After taking note of suggestions from several small cloud companies, the Ministry of Electronics and Information Technology (MeitY) said in a clarification issued on Friday that the threshold has been reduced from Rs 100 crore to Rs 500 crore. . Previous standards required small and medium-sized businesses to partner with other companies, which caused some industry insiders to complain.
Tarun Dua, Chief Executive Officer of E2E Networks, a Delhi-based artificial intelligence (AI) cloud services provider, said, “The relaxation in revenue eligibility allows us to qualify directly and we deeply appreciate this flexibility. “I’m doing it,” he told ET.
However, he said there are some new provisions that require additional clarification.
According to the proposed amendments, a bidder or lead partner must have an average annual turnover of more than Rs 50 billion for the last three financial years or since registration if less than three years have passed since registration .
EtechSunil Gupta, CEO of Yotta Data Services, which ordered 16,000 H100 GPUs from Nvidia and put 4,000 H100 GPUs into service a few months ago, agreed. “By lowering turnover thresholds and introducing greater flexibility in AI computing infrastructure requirements, the government has made it easier for a variety of players, including startups in the AI services space in particular, to contribute to this mission.”
Find a story that interests you
During pre-bid meetings, the company had suggested that it should abandon the creation of a firm purchase order for 1,000 GPUs and instead consider a letter of intent.
According to the proposed amendments, bidders would now be able to create purchase orders three months after becoming the lowest bidder, instead of at the time of bid submission.
However, MeitY is now seeking a bank guarantee of Rs 500-100 billion (Rs 10 billion for 10 GPUs) in lieu of the purchase order, if the purchase order is not submitted within three months of the loan. , the bank will forfeit the guaranteed amount.
“Given the fact that the GPU supply chain remains highly constrained and the supply of GPUs is mostly controlled by a single vendor, this “This could be a huge loss for bidders.” “These kinds of risks can only be taken by very large bidders, perhaps those with a turnover of Rs 200-crore to Rs 500-crore.”
“Essentially, this lower turnover threshold may allow a few more bidders to qualify for AI mission bidding, but not many startups will qualify.” said Parimal.
“This tender is not for small businesses or high-tech startups. It is fundamentally suited for large corporations and big tech.”
But by giving additional scores to companies that took business risks early on and built on existing large-scale GPU infrastructure, Mr. He said it rewards entrepreneurship.
He was referring to a new provision under which 1 mark will be awarded for every 20 AI calculation units above 50, up to a maximum of 15 marks. If half of the installed capacity consists of AI computing units with double precision (FP64) functionality, 5 bonus marks will be awarded over the 15 marks allocated on this basis.
“Last year, we took an early and bold bet to order large quantities of Nvidia chips and build an advanced computing infrastructure with a full stack of AI platform software capabilities,” said Gupta. “The focus on Make in India highlights the government’s commitment to self-reliance and sovereignty across the AI value chain, as well as data localization, which will undoubtedly drive the growth of India’s technology manufacturing sector. It will be promoted.”
Top chip companies, cloud makers and cybersecurity companies have expressed interest in the government’s Rs 10,000-crore GPU procurement bid, which will be at the heart of India’s AI mission. A number of companies attended the August 29 pre-bid meeting, including Nvidia, Intel, Advanced Micro Devices, Qualcomm, Microsoft Azure, Amazon Web Services, Google Cloud, and Palo Alto Networks. Representatives from similar companies also attended the meeting. Yotta Data Services, L&T, E2E Networks, Tata Communications, CtrlS, ST Telemedia Global Data Centers, NTT Data, Dell, Vigyanlabs.
The bid is attracting a lot of interest because the rapid growth of the AI industry has made the domestic GPU cloud market a commercially profitable and lucrative business, executives said. These GPUs will ultimately be in huge demand, they said, and today’s enterprises are looking to take advantage of this.
ET reported on September 11 that the company is rushing to form a consortium to partner not only with each other, but also with cloud service providers and large corporations so that small and medium-sized businesses can participate in tenders.
According to another clause introduced on Friday, bidders will have to follow the guidelines of the Ministry of Commerce and Industry’s Department for Promotion of Industry and Trade and MeitY’s Make in India initiative.
Yet another new provision relates to Power Usage Effectiveness (PUE), a metric that measures how efficiently a data center uses energy. It is calculated by dividing the total amount of power used by the data center by the amount of power used by its computing equipment. A lower PUE ratio indicates better energy efficiency.
According to the new clause, bidders must ensure that the data center where AI services on the cloud are provided has a PUE of 1.35 or less. This could be the basis for eliminating multiple players, said a cloud company executive on condition of anonymity.
“This time we focused on PUE, which is a big step forward in terms of sustainability,” said another data center executive, speaking on condition of anonymity.
Nvidia was not immediately available for comment, but a Tata Communications executive confirmed that the company had received the errata from MeitY and was reviewing it.
The bidding deadline will be extended from September 25th to October 16th.
#Small #businesses #chase #Indias #big #dreams