mission.

ET viewed a copy of the corrigendum issued on Wednesday.

mark the second extension of the deadlinewhich was originally scheduled for September 6 and was later delayed until October 16. During the previous extension, new clauses were added, and some companies expressed concerns, mainly regarding the requirement to achieve a effectiveness of energy use (PUE) less than 1.35.

Also read: Companies scramble to create consortia to be eligible for Rs 10,000 crore GPU plan

PUE is a measure of the energy efficiency of data centerscalculated by dividing the total power used by the data center by the power used by its computing equipment. A lower PUE indicates higher energy efficiency.

On October 9, ET reported that several data center operators wanted the PUE criteria removed, citing the need to increase capital expenditure, which would be passed on to customers. Data center executives said achieving a PUE of 1.3 is a technical challenge in India due to high ambient temperatures, especially with direct-to-chip cooling technology. When asked if the tender clauses had been modified, the government official said that, as of Wednesday, the tender remains unchanged, although concerns raised by some companies are being considered. The error correction did not reflect any additional modifications to the tender.

“The last date for submitting offers was extended until November 12. As of today, the tender clauses remain unchanged. Some companies have expressed their concerns and we are reviewing them,” the official said.

Sunil Gupta, CEO of Yotta Data Servicessaid at the Nasscom Future Forge 2024 event in Bengaluru on Wednesday: “There are six or seven big conglomerates that are willing to invest heavily in artificial intelligence infrastructure in India.”

Yotta Data Services is one of the bidders for the IndiaAI Mission GPU tender. Last November, it placed an order for 16,384 H100 GPUs and deployed the first phase with 4,096 GPUs.

Gupta said that apart from investment in GPUs, the country would require substantial investment in underlying data centers. “AI workloads are extremely high-density. A typical data center rack consuming 6 KW of power would require 100 KW per rack,” he said.

“Globally, data centers consume almost 3% of the world’s total energy production. If you introduce thousands of GPUs, this number risks reaching double-digit percentages. This is a global concern,” said Gupta, noting that Singapore had imposed a moratorium when data centers began consuming 7% of the country’s total energy.

However, Gupta argued that in India, where green energy is promoted, data centers consume between 1,500 and 1,800 MW (around 0.2% to 0.3% of the country’s total energy production), so It is not a major concern. He emphasized that technologies such as direct-to-chip cooling and immersion cooling could help reduce power consumption.