What India can learn from Chinese DeepSeek saga – Firstpost

The big story in artificial intelligence (AI) is the latest model from DeepSeek, a research firm solely owned and funded by the Chinese hedge fund High-Flyer, called R1.

Touted as ‘AI’s Sputnik moment,’ the model has garnered global attention for rivalling OpenAI’s performance benchmarks at reportedly only 10 per cent of the costs. This is a paradigm shift, from brute-force generative AI that requires internet-scale training data and billions of dollars in capital to more optimised, inference-based intelligence systems that democratise costs and make the technology an accessible commodity for all downstream applications.

DeepSeek isn’t a one-off. 2025, thus far, has had China showcase its capabilities in not only AI but also sixth-generation fighter jets and nuclear fusion. India must introspect on its slow progress and why an innovation like this did not come from one of its technology hubs.

Creative Insecurity

Introducing ‘creative insecurity’ in his book, The Politics of Innovation, Mark Zachary Taylor proposes that a nation’s science and technology (S&T) missions are likely to get high priority when faced with external threats. Extending this to a company level, DeepSeek is also a case study of external constraints leading to innovative breakthroughs.

High-Flyer’s high dependence on AI for its quantitative trading strategies makes it vulnerable to geopolitical tensions. What started in 2017 when a single Chinese company, ZTE, faced restrictions for violating US sanctions has now increased in severity to a fully blown export control regime and decoupling in cross-border research and knowledge transfer.

The need for self-reliance amidst this adversarial competition has led to China’s breakneck pace of progress in multiple technological building blocks that position it well in critical sectors.

Additionally, it should not be presumed that High-Flyer enjoys preferred status under the Xi Jinping administration. Evidence points to the contrary.

Xi Jinping has primarily held a dismissive view of online platforms, including social media, gaming, and entertainment platforms, comparing them to ‘spiritual opium’ and arguing that they jeopardise the promotion of people’s “common prosperity”. His slogan, “Walk with two legs”, reflects his view that even experimental research should be rooted in productive national power goals. These are defined as achieving self-reliance, military modernisation, economic advancements for the Chinese people, and fundamental breakthroughs that position China as a world leader.

High-Flyer’s sustenance was dependent on the whims of the establishment. The Great Firewall meant that Western firms were never allowed to operate in China, and domestic tech too is subject to censorship and regulation. With the Chinese stock market in decline for four consecutive years, Beijing led a crackdown on several quant-trading firms for profiting at the detriment of retail investors. While many firms were shut down, High-Flyer reportedly suffered huge losses.

Chinese R&D ecosystem

Despite these obstacles, DeepSeek benefitted from a supportive structural Chinese research and development (R&D) ecosystem that existed for several decades. Science and technology (S&T) initiatives in China gained momentum in the 1980s under Deng Xiaoping’s national programs focused on capability building in critical technology sectors. Economic reforms allowed for international linkages—Chinese students were sent abroad, and efforts were made to position China as a leading destination for foreign researchers. Foreign investments into the country were encouraged, and national S&T parks were established. This was carried out in conjunction with intellectual property espionage, reverse engineering and exploitative transfer of technical know-how from foreign corporations operating in China.

Additionally, there was infrastructure support for foreign-funded R&D centres and platforms that integrate domestic companies with global innovation networks. Since 2015, China has strengthened its intellectual property (IP) by setting up IP courts in major cities to streamline the judicial process in line with a national IP strategy.

Xi Jinping’s vision for a ‘Chinese Dream’ and national rejuvenation is rooted in technological supremacy. The New Generation Artificial Intelligence Development Plan, launched in 2017, set ambitious goals for China to become a global power in the field by 2030. DeepSeek’s AI lab, if not its parent company, is likely to have gained from positive network efforts such as high-quality talent, tax incentives that reward R&D, and regulatory sandboxes that encourage experimentation.

It is this cross-border diffusion of ideas at the confluence of high technical talent, private enterprise, and commitment to R&D from which DeepSeek emerged.

DeepSeek’s choice to open source their work is interesting, not only because this would not have received global attention without it, but also because it appears to be a move in establishing street capability. Big Tech’s arrogant position has now been exposed for what it is, and its technological moat has come under question. Building India-specific models has now become accessible. But India’s challenge in AI or any other field is not related to a lack of capital. What India lacks is an innovative architecture. For instance, despite a high presence on open-source platforms, the number of contributions from Indian developers is glaringly small.

Technology Diffusion—Lessons for India

Jeffrey Ding’s General Purpose Technology (GPT) diffusion theory offers an illuminating framework for analysing innovation architecture. It highlights that the path to innovation and eventual economic prowess is paved in an enabling environment where technological knowledge can flow freely between academia, industry, and government, and there is an effective skill infrastructure.

The lesson to take away from China is not one of state control. This is often just the visible aspect of China’s success, and it obscures the fact that China has created an ecosystem where private sector ingenuity can flourish despite private aspirations coming into friction with the state regularly. India, however, is severely lacking in private-sector R&D investment. Even the pharmaceutical industry, which is the leading sector in India in this regard, pales in comparison to the global average. In contrast, High-Flyer reportedly reinvested 70 per cent of its annual revenues into R&D.

The Chinese model of a statist top-down technological view is one way of doing things. India lacks the state capacity to micromanage innovation at the national level. India’s high public sector investment, at par with developed countries, has not propelled the pace of innovation to what India needs. India must pivot toward strengthening the private sector. The much-maligned profit motive deserves rehabilitation in India’s national discourse on innovation. Despite the Chinese State’s efforts to channel S&T towards national priorities, DeepSeek originated in private profit and not any lofty national goals. Profit-driven research is good and efficient. It is from the apparent chaos of variation and natural selection that innovations emerge.

The government’s role should only be that of an enabler.

The US has shown with the examples of the Defence Advanced Research Projects Agency (DARPA) and NASA that the defence and space sectors are often the cradle of innovation. As highlighted by the concept of creative insecurity, increasing competition in these sectors is necessary. Many experts made the case for more government R&D work to be outsourced to the private sector right from the design phase. They’ve compared NASA, which outsources around 80 per cent of its work, with ISRO, which still does a lot of its work in-house, and recommended a fundamental reboot of the ecosystem.

From the GPT theory, it follows that India should establish innovation clusters where research institutions, startups, and established companies coexist in close proximity, facilitating knowledge spillovers, collaboration, and utilising the benefits of networks. These clusters should have significant autonomy while receiving infrastructure support and research funding from the government. The government can create subtle environmental pressures through strategic funding, regulatory frameworks, and market incentives but should resist the temptation to micromanage innovation pathways.

Innovation is a sector that has not yet been truly liberalised in India despite the reforms of 1991. Unless this is unleashed, cutting-edge research contributions from India will not be a reality.

Shobhankita Reddy and Arindam Goswami are Research Analysts at the High Tech Geopolitics Programme at the Takshashila Institution, Bengaluru. The views expressed in the above piece are personal and solely those of the authors. They do not necessarily reflect Firstpost’s views.