Indian Prime Minister Narendra Modi, whereas talking on the Responsible AI for Social Empowerment (RAISE) 2020 summit, stated he desires the nation to change into an AI hub, nonetheless, one of many main drivers of AI, apart from information, is compute. For instance, the supercomputer hosted in Microsoft Azure, particularly designed for OpenAI, incorporates 1000’s of NVIDIA Graphics Processing Units (GPUs) to energy AI workloads and speed up coaching processes. But, at present, the worldwide AI business is dealing with a scarcity of those GPUs.

The market scarcity is important to the extent that even NVIDIA’s deliberate manufacturing of two million GPUs for 2024 is already sold out. Today, not simply AI labs however completely different organisations and international locations are attempting to get their arms on these GPUs. The UK Prime Minister Rishi Sunak is planning to spend USD 126.3 million to purchase AI chips as part of the nation’s plans to enhance its AI sources and change into a frontrunner in AI. Reportedly, his administration is already in talks with main AI chip makers similar to NVIDIA, AMD and Intel. 

Given India’s ambitions in AI, the scarcity may have each short-term and long-term implications for India. Moreover, owing to the intricate nature of GPUs, they’ll change into entwined with geopolitics. For occasion, the US authorities’s imposition of export controls prevents NVIDIA from promoting its AI chips to China. Considering the business’s heavy dependence on NVIDIA’s near-monopoly within the GPU market, may amassing GPUs, just like the UK, emerge as a rational transfer for India?

Will the GPU disaster impression Indian firms?

In India, there are quite a few startups and huge organisations engaged on or deploying AI fashions at completely different ranges. Today we’re within the age of generative AI. Its explosive progress, which includes coaching complicated fashions that require vital computational energy, has led to a surge in demand for AI-focused GPUs. In India, a number of firms have constructed or are within the technique of constructing proprietary Large Language Models (LLM). 

Leena AI, which is a conversational AI-backed platform, has developed WorkLM, the corporate’s proprietary LLM constructed particularly for enterprise worker expertise. In a earlier interplay with AIM, Mayank Kumar, co-founder & managing director at upGrad stated that the edtech agency is exploring the thought of constructing its personal proprietary LLM. Similarly, Indian IT large Tech Mahindra is already constructing an Indic-LLM that will have the flexibility to converse in over 40 Indic languages.

Moreover, there are over 4000 AI startups in India, in response to Tracxn. Given the quantity will solely enhance within the coming years, the computing energy required to coach fashions will even enhance considerably. Even although Indian companies can depend on hyperscalers like AWS and Azure for computational energy, shopping for GPUs brings vital benefits. In such a case, bigger firms might discover it comparatively simpler to amass GPUs,  however smaller companies may face challenges in acquiring them. Pawan Prabhat co-founder of Shorthills AI believes the federal government shopping for GPUs may not be the suitable thought as a result of, traditionally, India has executed higher in sectors the place the federal government has had much less management. “Rather it can help software companies offset the increased costs by giving some tax exemptions on purchases of GPUs physically or in the cloud for a limited period. Just like we have tax policies and SOPs to support startups, the government can be of help indirectly,” he informed AIM.

However, Ranjan Chopra, managing director and chief govt officer at Team Computers believes the Indian authorities ought to think about allocating funds to amass GPUs to advance its AI ambitions. If the federal government makes such an funding, it might align completely with India’s imaginative and prescient for technological management, fostering innovation, driving financial progress, and guaranteeing competitiveness within the international AI panorama. “Such a strategic investment not only accelerates AI progress but can also position India as an innovative powerhouse, supporting startups, fostering economic growth, and strengthening our global AI leadership,” he informed AIM. 

GPUs are the spine of AI Research 

Moreover, for India to realize management in AI, it should place a considerable emphasis on AI analysis, a essential part of which is GPUs. The authorities has established the National Research Foundation (NRF) with the goal of enhancing analysis efforts within the nation, together with the sphere of AI analysis. In a earlier interplay with AIM, Prof. Arnab Bhattacharya, Dept of Computer Science & Engineering, IIT Kanpur, stated that AI researchers in India usually battle with funds required to purchase {hardware} similar to GPUs/TPUs. 

Currently, there are AI supercomputers like AIRAWAT, powered by NVIDIA DGX A100 GPUs, empowering researchers in India. Nonetheless, sooner or later, the demand for computational energy is anticipated to extend considerably. In this gentle, the federal government has introduced plans to develop nine more supercomputers within the nation. However, the requirement for supercomputers within the nation is various, therefore, we are able to hope a portion of the brand new supercomputers might be devoted in direction of AI analysis. 

Kesava Reddy, CRO at E2E Networks informed AIM that ​​the federal government can discover methods like bulk buying to subsidise the price of GPUs for analysis and growth functions. “Potentially a PPP (Public-Private Partnership) model can be explored to make GPUs more accessible and affordable for AI research. This can potentially be done in collaboration with local Cloud Service Providers (CSPs), who enable the Indian market and are being used by local startups and enterprises.” 

Furthermore, startups who’re constructing generative AI applied sciences may profit from this. “Using money to get GPUs can help Indian AI startups and research projects have the right tools for their work. But we should also aim to make GPUs here in India, so we don’t have to rely on others in the long term,” Vaibhav Srivastava, senior data safety analyst/ advertising and marketing Lead at Innefu Labs, informed AIM. 

Making India self-reliant 

Similar to India’s deal with changing into self-reliant in semiconductors, the query that arises is ought to India goal to change into self-reliant in AI {hardware} as properly? Similar to the US-China situation, if tomorrow, the US forbids NVIDIA from transport chips to India, it may change into a possible disaster. 

Hence, “I strongly advocate that India should prioritise the local production of GPUs and AI chips, akin to the strategic focus on semiconductor manufacturing. Local production offers several strategic advantages. Firstly, it reduces India’s reliance on international markets, ensuring a more dependable supply. Recent global supply chain disruptions underscore the need for self-reliance in critical technology components,” Chopra stated.

Reddy additionally believes with India’s deal with ‘Make in India’, and native manufacturing, constructing a home business for GPUs and AI chips generally is a highly effective long-term technique. Much just like the USD 10 billion Production-Linked Incentive (PLI) scheme for semiconductors, the federal government may entice business giants similar to NVIDIA and AMD to ascertain manufacturing models inside the nation.

But Prabhat, however, is of the opinion that making a GPU or an AI chip at this cut-off date might be a tall order. India is making an attempt its hand at making a fab and it may not be a good suggestion to get into creating AI {hardware} proper now. “India has traditionally been very strong in software and given that we are also far off from creating a cloud infrastructure company like AWS, Azure, etc., it might be a better idea to collaborate with these chip companies and cloud companies to help them set up their units in India. In due time, we can also create our own GPU/ AI chips.” But within the short-term, the most important leverage would doubtlessly come from determining methods by means of which native startups and companies can entry superior GPUs, Reddy stated.