China’s AI Leap: Can India Catch Up with DeepSeek?

• India contributes only 1.4% to global AI research, far behind leaders like the U.S. and China, who dominate with over 50% combined output.
• India lacks supercomputing resources and a robust cloud infrastructure, making access to high-performance GPUs expensive and limited for researchers.
• Indian corporations prioritize short-term profits, leading to minimal investment in AI research and innovation compared to strategic efforts by Chinese tech giants like Alibaba and Tencent.
• Despite producing many engineers, India’s top AI talent migrates to countries offering better infrastructure, career growth, and living conditions, exacerbating a brain drain.
• India can advance by building local cloud infrastructure, encouraging corporate investments, enhancing AI education, fostering public-private partnerships, and supporting AI start-ups through initiatives like the IndiaAI Mission.

It was the technological version of an ambush. On January 28, 2025, Chinese start-up DeepSeek took the world by surprise by releasing an artificial intelligence assistant that is reportedly on par with industry-leading models such as ChatGPT and Copilot at a fraction of the cost.[1] With its ability to deliver accurate insights from vast data sets at unprecedented speeds, the Chinese company is threatening to upset the dominance of America in AI. For Indians, it also highlighted an ominous fact – the Chinese are light years ahead of India.

In recent years, AI has emerged as a driving force behind innovation across industries, from healthcare to finance and manufacturing to transportation. However, despite its burgeoning promise, India is falling behind the global leaders. As of recent studies, India ranks 14th globally in AI research, contributing only 1.4 percent of papers to top AI conferences between 2018 and 2023.[2] In contrast, the U.S. and China contribute over 50 percent to the world’s AI publications — 30.4 percent and 22.8 percent, respectively. This gap reflects a complex array of challenges, from inadequate infrastructure to a lack of investment and talent retention.

India cannot afford to lag too far behind China in AI because the future of technological dominance hinges on advancements in artificial intelligence. Falling behind means losing competitive advantage, market leadership, and innovation potential. China’s rapid strides in AI pose strategic and economic threats to India’s global standing.

India’s struggle to catch up with the AI powerhouses is not due to a lack of potential or talent. Rather, it is the convergence of several critical factors — ranging from infrastructure challenges and limited computing power to a lack of significant investment in AI research — that are inhibiting the country’s AI progress.

Compute Crunch

One of the most significant barriers to AI research and development in India is the scarcity and high cost of computing power.

Artificial intelligence, particularly deep learning, requires massive computational resources, often in the form of graphics processing units (GPUs), which are used to train AI models. High-end GPUs like the Nvidia H100 can cost up to $40,000 per unit, making them out of reach for many research institutions and start-ups in India. GPU imports to India were also restricted by the U.S. after the Biden administration capped the number of such units countries could acquire. [3]

Now chew on this – of the top 500 supercomputers in the world, India has only three. China and the US have 215 and 113, respectively.[4] For AI researchers, this means higher barriers to entry, with many universities and research institutes unable to access the computational power necessary for cutting-edge AI work.

Infrastructure Challenges

Due to the lack of computing infrastructure, many AI companies in India rely on cloud infrastructure. However, India’s cloud adoption is also at a very nascent stage. Large cloud providers in the U.S., like AWS, Microsoft Azure, and Google Cloud, have built massive data centers that provide AI researchers with scalable computing resources at reasonable costs. India, on the other hand, lacks a robust native cloud computing ecosystem, making it difficult for AI researchers to access the necessary infrastructure locally.

Additionally, concerns around data security and privacy in India have led many companies and academic institutions to avoid storing sensitive data on foreign servers, further complicating the issue. This absence of a comprehensive, locally available cloud infrastructure drives up costs for AI research and innovation, particularly for industries that rely on big data to train AI models.

Lack of Corporate Interest

An Indian AI scientist working at Nvidia’s India office told me that the lack of corporate investment in AI development is a key reason why the sector is not taking off in India. “Currently, the thinking in Corporate India is that investments in AI do not offer them value,” said the scientist who requested anonymity.

“Indian companies are not strategically driven or positioned; they are shareholder driven. Many of them also have large Indian and overseas investors who are focused on maximizing profit. These investor won’t allow Indian companies the bandwidth to invest the substantial sums required for AI development.”

Not only are entry costs prohibitive in AI, there is an ongoing – and escalating cost – as your users grow. Ashesh Rajhans, an IIM Ahmedabad alumnus who recently launched an AI startup in Bengaluru, explains: “Training an AI model can be very expensive in terms of compute power. But it is a one-time expense. The real expense comes when people use the AI model, and is known as inferencing. Inference is continuous and requires a lot of computer power, making it costly.”

This means the more users you attract, the higher your computing costs. And since 99 percent of users are not prepared to pay for using AI chatbots, there is little to zero revenue for the developers. Only mega-corporations like Microsoft, Google, Amazon, and Alibaba have the capacity to absorb these losses, which can amount to billions of dollars.

Rajhans recommends that large Indian corporations like the Ambanis and the Adani Group should invest in AI. “The Indian government can then offset their investment costs via subsidies or tax breaks. All countries do this; India should emulate them,” he adds.

Rajhans points out another problem faced by AI companies: “You can outspend everyone and be at the top of your game, but then a new upstart enters the arena and rivals your product at a fraction of the cost. Suddenly, your entire business model crumbles. That’s what DeepSeek did to OpenAI’s ChatGPT. Now OpenAI’s high-cost model – using billions of dollars and hundreds of engineers – is being questioned.”

So what about the rash of AI start-ups that have cropped up in India?[5] According to the Nvidia scientist, most of these companies are small-scale operations that lack the resources to make significant strides in the AI space. “Many of these start-ups rely on pre-existing AI models developed by companies like OpenAI or China’s Alibaba Cloud (Qwen);  they have not created their own large language models.” (Note: LLMs are a type of software trained on vast amounts of data that underpin AI applications.)

In contrast, China’s AI sector is strategically driven by companies like Alibaba and Tencent. These companies recognize the value of AI in improving efficiency and predicting market trends in their consumer-facing businesses. These companies invest heavily in LLMs, with a clear focus on leveraging them for competitive advantage. For instance, Alibaba has been incorporating AI tools like automated replies and chatbots to enhance customer service for its merchants. These features are valuable because they offer round-the-clock support, answering common queries even when live agents aren’t available.[6]

State of Academia

India’s academic institutions, including the prestigious Indian Institutes of Technology (IITs), are making commendable strides in AI research. IIT Bombay, for example, has an AI research division that is among the best in the country. However, even the most renowned institutions face significant limitations in terms of access to cutting-edge AI tools and computing power. Without easy access to high-performance GPUs, Indian students and researchers cannot compete with the best AI labs globally. As a result, many talented AI researchers eventually migrate to countries that offer better opportunities for advancement, further perpetuating the brain drain.

To illustrate, a postgraduate student from IIT Bombay joined a small Bengaluru-based AI start-up in 2022. However, he became frustrated by the lack of access to GPUs, which he had been promised when he was offered the job. He quit in disgust the following year. He told me he had been working on speech-to-text, an area where India has just a handful of professors. “If they are snapped up by some overseas university, research in speech-to-text in India would practically grind to a halt,” he said.

Talent Acquisition and Retention

India faces significant challenges in attracting and retaining top-tier AI talent. The country produces a large number of engineering graduates, but most of them lack the deep expertise required to push the boundaries of AI research. Plus, while Indian cities like Bengaluru have emerged as technology hubs, they still cannot compete with cities like San Francisco or Beijing in terms of living standards, infrastructure, and opportunities for career growth.

India’s inability to retain talent is exacerbated by the fact that its top AI researchers often seek opportunities abroad, particularly in the U.S.,  where the availability of resources, better living conditions, and greater career advancement prospects make them more attractive. As a result, India struggles to develop and retain a critical mass of AI talent, further hindering the country’s progress in the field.

Location Conundrum

As Germany found out of its peril in World War II, complete economic independence is a chimera. No country can do it alone. America’s lead in AI is possible because of its world-class universities, research labs, and highly liveable cities that attract the best engineers and scientists from all over the world. Similarly, France’s AI industry is centered around Paris, one of the world’s most popular cities.

Cross-border talent seeks liveable locations, and India has a huge problem in this regard. All large Indian cities are polluted and crowded, with extremely poor air quality. Where the living conditions are good, for instance, in smaller cities like Chandigarh, there are not many jobs.

AI requires a collaborative effort. “You need foreign researchers and academics to stay long-term in the country and help develop the AI sector, but such people are extremely selective about where they live,” said the Nvidia scientist.

If India hopes to attract global IT talent, it needs to create more tech-driven enclaves like GIFT City in Gujarat, where the infrastructure and living standards are on par with those of the West.[7] The best places for building brand new cities on the GIFT City model are in the Western Ghats – in the stretch between Ratnagiri and Karwar. They have the ideal mix – mountains, beaches, fertile plains, tropical climate, and adventure tourism – that today’s highly driven engineers seek.

India faces another major roadblock preventing skilled foreign workers from moving to India. “We have way too many protests and processions and a seemingly endless cycle of elections,” said the Nvidia scientist. “All this can be quite intimidating for someone from a developed country. They find India to be a chaotic place. We need to sort out this mess first. As well as clean cities, we must create a peaceful and stable environment too.”

How India Can Catch Up in AI

Despite these challenges, India is not without hope. The country can take several steps to catch up with global AI leaders, which involve both short-term actions and long-term strategic shifts.

• Investment in AI Infrastructure: Build a robust cloud computing infrastructure that can support AI research at scale. The government can play a significant role here by incentivizing global cloud providers to set up data centers in India. Additionally, Indian universities and research institutions must be provided with grants or subsidies to access cutting-edge GPUs and computing power through collaborations with global tech giants or state-sponsored initiatives.
• Corporate Investment in AI: Indian corporates need to recognize the long-term strategic value of AI and start investing in it more aggressively. Large IT companies like Infosys, Tata Consultancy Services (TCS), and Wipro have significant financial resources that could accelerate India’s AI development. For instance, Infosys posted a gross profit of Rs 29,000 crore in 2024[8], and TCS posted Rs 24,000 crore.[9] These companies can set up dedicated AI research centers, collaborate with universities, and invest in AI start-ups to create a thriving ecosystem.
• Focus on Talent: India must improve its AI education at all levels — primary, secondary, and tertiary. AI should become a core subject in technical curricula across engineering institutes. The government and industry can collaborate to provide scholarships, fellowships, and internships to develop a pipeline of AI talent.
• Public-Private Partnerships: The Indian government can play a critical role in advancing AI by creating policies that encourage collaboration between public institutions and private industry. Government-led AI initiatives could provide research grants, tax breaks, or public sector funding to start-ups and established AI companies.
• Global Collaborations and Networks: Focus on building stronger collaborations with international research bodies and universities. This would help elevate the standard of AI research in the country, expose Indian researchers to global trends, and foster the exchange of ideas.
• Top Down Approach: The jolt provided by DeepSeek has woken up the Indian government. Electronics and Information Technology Ashwini Vaishnaw announced that as many as six firms would develop Indian foundational AI models using newly-provisioned GPUs by the IndiaAI Mission.[10]

The IndiaAI Mission, under the IT Ministry, will make a cluster of 18,693 GPUs available virtually to start-ups and researchers immediately for AI research and building foundational models. Almost half of the ₹104 billion outlay for the IndiaAI Mission is dedicated to making this computing facility available.

Algorithmic efficiency could be the game changer, allowing India to build models faster and cheaper. Vaishnaw claims the foundational models are set to rival the likes of DeepSeek and OpenAI. “We’re inviting proposals to create our own foundational model tailored to the Indian context — one that embraces our languages and culture, eliminates biases, and is built on datasets that reflect the realities of our citizens,” he said.

Conclusion

While India faces significant challenges in catching up to the U.S. and China in AI, these obstacles are not insurmountable. With strategic investments in infrastructure, corporate commitment to long-term development, and a focus on talent retention, India can carve out a leadership role. The time to act is now, as the potential for AI to drive economic growth and social progress in India is enormous. By taking proactive steps, India can ensure that it doesn’t just keep pace with AI advancements globally but plays a significant role in shaping the future of this transformative technology.

Citations

[1] What is DeepSeek and why is it disrupting the AI sector? (Reuters); https://www.reuters.com/technology/artificial-intelligence/what-is-deepseek-why-is-it-disrupting-ai-sector-2025-01-27/

[2] India 14th in AI research with just 1.4% share of papers: study (The Economic Times); https://economictimes.indiatimes.com/tech/artificial-intelligence/india-14th-in-ai-research-with-just-1-4-share-of-papers-study/articleshow/112554830.cms? UTM_Source=Google_Newsstand&UTM_Campaign=RSS_Feed&UTM_Medium=Referral

[3] India to take up AI chip export curbs with Trump Govt. (The Economic Times); https://economictimes.indiatimes.com/tech/technology/india-to-take-up-bidens-ai-chip-export-curbs-with-trump-govt/articleshow/117406695.cms?from=mdr

[4] https://www.top500.org/lists/top500/2020/11/

[5] Indian GenAI Startup Tracker: 60+ Startups Putting India On The Global AI Map (Inc42); https://inc42.com/startups/indian-genai-startup-tracker/

[6] Alibaba bets on AI to fuel cloud growth as it expands globally to catch up with U.S. tech giants (CNBC); https://www.cnbc.com/2024/05/23/alibaba-bets-on-ai-to-fuel-cloud-growth-as-it-expands-globally.html

[7] GIFT Special Economic Zone (SEZ) – A Global Financial Hub (Gujarat International Finance Tec-City); https://giftsez.com/

[8] Financial highlights of the Third Quarter ended December 31, 2024 (Infosys); https://www.infosys.com/about/last-quarter.html

[9] TATA Consultancy Services Limited Unaudited condensed consolidated interim statement of financial position 2024 (TCS);  https://www.tcs.com/content/dam/tcs/investor-relations/financial-statements/2024-25/q2/IFRS/Extracts%20from%20Consolidated,%20Unaudited%20-%20INR.pdf

[10] India to build foundational AI model in months, GPUs to be made available to start-ups, academia at subsidised rates: IT Minister (The Hindu); https://www.thehindu.com/incoming/india-to-build-foundational-ai-model-in-months-gpus-to-be-made-available-to-start-ups-academia-at-subsidised-rates-it-minister/article69159786.ece