In a move to propel India and help local startups in AI innovation, Union Minister Rajeev Chandrasekhar revealed a plan to establish a cluster of 25,000 GPUs.
This initiative, which should be carried out within the framework of a public-private partnership (PPP), is under discussion at the highest levels of the Ministry of Electronics and Informatics. Chandrasekhar announced the ambitious effort in September, highlighting his commitment to fostering real-world applications of AI.
The minister said ongoing discussions about AI almost always focus on applications like ChatGPT. “Our mission is about real-world AI use cases. We are looking at health, governance, education and creating AI-specific integrated circuits for these applications,” he said in a statement.
Abundant data, need GPU
At its core, this initiative is not just about improving India’s AI prowess; it is also about safeguarding the country’s data sovereignty. The scarcity of GPUs in the country has pushed many companies to rely on cloud-based solutions abroad. Aware of the urgency of tackling this issue, one of the seven working groups on AI established by MeitY strongly recommended the creation of a cluster of 25,000 GPUs.
Experts say India’s abundant data resources and human capital require supercomputing power to compete in the global AI arena. This massive GPU cluster is a crucial step toward achieving this goal. For context, the fastest supercomputer in India currently, “Airawat‘, has only 640 GPUs, ranked 75th in the world. In contrast, the world’s best supercomputers have more than 30,000 GPUs.
Once the proposal is finalized, the government will launch a standard tender process to invite private companies to participate in the creation of the GPU cluster. Notably, discussions with tech giants like NVIDIA, including a meeting between its founder and CEO Jensen Huang and Prime Minister Modi, highlighted the potential for collaboration.
Huang said AIM what India will do Tens of thousands of GPUs in order to build an infrastructure – approximately 1,00,000 GPUs. “We are going to release the fastest computers in the world. These computers aren’t even in production (until now). India will be one of the first countries in the world (to get them),” Huang said, confirming that it would be faster than anything the world has ever seen.
If we are to believe the figures mentioned by the founder of NVIDIA, the cluster of 25,000 GPUs could very well be part of a larger delivery containing 1,00,000 GPUs.
However, while this first initiative is undeniably important, it only marks India’s first foray into the arena of countries promoting AI research and development capabilities. Compared to companies like OpenAI, which have more than 20,000 GPUs and a $10 billion investment from Microsoft, the government will need partnerships with the private sector to fully exploit the potential of this computing power.
The estimated cost of this ambitious project is between INR 8,000-10,000 crore and is currently under deliberation at the highest levels of Meity. Indian AI startups, industry players and prominent CEOs have consistently advocated for such investments in computing capacity to address the scarcity and prohibitive cost of GPUs.
Issues Related to Local AI Advancement
National initiatives to build supercomputers and plans to train LLMs in several Indian languages are already underway. However, many problems arise.
When businesses seek access to GPUs from cloud service providers or GPU manufacturers, they deal with prolonged wait times, sometimes extending over months. To address this bottleneck, companies are urging the government to invest in critical IT infrastructure for AI systems and applications. Without this support, India risks falling behind in the global AI race, which encompasses applications from banks to space stations, all powered by algorithmic intelligence.
“India’s leading startups are grappling with the challenge of gaining access to clusters of 1,000 GPUs, often diverting valuable funds from their fundraising efforts, managing director of PeakXV Partners (formerly Sequoia India). Rajan Anandan, said; highlighting the need for affordable access to these clusters, suggesting a pyramid approach: free access for academic institutions, subsidized access for startups and commercial access for large companies.
IBM CEO Arvind Krishna recently reiterated the same saying: “In many emerging technologies, it often requires the government to step in first before others follow. »
“The government should set up a national computing center for artificial intelligence,” Krishna stressed.
India is home to over 60 active genAI startups as of May 2023, having received approximately $475 million in funding between 2021 and 2023. Although the Indian AI ecosystem is thriving, it lags behind other countries like the United States and Israel in terms of fundamental AI models and funding.
What about other countries?
In contrast, governments in other countries have committed substantial funds to securing access to GPUs for research purposes. The UK, Saudi Arabia, UAE and Chinese tech giants have all invested significantly in acquiring GPUs to bolster their AI capabilities. Even the United States offered a 50% discount to researchers working on supercomputing projects.
Lack of access to local GPU could push Indian companies to opt for foreign cloud providers, leading to data localization issues. Sudipta Ghosh, Partner and Head of Data and Analytics at PwC India, highlighted the importance of regulatory frameworks for responsible and ethical AI. Such frameworks would not only build public trust but also ensure transparency and accountability.
The scarcity of GPUs has also made them more expensive in India, leading suppliers to be cautious about shipping them to a country where demand, payment capabilities and ticket sizes are comparatively smaller.
Meeting this challenge could require domestic GPU manufacturing, potentially supported by government incentives. Prashant Garg, partner at EY, suggests following the model of attracting global automakers to set up shop in India.
Efforts are already underway to provide GPU access to AI startups through collaborations between Nasscom and the Center for Development of Advanced Computing (CDAC). However, experts agree that India needs to invest in basic research and attract top AI scientists to drive innovation.
Meanwhile, collaborations between US GPU maker NVIDIA and Indian giants Reliance Jio and Tata Sons show promise in providing IT infrastructure to emerging companies.
Looking ahead, other semiconductor giants like AMD, Micron, SOLIS-IDC, Foxconn and STMicroelocronics could also snowball with many GPU manufacturers. With loose regulations and favorable trade conditions, India could manufacture GPUs locally.
