The artificial intelligence (AI) flight is taking off, and DeepSeek is the final call for India to show up at the boarding gate. Since its private sector is too risk-averse to back research projects with uncertain payoffs, the state would have to step up.
The Chinese startup’s AI models, which it began offering last month as open-source licenses, have been built at a fraction of the cost of resource-intensive rivals like OpenAI’s ChatGPT or Google’s Gemini. The US tech industry and Wall Street investors are rightly viewing DeepSeek as a disruptor. Silicon Valley firms are putting down hundreds of billions of dollars to fend off the challenge. Italy, Australia, South Korea, Taiwan, Texas and New York have banned DeepSeek’s AI assistant from government devices.
India is watching the economic and political contest from the sidelines. That is dangerously complacent.
Illustration: Constance Chou
Unlike in manufacturing or transportation, where it has already ceded a large lead to its neighbor, here is a race that is still wide open, but unless the most-populous nation puts its globally acknowledged edge in software programming to work, the moment would pass it by.
DeepSeek is threatening the one comparative advantage the country has assiduously worked toward in the 21st century: code-writing on an industrial scale.
“Let the code write itself,” DeepSeek Coder says. The AI-based programming assistant, which has already made a splash in the development community, is bound to become more capable in the weeks and months ahead. Using such tools, a small, highly skilled section of India’s 5 million code-writers would become tremendously more productive, but for the majority, this would be bad news.
Homegrown models, if they are cheaper to license than foreign AI, would get deployed faster. A surge in efficiency across industries would absorb the workers displaced from code-writing jobs. Putting shiny new tools in the hands of the more than 2.5 million PhDs minted annually in science, technology, engineering and medicine could lead to breakthroughs that end up lifting the economy. This is the link that policymakers are refusing to see. The US$1.2 billion that New Delhi committed last year to the India AI Mission is a fraction of a US$24 billion subsidy program for manufacturers.
It has been nearly two years since Sam Altman was asked at a conference if an Indian start-up could build a foundational model, trained on vast datasets and capable of multiple applications, for US$10 million. The founder of OpenAI said that such an undertaking would be “completely hopeless.”
It is only natural for Altman to say that — nobody wants their second-largest market by number of users to become a competitor, but what explains the lack of confidence on the Indian side? Even with DeepSeek calling Silicon Valley’s bluff on costs, Indian tech companies are reluctant to take up foundational work in generative AI because success is not guaranteed.
There is some investment going into adapting existing models to handle local languages, but foreign alternatives that would offer the same Indian-language capabilities plus a whole lot more would likely overshadow these efforts, said Nilesh Jasani, founder of GenInnov, a Singapore-based global innovation fund.
China has ramped up its share of top-tier global AI talent to 47 percent, compared with 18 percent for the US. India’s figure is just 5 percent because most of its talent ends up migrating, largely to the US.
Multinationals like GE Aerospace are doing sophisticated applied research in India using local engineering talent, but there simply is not a supporting environment for deep fundamental research, and no urgency to build one.
This is the situation eight years after a team of Google researchers gave shape to the language-processing architecture behind today’s models. “Attention Is All You Need,” was the title of their paper. Clearly, local tech policymakers were not very attentive to the technology’s potential, even though two of the eight scientists involved in that pathbreaking project are Indian-born.
If the public sector has been lacking in attention, the private sector has not been short on intent. Mukesh Ambani, the country’s richest tycoon, recently announced that he would build the world’s largest data center. He is buying Nvidia Corp chips, which would be crucial for training models.
However, infrastructure alone would not be enough. Breakthroughs in natural language processing would come from hundreds of attempts at model-building, each costing a few million dollars. The software outsourcing companies of Bengaluru, India’s Silicon Valley, should have been at the forefront of this initiative because of the threat generative AI poses to their bread-and-butter activity of code-writing for global corporations.
Yet they are not keen on taking bold bets. Their current business is still rewarding large shareholders with fat dividend checks and share buybacks, with little new investment. There is a limited appetite for moonshots.
The idea that India is the next China — running just about a decade behind it — was popular about 20 years ago. Back then, people wrote books about “Chindia.” The South Korean financial industry even launched funds covering the two most populous nations as a single theme.
Two decades later, there is not a lot left of that illusion of comparability. While India did attend to its basic infrastructure shortages like electricity and roads, China raised the gap between the two economies by lifting its technology game. Its arc of global dominance, which just 10 years ago was limited to a few industries like drones and solar panels, has extended to electric vehicles, high-speed trains and now, generative AI.
So when Indian tech companies say that they would profit from growing AI adoption by making customized digital assistants for global corporations, they are ignoring the inevitable progress of generative AI to artificial general intelligence. Models that rival human cognitive abilities would handle most programming tasks on their own. Either OpenAI, or DeepSeek, would get there.
Or, another model would succeed. For the South Asian country to not become a permanent importer of AI tools, it needs to own foundational technologies, built with massive government support for research universities and institutions.
Above all, tech policymakers have to shake off their defeatist fatalism and heed the challenge facing them. Attention is everything.
Andy Mukherjee is a Bloomberg Opinion columnist covering industrial companies and financial services in Asia. Previously, he worked for Reuters, the Straits Times and Bloomberg News. This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
A gap appears to be emerging between Washington’s foreign policy elites and the broader American public on how the United States should respond to China’s rise. From my vantage working at a think tank in Washington, DC, and through regular travel around the United States, I increasingly experience two distinct discussions. This divergence — between America’s elite hawkishness and public caution — may become one of the least appreciated and most consequential external factors influencing Taiwan’s security environment in the years ahead. Within the American policy community, the dominant view of China has grown unmistakably tough. Many members of Congress, as
After declaring Iran’s military “gone,” US President Donald Trump appealed to the UK, France, Japan and South Korea — as well as China, Iran’s strategic partner — to send minesweepers and naval forces to reopen the Strait of Hormuz. When allies balked, the request turned into a warning: NATO would face “a very bad” future if it refused. The prevailing wisdom is that Trump faces a credibility problem: having spent years insulting allies, he finds they would not rally when he needs them. That is true, but superficial, as though a structural collapse could be caused by wounded feelings. Something
Former Taipei mayor and Taiwan People’s Party (TPP) founding chairman Ko Wen-je (柯文哲) was sentenced to 17 years in prison on Thursday, making headlines across major media. However, another case linked to the TPP — the indictment of Chinese immigrant Xu Chunying (徐春鶯) for alleged violations of the Anti-Infiltration Act (反滲透法) on Tuesday — has also stirred up heated discussions. Born in Shanghai, Xu became a resident of Taiwan through marriage in 1993. Currently the director of the Taiwan New Immigrant Development Association, she was elected to serve as legislator-at-large for the TPP in 2023, but was later charged with involvement
Out of 64 participating universities in this year’s Stars Program — through which schools directly recommend their top students to universities for admission — only 19 filled their admissions quotas. There were 922 vacancies, down more than 200 from last year; top universities had 37 unfilled places, 40 fewer than last year. The original purpose of the Stars Program was to expand admissions to a wider range of students. However, certain departments at elite universities that failed to meet their admissions quotas are not improving. Vacancies at top universities are linked to students’ program preferences on their applications, but inappropriate admission
RSS