Can India Catch up to China in AI Capabilities
Unlike China, India does not benefit from its AI talent since many top graduates pursue jobs in the U.S. says Sunil Mani
(Image courtesy Wikimedia Commons.)
August 9, 2025
By Sunil Mani*
Last month, the Government of India announced a Research, Development, and Innovation scheme to promote private-sector-led research and innovation in key technologies, including artificial intelligence (AI), semiconductors, quantum computing, biotechnology, and advanced agricultural techniques. The $11.4 billion budget for the plan is a big sum compared to previous govrenment spending on similar programs. But, in the AI field, the question is, will the funding be enough to accelerate India’s capabilities and narrow the gap with China?
It is roughly three years since the importance of AI first got wide public attention following OpenAI's introduction of ChatGPT. The research, development, and applications of AI tools and the nurturing of AI talent are now key necessities for the defense of a country as well as for its economic competitiveness and growth.
Overall, in the AI race, it appears that the United States has a slight lead over China. Nvidia, an American company with a market value of over $4.5 trillion, leads in the supply of advanced chips required to build supercomputers. The leader among AI tools is San Francisco based OpenAI, with an estimated value of $500 billion. In the defense field, AI based U.S. companies include Palantir, which has a market value of $442 billion, and Anduril, with an estimated value of $31 billion.
China though has a lead over the U.S. in several AI applications, including surveillance, facial recognition, smart cities, manufacturing automation, robotics, and consumer internet platforms. China’s AI development is far ahead of that in India, adding a new dimension to their geo-political rivalry.
As in other major parts of its economy, China's advance in AI is driven by three key elements: ambitious government policies with clear goals and deadlines; massive government funding to develop talent and promote research mostly at universities and at some institutes; and financial and other support for large companies, mostly private, which emerge as winners after intense competition among hundreds of startups.
On the policy side, AI is a core component of national strategies such as "Made in China 2025" and its successors. The policies are implemented by close co-ordination between government agencies and companies to develop research, computer networks, data centers, power sources, and other infrastructure and supplies. Also, they work with universities to build talent in AI, machine learning, math, and related areas. All of this enables the rapid transformation of AI research into large-scale deployment of tools for commercial, defense, and other government projects.
Between 2020 to 2025, the Indian governments spent roughly $1.5 billion on AI research at universities and research institutes. Some estimates suggest that China’s comparable budget was almost 50 times larger.
In March 2024, the government of India launched its IndiaAI Mission, 2024-2030. It seeks to develop AI infrastructure, including more than 10,000 advanced computer chips, support creation of large language models and fund startups, academic research and data gathering. The $1.3 billion budget marks a significant step up in India's AI effort. But the funding is tiny compared to the sums being spent by China.
Currently, according to a RAND study, China is increasing public funding for its AI industry through specialized industry funds, bank loan programs, and local government funding: including $8.2 billion for startups, and Bank of China’s five-year, $138 billion financing program.
Massive funding is necessary since the development of AI is capital and technology intensive. A major expense is buying supercomputers. In December last year, for instance, Stanford University budgeted $30 million to purchase a supercomputer, recruit a team of data scientists, and for operational expenses.
As of June 2025, China had 47 of the Top 500 supercomputers in the world, while India had just six. This puts India at a disadvantage since supercomputers are the foundation for AI development. They provide the immense computational power needed to train large AI models like ChatGPT or China's DeepSeek—often requiring thousands of high-performance chips, running for weeks or months. They also enable the processing of massive datasets of text, images, scientific research, and other material, at high speeds, as well as support complex AI-driven simulations in areas such as autonomous vehicles, climate modelling, and drug discovery.
(Photo: Courtesy IIT, Bombay.)
In the field of talent, in 2022, China had about 78,000 domestic AI researchers, more than five times larger than the 15,000 in India, according to a Stanford University study. AI capabilities in some areas are doubling every four months in China and the US, and so is the need for AI talent. It is hence likely that the talent gap between China and India has widened since 2022.
In theory, India’s key advantage in AI is its vast potential pool of talent, including hundreds of thousands with good aptitude for science, engineering, math, and statistics (STEM). But the capacity to train them is limited due to shortages of faculty, laboratories, computers and other technical equipment, as well as campuses.
The government-run universities and research institutions include 23 Indian Institutes of Technology (IIT), seven Indian Institutes of Science Education and Research, the Indian Institute of Science, Bengaluru, the Indian Statistical Institute, and 31 National Institutes of Technology. In addition, there are private universities such as the Manipal Institute of Technology, and three BITS campuses. In 2025, these top STEM institutes in India are estimated to have graduated around 30,000 students, including graduates, postgraduates, and PhDs, with about two thirds of them from the IITs.
In contrast, this year Chinese universities are estimated to have graduated roughly 70,000 PhDs in STEM, with about half of them from top quality universities including Tsinghua and Beijing, according to a Georgetown University study. The total output of top STEM graduates, post graduates, and PhDs in China is at least five times greater those from the top Indian universities.
Also, there is a wide gap in quality even among the top universities in India and China. Not one of the top 50 engineering colleges in India is among the 100 best in the World, as Ashok Nag points out in a Global Indian Times story. The highest-ranking Indian institute holds a position of 145, while five of the top ten are from China. Overall, while 42 Indian institutes secured a spot on the U.S. News Global List of the 1000 best engineering schools, for 2022-23, there were 193 institutes from China on the list.
Equally important, there is a major difference in the potential AI talent pool of the two countries. China retained nearly all of its top 0.1% most cited AI researchers domestically, compared to India retaining less than two thirds (62%) of its top talent. This is because many top-quality graduates from India take up AI research posts or jobs in the U.S. or other Western countries, which hurts India's development of AI capabilities. In sharp contrast, leading researchers in China stay on and strengthen the AI talent pool in the country, with many of them founding startups.
Not surprisingly, China's lead in AI is also reflected in research publications and patent applications. In 2023, for instance, more than two thirds of the world's AI patents were granted to researchers in China, while those from India accounted for less than half a percent, according to a Stanford study.
(Image: courtesy Wikimedia Commons.)
There is a sizeable gap in the AI funding and capabilities of Chinese and Indian companies. Huawei and Baidu are among the leading Chinese AI firms. In 2024, Huawei’s revenues were $118 billion. Based in Shenzhen, it is a leader in developing China’s AI infrastructure and plays a key role in the country’s strategy to reduce dependence on foreign technology. For instance, its Ascend processors, used by Chinese data centers, are designed to replace advanced computer chips from Nvidia. Huawei also provides cloud-based AI tools and services to major Chinese enterprises as well as to the government, including for weather forecasting.
Each year, Huawei invests between $5 to $7 billion in AI-related research and development (R&D), according to published statements. The actual figure is likely far higher given that government funding in defense areas are typically not disclosed by companies.
Baidu, based in Beijing, leads in AI-powered search, autonomous driving via its Apollo platform, and large language models like Ernie Bot, which competes with OpenAI's GPT. The company operates Paddle, China's largest open-source AI framework, and applies AI to healthcare diagnostics, cloud computing, and other areas. Baidu spends roughly $3 billion annually on AI R&D, making up about a fifth of its revenue. Its Apollo self-driving initiative has received roughly $2 billion in funding since 2017, according to Crunchbase. Baidu is valued at around $30 billion
Compared to the Chinese companies, Indian AI companies are miniscule. Zoho, a privately owned AI-powered software company, is valued between $5 to $7 billion. Other Indian AI companies include CropIn, in the field of agriculture, Niramai, in healthcare, Locus, in logistics, Sigmoid, in financial fraud detection, and Qure.ai, in radiology diagnostics.
There are more than 1,500 AI startups in India, together employing more than 50,000. However, given limited domestic supercomputing infrastructure, they depend on expensive cloud services such as those from Amazon Web Services or Microsoft’s Azure. Some Indian startups, with limited funding, use older computer chips which restricts their capabilities.
One vulnerability of China's AI strength, is dependence on imports of advanced computer chips, especially from the U.S., as a recent UNCTAD report notes. In fact, China is unable to buy the latest computer chips from Nvidia due to U.S. export restrictions. However, this has accelerated the push by Huawei and other Chinese companies to develop similar chips locally. Also, some Chinese companies are coming up with unique solutions.
In January this year, for instance, Chinese AI company DeepSeek introduced R1, a tool that rivals ChatGPT, developed at a fraction of the cost. DeepSeek, based in Hangzhou, built the tool by using earlier versions of Nvidia's chips and its own software. It is likely that other Chinese companies are working on similar innovations. Also, rapid progress in chip design and algorithmic efficiency by Huawei and other Chinese companies is expected to make them strong competitors in AI hardware, within the next few years.
China's lead over India in AI is likely to persist, possibly even widen, based on comparative data on publications by researchers in the two countries, which provides an early indication. Although somewhat dated, a study of papers published from 2019–2023 show China's dominance across various AI fields. The contribution of Indian researchers, while improving, remains considerably less influential on the global stage.
Regarding propsects for the rapid advance of AI in India, optimistically assume that half of the $11.4 billion in funding for new technology businesses, announced by the Indian government last month, is allotted to AI. Then, will Indian AI companies boost salaries to hire top STEM graduates who currently pursue careers in the U.S. or with multinational companies in India?